The American Naturalist

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

sent to the Editor of THE AMERICAN NATURALIST, Garrison- -on-Husson, New York.

icles containing summaries of research work bearin ng on the

pr parres aie organic evolution are especially welcome, and will be given srererince

n publica

One hundred reprints of ppb are supplied to authors free of charge.

Further reprints 3 will be supplied atc

Su ns and paa at should be sent to the publishers. The

subscription arias is four dollars a year. Foreign postage is fifty cents and

Canadian postage twenty-five cents additional. The charge for single copies is

forty. cents. The advertising rates are Four Dollars for a page.

THE SCIENCE PRESS

Lancaster, Pa. Garrison, N. Y.

_ NEW YORK: Sub-Station 84

Entered tter, April 2, 1908, at t Lancaster, Pa., under the Act of

Congress of March 3, 1879.

BO o KS FOR S ALE | 1APAN NATURAL HISTORY SPECIMENS

Perfect Condition and Lowest Prices.

Specialty: Bird Skins, Oology, Entomology, Marine

NATURAL HISTORY Animals and others. Catalogue free. Correspond-

Catalogue on ence solicited.

JOHN D. SHERMAN, w] T. FUKAI, Naturalist,

403 Seneca MOUNT VERNON, N N. Y. Konosu, Saitama, Japan

PUBLISHED BY THE CAMBRIDGE UNIVERSITY PRESS

| MIMICRY IN BUTTERFLIES

E Bound in Buckram. Royal 8 Svo, Pp. vii+188. With 16 plates (12 of which

eG lored). Price $4.50.

oe > REGINALD (niona PUNNETT, FES, Fellow of Gonville and Caius College,

X ye ARTHUR t BALFOUR, Professor of Genetics in the University of =

| -EXTRACT FROM THE PREFACE

- - ‘ -2E SR has — written tothe ne T it may appeal to > several

RET remedy this deficiency mney that athe |

3 ib tht i mill be found o : of interest to those who ho live in or visit mra : =

THE

AMERICAN NATURALIST

Vor. L: January, 1916 No. 589

THE EVOLUTION OF THE CELL?

BY THE LATE PROFESSOR E. A. MINCHIN, F.R.S.

Waen addressing an audience of biologists it would be

superfluous to insist upon the importance of the study of

the cell and its activities. It is now recognized almost

universally that the minute corpuscles known by the some-

what unsuitable term ‘‘cells’’ are the vital units of which

the bodies of animals and plants are built up, and that all

distinctive vital processes—metabolism, growth and re-

production, sexual phenomena and_heredity—reduce

themselves ultimately to activities taking place in, and

carried on by, the individual cells which build up the

body as a whole. Each cell must be regarded as a living,

individual organism which, however much it may be spe-

cialized for some particular function or form of vital ac-

tivity, is capable of maintaining its life and existence in a

suitable environment by carrying on all the necessary

processes of metabolism which are the essential and dis-

tinctive characteristics of living beings. In the case of

cells composing the complex body of the higher anirnals

and plants the cells are mutually interdependent, and,

with the exception of the mature germ-cells, can not main-

tain their existence apart from their fellows; that is to

say, the only natural? environment suitable for their con- —

1 Address by the President to the Zoological Section of the British Asso-

ciation for the Advancement of Seianee. Manchester, 1915. :

2 Tt is not necessary to do more than refer here to the investigations that

have been carried on in recent ss at vem esr sodasi oe ; 8

plication of tissue-cells body in artificial cult

6 THE AMERICAN NATURALIST [ Vou. L

tinued existence is the complex body or cell-common-

wealth of which they form an integral part. But in the

simplest forms of life the whole body of the living indi-

vidual may reach no higher degree of complexity than

the single cell, which is then seen as an organism physio-

logically complete in every respect, living a free and in-

dependent life in Nature and competing with other or-

ganisms of all kinds, simple or complex, in the universal

struggle for existence amongst living beings. This state-

ment of the ‘‘cell-theory’’ is that with which, I believe, the

majority of modern biologists would agree; not without,

however, some dissentients, amongst whom I personally

am not to be numbered.?

The fundamental importance of the cell as a complete

living organism, whether maintaining itself singly and in-

dependently or in union with other similar but individually

specialized units, has made it the object of intensive and

concentrated study, not only by those who group them-

selves according to their special points of view as zoolo-

gists, botanists, physiologists, ete., but also by a class of

investigators who take the cell itself as the subject of a

branch of biological investigation termed cytology, which

deals with cells in a general manner independently of

their provenance, whether animal or vegetable. Some

knowledge of the cell and its activities is necessary at the

present time for every one concerned with the study of

living things, whether that study is pursued for its own

sake and with disinterested objects, or with the intention

of applying scientific principles to practical aims, as in

medicine or agriculture. One might have expected, there-

fore, that at least some elementary understanding of the

nature and significance of the cell, and the importance of

cellular activities in the study of life and living things,

would have formed at the present time an indispensable

part of the stock of knowledge acquired by all intelligent

persons who are ranked as ‘‘educated’’ in popular esti-

These experiments afford strong support to the view that the cell is to be

regarded primarily as an independent living organism,

3 See Appendix A.

No. 589] THE EVOLUTION OF THE CELL T

mation. Unfortunately this is so far from being the case

that it is practically impossible, in this country at least,

to find any one amongst the educated classes to whom the

words ‘‘cell’’ and ‘‘cytology’’ convey any meaning at all,

except amongst those who have interested themselves

specially in some branch of biology. Consequently, any

discussion concerning the cell, although it may deal with

the most elementary processes of life and the fundamental

activities and peculiarities of living beings, ranks in popu-

lar estimation as dealing with some abstruse and recon-

dite subject quite remote from ordinary life and of inter-

est only to biological specialists. It must, however, be

pointed out that the general state of ignorance concern-

ing these matters is doubtless in great part due to the fact

that an objective acquaintance with cells can not be ob-

tained without the use of expensive and delicate optical

instruments.

I propose in this address to deal with an aspect of cytol-

ogy which appears to me not to have received as yet the

attention which it deserves, namely, the evolution of the

cell itself and of its complex organization as revealed by

the investigation of cytologists. Up to the present time

the labors of professed cytologists have been directed

almost entirely towards the study of the cell in its most

perfect form as it occurs in the Metazoa and the higher

plants. Many cytologists appear indeed to regard the

cell, as they know it in the Metazoa and Metaphyta, as

the beginning of all things, the primordial unit in the

evolution of living beings.* For my part I would as soon

postulate the special creation of man as believe that the

Metazoan cell, with its elaborate organization and its ex-

traordinarily perfected method of nuclear division by

karyokinesis, represents the starting-point of the evolution

4 For example, my friend Dr. C. E. Walker, in an article in Science Prog-

ress (Vol. VII, p. 639), after stating that ‘‘The unit of living matter, so

far as we know, is the cell,’’ proceeds to deal with ‘‘that form in which it is

found in the multicellular and the majority of unicellular organisms, both

animal and vegetable’’ and then describes the typical cell of the cytologist,

with nucleus, cytoplasm, centrosome, chrondriosomes, and reproduction with

fully developed karyokinesis,

8 THE AMERICAN NATURALIST [ Vou. L

of life. So long, however, as the attention of cytologists

is confined to the study of the cells building up the bodies

of the higher animals and plants, they are not brought

face to face with the stages of evolution of the cell, but are

confronted only with the cell as a finished and perfected

product of evolution, that is to say, with cells which, al-

though they may show infinite variation in subordinate

points of structure and activity, are nevertheless so fun-

damentally of one type that their plan of structure and

mode of reproduction by division can be described in gen-

eral terms once and for all in the first chapter of a bio-

logical text-book or in the opening lecture of a course of

elementary biology.

One of the most striking features of the general trend

of biological investigation during the last two decades

has been the attention paid to the Protista, that vast as-

semblage of living beings invisible, with few exceptions,

to the unassisted human vision and in some cases minute

beyond the range of the most powerful microscopes of

to-day. The study of the Protista has received in recent

years a great stimulus from the discovery of the im-

portance of some of the parasitic forms as invaders of

the bodies of men and animals and causers of diseases

often of a deadly nature; it has, however, yielded at the

same time results of the utmost importance for general

scientific knowledge and theory. The morphological

characteristic of the Protista, speaking generally, is that

the body of the individual does not attain to a higher de-

gree of organization than that of the single cell. The ex-

ploitation, if I may use the term, of the Protista, though

still in its initial stages, has already shown that it is

amongst these organisms that we have to seek for the

forms which indicate the evolution of the cell, both those

lines of descent which lead on to the cell as seen in the

Metazoa and Metaphyta, as well as other lines leading in

directions altogether divergent from the typical cell of

the text-book. We find in the Protista every possible condi-

tion of structural differentiation and elaboration, from cells

No. 589] THE EVOLUTION OF THE CELL 9

as highly organized as those of Metazoa or even, in some

cases, much more so, back to types of structure to which

the term cell can only be applied by stretching its mean-

ing to the breaking-point. Already one generalization of

eytologists has been torpedoed by the study of the Pro-

tista. The dictum ‘‘Omnis nucleus e nucleo’’ is perfectly

valid as long as it is restricted to the cells of Metazoa and

Metaphyta, to the material, that is to say, to which the

professed cytologist usually confines his observations.’

But in the Protista it is now well established that nuclei

ean arise de novo, not from preexisting nuclei but from

the extranuclear chromatin for which Hertwig first coined

the term ‘‘chromidia.”’

It is clear, therefore, that the results already gained

from the study of the Protista have brought about a new

situation which must be faced frankly and boldly. It is

impossible any longer to regard the cell as seen in the

Metazoa and as defined in the text-books as the starting-

point of organic evolution. It must be recognized that

this type of cell has a long history of evolution behind it,

which must be traced out, so far as the data permit. The

construction of phylogenies and evolutionary series is of

course purely speculative, since these theories relate to

events which have taken place in a remote past, and which

ean only be inferred dimly and vaguely from such frag-

ments of wreckage as are to be found stranded on the

sands of the time in which we live. Many important

stages of evolution may be totally submerged and no

longer available for study and consideration. The ex-

tent to which such speculations will carry conviction to

a reasonable mind will depend entirely on the stores of

5 Vejdovsky (‘‘Zum Problem der Vererbungstriiger,’’ Prag, 1911-1912, .

120) has already maintained, for the cells of Me F1

aphorism ‘‘ Omnis nucleus e nucleo’’ should be changed to ‘‘Omnis case

e chromosomatis’’ [sic], on the ground that the nucleus, as such, is not an

original cell-component ‘‘ but is produced secondarily from the chromosomes

of the mother-cell.’’ If this is true, there is but little difference in detail,

and none in principle, between the formation of ‘‘secondary’’ nuclei from

chromidia and the reconstruction of a daughter-nucleus from chromosomes

in the most perfected form of karyokinesis.

10 THE AMERICAN NATURALIST (Vou. L

data that can be collected and which must be the last ap-

peal for the cogency of all arguments and judgments.

The study of the Protista is as yet in its infancy; groups

have been recognized and have received ponderous desig-

nations, although their very existence is yet in doubt, as

in the case of the so-called Chlamydozoa; and our knowl-

edge of the affinities and mutual relationships of the

groups is still very imperfect. All attempts, therefore,

to trace the evolution of the Protista must be considered

as purely tentative at present. If I venture upon any

such attempt, it is to be regarded as indicating a firm be-

lief on my part that the evolution of the cell has taken

place amongst the Protista, and that its stages can be

traced there, rather than as a dogmatic statement that

the evolution has taken place in just the manner which

seems to me most probable. When we reflect on the ir-

reconcilable differences of opinion amongst zoologists

with regard to the origin and ancestry of vertebrates, for

example, we may well be cautious in accepting pedigrees

in Protista.

Before, however, I can proceed to deal with my main

subject, it is absolutely necessary that I should define

clearly the sense in which I propose to use certain terms,

more especially the words ‘‘cell,’’ ‘‘nucleus,’’ ‘‘chroma-

tin,’’ ‘‘protoplasm’’ and ‘‘eytoplasm.’’ Unless I do so

my position is certain to be misunderstood, as, indeed, it

has been already by some of my critics.

` The term cell was applied originally by botanists to the

single chambers or units of the honeyecombed structure

seen in the tissues of plants. The application of the term

to such structures is perfectly natural and intelligible,

since each such cell in its typical form is actually a closed

space limited by firm walls, and containing a relatively

large quantity of fluid cell-sap and a small quantity of

the slimy protoplasmic substance. When these struc-

tures were first discovered, the limiting membrane or wall

of the cell was regarded as essential, and less importance

was attached to its contents. With increased knowledge,

No. 589] THE EVOLUTION OF THE CELL 11

however, and especially when animal tissues came to be

studied, it became apparent that the cell-wall, like the

fluid cell-sap, was a secondary product, and that the es-

sential and primary part of the cell was the viscid proto-

plasmic substance, in which a peculiar body, the ‘‘nu-

cleus,’’ or kernel, was found to be universally present.

Consequently the application and meaning of the term

cell had to undergo an entire change, and it was defined as

a small mass or corpuscle of the living substance, proto-

plasm, containing at least one nucleus. To these essen-

tial constituents other structures, such as a limiting mem-

brane or cell-wall, and internal spaces—vacuoles—filled

with watery fluid, might be added as products of the sec-

retory or formative activity of the living substance; but

such structures were no longer regarded as essential to

the definition of the cell, since in many cases they are not

present. It is to be regretted in some respects that with

this changed point of view the term ‘‘cell,’’ used orig-

inally under a misapprehension, was not replaced by some

other term of which the ordinary significance would have

been more applicable to the body denoted by it.®

The chief point that I wish to establish, however, is

that the term cell was applied originally to the protoplas-

mic corpuscles building up the bodies of the Metazoa and

Metaphyta, each such corpuscle consisting of a minute

individualized mass of the living substance and contain-

ing a nucleus. Hence a complete cell is made up of two

principal parts or regions, the nucleus and the remainder

of the protoplasmic body, termed the cytoplasm. By

some authors the term protoplasm is restricted to the

cytoplasmic portion of the cell, and protoplasm is then

contrasted with nucleus; but it is more convenient to con-

sider the whole cell as composed of protoplasm divided

into two regions, nucleus and cytoplasm.

We come now to the consideration of the body termed

6‘*Nothing could be less appropriate than to call such a body a ‘cell’;

yet the word has become so firmly established that every effort to replace it

by a better has failed, and it probably must be accepted as part of the

established nomenclature of science.’’—E. B. Wilson, ‘ ‘The s p. 19.

| THE AMERICAN NATURALIST [ Vou. L

the nucleus, which undoubtedly possesses an importance in

the life and functions of the cell far greater than would

be inferred from the name given to it. A nucleus, as

seen in its typical form, has a limiting membrane enclos-

ing a framework composed of a substance termed ‘‘linin.”’

The framework has the form of a network, which is prob-

ably to be interpreted, primitively at least, as the optical

expression of an alveolar structure similar to that seen

also in the cytoplasm, but of coarser texture, and the ap-

parent ‘‘threads’’ of the linin-framework may then be

the optical sections of the partitions between neighboring

alveoli. Such an interpretation does not exclude the pos-

sibility of the formation of real threads or fibers in the

framework in certain cases or during particular periods

of nuclear activity; just as fibrous structures may arise

in the alveolar cytoplasm also. The cavities of the frame-

work contain a watery fluid or nuclear sap, probably of

the same nature as the fluid enchylema or cell-sap con-

tained in the alveolar framework of the cytoplasm. At

the nodes of the alveolar framework are lodged grains or

masses of chromatin, a substance which must engage our

most particular attention, since it is the essential constit-

uent of the nucleus, universally present in all nuclei,

whether of the simplest or of the most complex types.

In addition to the chromatin-grains, which are distributed

in various ways over the linin-framework, there are to be

found usually one or more masses termed nucleoli, com-

posed of a material which differs from chromatin in its

reactions and has been termed plastin.

In the foregoing paragraph I have described in general

terms the typical nucleus of the text-books, as found com-

monly in the cells that build up the bodies of ordinary

animals and plants. The minutie of the details of struc-

ture and arrangement of the constituent parts may vary

infinitely, but the type remains fairly constant. When

we come, however, to the nuclei of the Protista, such pro-

nounced modifications and variations of the type are met

with that a description in general terms is no longer pos-

No. 589] THE EVOLUTION OF THE CELL 13

sible. I shall deal with some of these types later in my

attempts to reconstruct the evolution and phylogeny of

the cell. I will draw attention now only to a few salient

points. In the Protist cell the chromatin is not neces-

sarily confined to the nucleus, but may occur also as extra-

nuclear grains and fragments termed chromidia, scattered

through the protoplasmic body; and the chromatin may

be found only in the chromidial condition, a definite nu-

cleus being temporarily or permanently absent. Fur-

ther, when a true nucleus is present in the Protist body,

it seldom contains a nucleolus of the same type as that

seen in the nuclei of tissue-cells, that is to say, a mass of

pure plastin, but in its place is found usually a conspicu-

ous body which shows reactions agreeing more or less

closely with those of chromatin and which consists of a

plastin-basis more or less densely impregnated with

chromatin. Such a body is termed a karyosome (or

chromatin-nucleolus) to distinguish it from the true nu-

cleoli (plastin-nucleoli) characteristic of tissue-cells. Ac-

cording as the plastin or the chromatin predominates in

the composition of a karyosome, its reactions may re-

semble more nearly those of a true nucleolus in the one

case, or those of chromatin in the other. The so-called

karyosomatic type of nucleus is very common in the Pro-

tista, but by no means of invariable occurrence; in many

cases the nucleus consists of a clump of small grains of

chromatin, with no distinct karyosome, or with a karyo-

some which consists mainly of plastin. Thus two ex-

treme types of nuclear structure can be distinguished and

may be termed provisionally the karyosomatic type and

the granular type, ignoring for the sake of convenience in

nomenclature the types of structure transitional between

the two; as, for example, types in which a distinct karyo-

some is seen together with more or fewer peripherally

arranged grains of chromatin.

In either the karyosomatic or the granular type of

Protist nucleus we may find great simplification of the

complex type of nuclear structure seen in the tissue-cells

14 THE AMERICAN NATURALIST [ Vor. L

of animals and plants. Thus in the first place a distinct

nuclear membrane may be entirely absent and the chro-

matin-elements, whether occurring in the form of a com-

pact karyosome or of a clump of grains, are lodged simply

in a vacuole in the cytoplasm, that is to say in a cavity

containing a watery fluid of nuclear sap in which the

mass or masses of chromatin are suspended. It is a moot

point, to which I shall return again, whether in nuclei

of this simple type the linin-framework may sometimes

be absent altogether, or whether it is invariably present

in at least a rudimentary form, appearing as delicate

threads (in optical section) extending from the chromatin-

masses to the limiting wall of the nuclear vacuole, or be-

tween the grains of chromatin themselves. When such a

framework can be detected, the nucleus acquires the ap-

pearance, in preserved preparations at least, of possess-

ing a definite structure and is often termed a resting

nucleus; many observations have shown, however, that

the nucleus during life is undergoing continual internal

movements and re-arrangements of its parts and is by

no means at rest. The linin-framework can not, there-

fore, be regarded in any way as a rigid skeleton, but must

be interpreted as an alveolar framework similar to that

of the general protoplasm and equally liable to move-

ment, displacement and change.

From this survey, necessarily most brief and super-

ficial, of the manner in which the nuclei of Protists may

vary from the type of nucleus described in the text-books,

it is at once evident that the essential part of the nucleus

is the chromatin, and that the other structural constitu-

ents of the nucleus, namely, membrane, framework, and

plastin or nucleolar bodies, are to be regarded as acces-

sory components built up round, or added to, the primary

nuclear material, the chromatin. Even with regard‘to

the nuclei of Metazoa it is maintained by Vejdovsky that

at each cell-generation the entire nucleus of the daughter-

cell is produced from the chromosomes alone of the

No. 589] THE EVOLUTION OF THE CELL 15

mother cell.” The simplest body which can be recognized

as a nucleus, distinct from the chromidia scattered with-

out order or arrangement throughout the protoplasmic

body, is a mass of chromatin or a clump of chromatin-

grains supported on a framework and lodged in a special

vacuole in the cytoplasm. The complexity seen in the

most perfect type of nucleus takes origin by progressive

elaborations of, and additions to, a structure of this

simple and primitive type.

This brings me to a point which I wish to emphasize

‘most strongly, namely, that the conception of a true cell-

nucleus is essentially a structural conception. A nucleus

is not merely an aggregation of chromatin; it is not

simply a central core of some chemical substance or

material differing in nature from the remainder of the

protoplasm. As Dobell has well expressed it, a pound

of chromatin would not make a nucleus. The concepts

‘‘nucleus’’ and ‘‘chromatin”’ differ as do those of ‘‘table’’

and ‘‘wood.’’ Although chromatin is the one universal

and necessary constituent entering into the composition

of the cell-nucleus, a simple mass of chromatin is not a

nucleus. <A true nucleus is a cell-organ, of greater or

less structural complexity, which has been elaborated

progressively in the course of the evolution of the cell;

7 Walker, on the other hand, considers that ‘‘it seems quite possible that

the chromatin is merely a secretion of the linin.’’ (Science Progress, Vol.

VII, p. 641.) I doubt whether there are many cytologists who would admit

this possibility, and I think that very few protistologists would assent to

any such notion, since in the nuclei of Protista the linin-framework is in

many cases very little in evidence, if present at all.

8 Professor Armstrong writes: ‘‘Every organism must possess some kind

of nucleus, visible or invisible: some formative center round which the

various templates assemble that are active in directing the growth of the

organism.’’ (Science Progress, Vol. VII, p. 328.) I need hardly point

out that a chemical nucleus of this kind is not in the least what the biologist

Saker means by the term cell-nucleus. The one is a subjective postu-

te necessary for the comprehension of the activities of any speck of living

aaa or any portion, however minute, of a living organism; the other is a

concrete structure, known to us by actual observation, and as much an in-

tegral part of the true cell, considered as a definite type of organism, as a

backbone or its morphological equivalent is essential to the definition of a

true verte l

16 THE AMERICAN NATURALIST [Vou. L

it is as much an organ of the cell as the brain is an organ

of the human body. As a definite cell-organ, it performs

in the life and economy of the cell definite functions,

which it is the province of the cytologist to observe and

to study, and if possible to elucidate and explain. As an

organ of the cell, however, it has no homologue or ana-

logue in the body of the multicellular animals or plants;

there is no organ of the human body, taken as a whole,

similar or comparable to the nucleus of the cell. Conse-

quently, in studying the functions of the nucleus the

human cytologist finds himself in the same difficult posi-

tion that an intelligent living being lacking the sense of

sight would be when trying to discover the function of

visual organs in other organisms possessing that sense.

There is no organ of known and understood functions

with which the cytologist can compare the cell-nucleus

directly.

The foregoing brief consideration of the nucleus leads

me now to discuss in more detail the nature and proper-

ties of the essential nuclear substance, the so-called chro-

matin. To define, or characterize adequately, this sub-

_ stance is a difficult task. The name chromatin is derived

-from the fact that this substance has a peculiar affinity

for certain dyes or stains, so that when a cell is treated

with the appropriate coloring reagents—with so-called

nuclear stains—the chromatin in the nucleus stands out

sharply, by reason of being colored in a different manner

from the rest of the cell. In consequence, the statement

is frequently made, in a loose manner and without reflec-

tion, that chromatin is recognized by its staining reac-

tions, but in reality this is far from being true. When a

preparation of an ordinary cell is made by the methods

of technique commonly in use, the chromatin is recog-

nized and identified by its position in a definite body with

characteristic structure and relations to the cell as a

whole, namely the nucleus, and this is equally true

whether the chromatin has been stained or not. When

No. 589] THE EVOLUTION OF THE CELL 17

the cell has been stained with one of the dyes ordinarily

in use for coloring the chromatin, there are often seen in

the cytoplasm grains that are colored in exactly the same

manner as the chromatin-grains lodged in the nucleus.

Is an extranuclear grain which stains like chromatin to

be identified, ipso facto, as chromatin? By no means; it

may or it may not be chromatin. Simple inspection of a

stained preparation is altogether inadequate to deter-

mine whether such a body is or is not chromatin. Any

so-called chromatin-stain colors many bodies which may

occur in a cell besides the chromatin, and it may be nec-

essary to try a great many different stains before a com-

bination is found which will differentiate a given cyto-

plasmic enclosure from a true chromatin-grain by its

color-reactions. The so-called volutin-grains, for ex-

ample, which are found commonly in the cytoplasm of

many Protists, are identified by the fact that they have a

stronger affinity for ‘‘chromatin-stains’’ than chromatin

itself. a

When, moreover, chromatin is compared with regard

to its staining-reactions, both in different organisms, and

in the same organism at different times, it is found to

react very differently to one and the same stain. A stri-

king example of this capriciousness is seen when a pre-

served film is made of the blood of some vertebrate which

has nucleated blood-corpuscles, such as a bird or fish, and

which contains also parasitic trypanosomes. It is easy

to stain the nuclei of the blood-corpuscles with various

stains, as, for example, carmine-stains such as picro-car-

mine or alum-carmine, which will not color the nuclei of

the trypanosomes in the slightest. Moreover, every cy-

tologist knows that the ‘‘chromaticity’’ of the chromatin

varies enormously in different phases of the nuclear cycle

of generation; it is often difficult to stain the chromatin

in the ‘‘resting’’ nucleus, but the first sign of impending

nuclear division is a marked increase in the staining

powers of the chromatin. There is no dye known which

can be relied upon to stain chromatin always, or wherever

18 THE AMERICAN NATURALIST [ Vou. L

it occurs. Methyl-green has been claimed to be the most

reliable and certain of nuclear stains, but R. Hertwig, in

his classical researches upon Actinospherium, showed

that it sometimes fails to stain chromatin. It is perfectly

conceivable that there might be varieties of chromatin

which could not be stained by any dye whatsoever.

I have felt bound to insist strongly upon the inadequacy

of staining-methods for the detection and identification

of chromatin, well known though these facts are to every

eytologist, because here also I note a tendency amongst

biological chemists to regard staining-properties as the

sole criterion of chromatin. In reality such properties

are of entirely secondary importance. To use the ter-

minology of formal logic, staining-properties are an ‘‘ac-

cident,’’ though it may be an ‘‘inseparable accident,’’ of

chromatin, not a ‘‘difference’’ which can be used to frame

a logical definition, per genus et differentias, of this sub-

stance. If chromatin were nothing more than ‘‘stainable

substance,’’ as Professor Armstrong terms it,® some of

the most important results of cytological investigation

would be deprived of all real significance and reduced to

the merest futilities.

What then is the true criterion of the chromatin-sub-

stance of living organisms? From the chemical point of

view the essential substance of the cell-nucleus would ap-

pear to be characterized by a complexity of molecular

structure far exceeding that of any other proteins, as

well as by certain definite peculiarities. Especially char-

acteristic of chromatin is its richness in phosphorus-com-

pounds, and it stands apart also from other cell-elements

in its solvent reactions, for example, resistance to peptic

digestion. E. B. Wilson, in his well-known treatise, has

emphasized the ‘‘cardinal fact... that there is a definite

and constant contrast between nucleus and cytoplasm.”’

The outstanding feature of the nucleus is the constant

presence in abundance of nuclein and nucleoproteins.

Nuclein, which is probably identical with chromatin, is a

9 Science Progress, Vol. VII, p. 327.

No. 589] THE EVOLUTION OF THE CELL 19

complex albuminoid substance rich in phosphorus. It is

the phosphorus-content of chromatin that is its most

characteristic chemical peculiarity as contrasted with the

cytoplasm. How far these features are common, how-

ever, to all samples of chromatin in all types of living

organisms universally, can not, I think, be stated definitely

at present; at any rate, it is not feasible for a cytologist

of these days to identify a granule in a living organism

or cell as chromatin solely by its chemical reactions,

although it is quite possible that at some future time

purely chemical tests will be decisive upon this point—a

consummation devoutly to be wished.

The only criterion of chromatin that is convincing to

the present-day biologist is the test of its behavior, that

is to say, its relations to the life, activity and develop-

ment of the organism. I may best express my meaning

by objective examples. If I make a preparation of Arcella

vulgaris by suitable methods, I see the two conspicuous

nuclei and also a ring of granules lying in the cytoplasm,

stained in the same manner as the chromatin of the

nuclei. Are these extranuclear granules to be regarded

also as chromatin? Yes, most decidedly, because many

laborious and detailed investigations have shown that from

this ring of granules in Arcella nuclei can arise, usually

termed ‘‘secondary’’ nuclei for no other reason than that

they arise de novo from the extranuclear chromatin and

quite independently of the ‘‘primary’’ nuclei. The sec-

ondary nuclei are, however, true nuclei in every respect,

as shown by their structure, behavior and relations to

the life-history of the organism; they may fuse as nuclei

of gametes (pronuclei) in the sexual act and they become, |

with or without such fusion, the primary nuclei of future

generations of Arcella; they then divide by karyokinesis

when the organism reproduces itself in the ordinary way

by fission, and are replaced in their turn by new secondary

nuclei at certain crises in the life-history. In view of

these facts it can be asserted without hesitation that the

ring of staining granules in ne is oopan of, or at

20 THE AMERICAN NATURALIST [Vou L

least contains, true chromatin-grains, extranuclear chro-

matin for which R. Hertwig’s term chromidia is now used

universally. It is interesting to note that until the life-

history of Arcella was studied in recent times the con-

spicuous ring of chromidia was generally overlooked and

is not shown in some of the older pictures of the organism.

If, on the other hand, I make a preparation of some

unidentified amceba occurring casually in pond-water or

in an infusion, and find in its cytoplasm certain grains

staining in same manner as the chromatin of the nucleus,

it is quite impossible, without a knowledge of the life-

history of the organism, to assert definitely that the grains

in question are or are not true chromidia. They might

equally well turn out to be volutin or any other substance

that has an affinity for the particular chromatin-stains

used in making the preparation.

The fact that at the present time the only decisive

criterion of what is or is not chromatin is supplied only

by its behavior in the life-history and its relation to the

organism, makes it much easier to identify the chromatin

in some cases than in others. In those Protista or cells

which contain, during the whole or a part of the life-

history, one or more true nuclei, recognizable as such un-

mistakably by their structure and their ch teristic

relations to the reproductive and sexual phenomena of

the organism, the chromatin can be identified with cer-

tainty. If chromidia occur in the cell-body in addition

to true nuclei or even if the nuclei are temporarily absent

during certain crises of the life-history and the chromatin

occurs then only in the form of chromidia, there is still

no difficulty in identifying the scattered chromatin-grains

by the fact that they contribute, soon or later, to the

formation of nuclei.

On the other hand, in the simplest Protist organisms

which do not contain definite, compact nuclei recognizable

by their structure and behavior, the identification of the

chromatin may become correspondingly difficult. In the

absence of definite chemical criteria the term chromatin

No. 589] THE EVOLUTION OF THE CELL 21

acquires then a greater or less degree of vagueness and

uncertainty of application, and it is not easy to avoid a

tendency to a petitio principii in attempting to define or

identify it. To a large extent we are thrown back upon

the staining-reactions, which I have already shown to be

very unreliable, backed up by analogies with those forms

which possess definite nuclei. Since in the cells of all

animals and plants, and in all Protista which possess a

true nucleus, the chromatin is the one constituent which

is invariably present, as I shall point out in more detail

subsequently, there is at least a strong presumption,

though not of course amounting to absolute proof, that

it is present, or at least is represented by some similar

and genetically homologous constituents, in the forms of

simpler structure also. If then in Protista of primitive

type we find certain grains which exhibit the character-

istic staining-reactions of chromatin to be constantly

present in the organism, grains which grow and divide as

a preliminary to the organism multiplying by fission and

which are partitioned amongst the daughter-organisms

during the process of fission, so that each daughter-indi-

vidual reproduces the structure of the parent-form from

which it arose; then there is very strong prima facie evi-

dence, to say the least, for regarding such grains as

homologous with the chromatin-grains of ordinary cells.

Having now defined or explained, as well as I am able,

the terms of which I am about to make use, I return to my

main theme, the cell and its evolution. To summarize the

points already discussed, a typical cell is a mass of proto-

plasm differentiated into two principal parts or regions,

the cytoplasm and the nucleus, or, it may be, two or more

nuclei. The cytoplasm may or may not contain chro-

matin-grains in addition to other enclosures, and may

possess cell-organs of various kinds. The nucleus, highly

variable in minute structure, possesses one invariable con-

stituent, the chromatin-material in the form of grains and

masses of various sizes.

The cell, therefore, in its complete and typical form, is

22 THE AMERICAN NATURALIST [ Vou. L

an organism of very considerable complexity of structure

and multiplicity of parts. The truth of this proposition

is sufficiently obvious even from simple inspection of the

structural details revealed by the microscope in cells in

the so-called ‘‘resting condition,’’ but still more so from

a study of their activities and functions. The vital proc-

esses exhibited by the cell indicate a complexity of or-

ganization and a minuteness in the details of its mechan-

ism which transcend our comprehension and baffle the

human imagination, to the same extent as do the immen-

sities of the stellar universe. If such language seems

hyperbolic, it is but necessary to reflect on some of the

established discoveries of cytology, such as the extraor-

dinary degree of complication attained in the process of

division of the nucleus by karyokinesis, or the bewildering

series of events that take place in the nuclei of germ-cells

in the processes of maturation and fertilization. Such

examples of cell-activity give us, as it were, a glimpse

into the workshop of life and teach us that the subtlety

and intricacy of the cell-microcosm can searcely be exag-

gerated.

On the assumption that an organism so complex and

potent was not created suddenly, perfect and complete ~

as it stands, but arose, like all other organisms, by pro-

gressive evolution and elaboration of some simpler form

and type of structure, it is legitimate to inquire which of

the various parts of the cell are the older and more prim- .

itive and which are more recent acquisitions in the course

of evolution. But it must be clearly pointed out, to start -

with, that the problem posed in such an inquiry is per-

fectly distinct from, and independent of, another point

which has often been discussed at length, namely, the

question whether any parts of the cell, and if so which

parts, are to be regarded as ‘‘living”’ or ‘‘active’’ in dis-

tinction to other parts which are to be regarded as ‘‘not-

living” or ‘‘passive.’’? This discussion, in my opinion, —

is a perfectly futile one, of which I intend to steer clear.

e may agree that in any given cell or living organism,

No. 589] THE EVOLUTION OF THE CELL 23

simple or complex in structure, all the parts are equally

‘‘living’’ and equally indispensable for the maintenance

of life, or at least for the continuance of the vital func-

tions in the normal, specific manner, without losing the

right to inquire which of those parts are the phyloge-

netically older. A simple analogy will serve to point my

meaning. A man could not continue to live for long if

deprived either of his brain, his digestive tract, his lungs,

his heart, or his kidneys, and each of these organs is both

‘‘living’’ in itself and at the same time an integral part

of the entire organization of the human body; yet no one

would think of forbidding comparative anatomists to dis-

cuss, from the data at their command, which of these

organs appeared earlier, and which later, in the evolution

of the phylum Vertebrata. Moreover, speculative though

such discussions must necessarily be, there is no one

possessing even a first-year student’s knowledge of the

facts who would controvert the statement that the diges-

tive tract of man is phylogenetically older than the lung.

Speculative conclusions are not always those that carry

the least conviction.

The evolution of the cell may be discussed as a morpho-

logical problem of the same order as that of the phy-

_ logeny of any other class or phylum of living beings, and

by the same methods of inquiry. In the first place there

is the comparative method, whereby different types of

cell-structure can be compared with one another and with

organisms in which the cell-structure is imperfectly de-

veloped, in order to determine what parts are invariable

and essential and what are sporadic in occurrence and

of secondary importance, and if possible to arrange the

various structural types in one or more evolutionary

series. Secondly there is the developmental or ontoge-

netic method, the study of the mode and sequence of the

formation of the parts of the cell as they come into exist-

ence during the life-history of the organism. Both these

methods, which are founded mainly on observation, re-

quire to ms checked and opironaa sa the DEET

24 THE AMERICAN NATURALIST [ Vou. L

methods of investigating both the functions and behavior

of the organism and of its parts.

So long as cytologists limit their studies to the cells

building up the tissues of the higher animals and plants,

the comparative method has a correspondingly limited

scope, and that of the ontogenetic method is even more

restricted. Both methods receive at once, however, an

enormously extended range when the Protista are taken

into consideration. Then, moreover, we see the dawning

possibility of another method of investigation, that,

namely, of the chemical evolution of the organisms. Al-

ready some of the simpler Protista, the Bacteria, are

characterized and classified largely by their chemical

activities; but in more complex organisms, in those which

have attained complete cell-structure, such as Protozoa,

the data of chemistry do not as yet supply the evolutionist

with a helpful method of investigation.

The problem of cell-evolution may be attacked by the

help of the methods outlined in the foregoing remarks,

beginning with the consideration of the primary struc-

tural differentiation of the typical cell, the distinction of

nucleus, or rather chromatin, and cytoplasm. Since all

cells known to us exhibit this differentiation, we have three

possibilities as regards the manner in which it has come

about, which may be summarized briefly as follows: either

the cytoplasmic and chromatinic constituents of the cell

have arisen as differentiations of some primitive sub-

stance, which was neither the one nor the other; or one

of these two substances is a derivative of the other, in the

course of evolution, either cytoplasm of chromatin, or

chromatin of cytoplasm.

The idea of a primitive, undifferentiated protoplasmic

substance was first put forward by Haeckel, who em-

ployed for it the term ‘‘plasson’’ invented by Van Bene-

den’® to denote ‘‘la substance constitutive du corps des

Monéres et des cytodes . . . le substance formative par

10 Bull. de 1’ Acad. Roy. de Belgique, Second Series, Vol. XXXI (1871), p.

346.

No. 589] THE EVOLUTION OF THE CELL 25

excellence.” The simplest elementary organisms were

not cells, but cytodes, ‘‘living independent beings which

consist entirely of a particle of plasson; their quite homo-

geneous or uniform body consists of an albuminous sub-

stance which is not yet differentiated into karyoplasm

and cytoplasm, but possesses the properties of both com-

bined.’’* It is emphasized'* that a sharp distinction

must be drawn between protoplasm and plasson, the lat-

ter being a homogeneous albuminous formative substance

(‘‘Bildungsstoff’’) corresponding to the ‘‘ Urschleim’’ of

the older nature-philosophy.

Haeckel, as was usual with him, did not content him-

self with putting forward his ideas as abstract specula-

tions, but sought to provide them with a concrete and

objective foundation by professing to have discovered,

and describing in detail, living and existing organisms

which were stated to remain permanently in the condi-

tion of cytodes. In consequence, a purely speculative

notion was permitted to masquerade for many years

under the false appearance of an objective phenomenon

of nature, until the error was discovered gradually and

the phantom banished from the accepted and established

data of biology. Organisms supposed to be of the nature

of cytodes constituted Haeckel’s systematic division,

Monera, of which there were supposed to be two sub-

divisions, the Phyt andthe Zoomonera. The Phy-

tomonera were stated to have the plasson colored green

and to live in a plant-like manner; the Zoomonera were

colorless amceboid masses of plasson which nourished

themselves in the animal manner. The Bacteria were

also included by Haeckel in his Monera, apparently, or

at all events ranked as cytodes.’* Most importance, how-

ever, was attributed by Haeckel to the large amceboid

forms of Monera, described as without nuclei or con-

tractile vacuoles, but as representing simply structureless

11 Anthropogenie, sixth edition, Leipzig, 1910, p. 119.

12 Ibid., p. 532.

13 Ibid., p. 119.

26 THE AMERICAN NATURALIST [ Vou. L

contractile masses of albumin (‘‘Hiweiss’’), perfectly

homogeneous;!* examples of these were announced to

exist under the names ‘‘Protameba’’ and ‘‘Protogenes,”’

denoting forms of life which Haeckel claimed to have

discovered, but which have never been found again by

any other naturalist. These organisms, as described by

Haeckel, were by no means such as the modern micro-

scopist would call minute; on the contrary, they were

relatively large, and some of the forms added to the

Monera by Haeckel’s contemporaries might even be

termed gigantic, as, for example, the supposed organism

Bathybius, discovered in the bottles of the Challenger

Expedition, which was believed to cover large areas of

the floor of the ocean with a layer of primordial proto-

plasm, but which proved finally to be a precipitation by

alcohol of the gypsum in sea-water.

The theory of plasson and of the cytodes of Haeckel

may be considered first from the purely speculative stand-

point of the origin of the living substance, a problem with

which I wish to become entangled here as little as pos-

sible, since it is my object to confine myself so far as

possible to deductions and conclusions that may be drawn

from known facts and concrete data of observation and

experiment. If, however, we postulate a chemical evolu-

tion of protoplasm, and believe that every degree of com-

plexity exists, or at least has existed, between the simplest

inorganic compounds and the immensely complicated

protein-molecules of which the living substance is com-

posed, then no doubt chemical compounds may have ex-

isted which in some sense were intermediate in their

properties between the two constituents, cytoplasm and

chromatin, found in all known samples of the living sub-

stance of organisms. In this sense and on such a hypoth-

esis, a substance of the nature of plasson may perhaps be

recognized or postulated at some future time by the bio-

chemist, but this is a subject which I am quite incompe-

tent to discuss. To the modern biologist, who can deal

14 See his ‘‘Prinzipien der generellen Morphologie,’’ Berlin, 1906, p. 61.

No. 589] THE EVOLUTION OF THE CELL 27

only with living things as he knows them, Haeckel’s

plasson must rank as a pure figment of the imagination,

altogether outside the range of practical and objective

biology at the present time. All visible living things

known and studied up to the present consist of proto-

plasm, that is to say, of an extremely heterogeneous sub-

stance of complex structure, and no living organism has

been discovered as yet which consists of homogeneous

structureless albuminous substance. Van Beneden, who

is responsible for the word plasson, though not for the

cytode-theory, was under the impression that he had ob-

served a non-nucleated homogeneous cytode-stage in the

development of the gregarine of the lobster, Gregarina

(Porospora) gigantea. Without entering into a detailed

criticism of Van Beneden’s observations upon this form,

it is sufficient to state that the development of gregarines

is now well known in all its details, and that in all phases

of their life-cycle these organisms show the complete

cell-structure, and are composed of nucleus and cyto-

plasm. Moreover, all those organisms referred by Haeckel

to the group Monera which have been recognized and

examined by later investigators have been found to con-

sist of ordinary cytoplasm containing nuclei or nuclear

substance (chromatin). In the present state of biological

knowledge, therefore, the Monera as defined by Haeckel

must be rejected and struck out of the systematic roll as

a non-existent and fictitious class of organisms.

Since no concrete foundation can be found for the view

that cytoplasm and chromatin have a common origin in

the evolution of living things, we are brought back to the

view that one of them must have preceded the other in

phylogeny. The theories of evolution put forward by

Haeckel and his contemporaries, if we abolish from them

the notion of plasson and substitute for it that of ordi-

nary protoplasm, would seem to favor rather the view

that the earliest forms of life were composed of a sub- _

stance of the nature rather of cytoplasm, and that the i.

nuclear substance or ch om: :

Er mpana ae in evolu- a

28 THE AMERICAN NATURALIST [Von L

tion as a product or derivative of the cytoplasm. I have

myself advocated a view diametrically opposite to this,

and have urged that the chromatin-substance is to be re-

garded as the primitive constituent of the earliest forms

of living organisms, the cytoplasmic substance being a

later structural complication. On this theory the earliest

form of living organism was something very minute,

probably such as would be termed at the present day

‘‘ultra-microscopic.’’ After I had urged this view in the

discussion on the origin of life at the Dundee Meeting of

the British Association in 1912 a poem appeared in

Punch,'* dividing biologists into ‘‘cytoplasmists’’ and

‘chromatinists.’’ I must confess myself still a whole-

hearted chromatinist. But before I consider this point

I may refer briefly to some other speculations that have

been put forward with regard to the nature of the earliest

form of life. It is manifestly quite impossible that I

should undertake here to review exhaustively all the

theories and speculations with regard to the origin of life

and the first stages in its evolution that have been put

forward at different times. I propose to limit myself to

the criticism of certain theories of modern times which,

recognizing the fundamental antithesis between chroma-

tin and cytoplasm, regard these two cell-constituents as

representing types of organisms primitively distinct, and

suggest the hypothesis that true cells have arisen in the

beginning as a process of symbiosis between them.

Boveri, whose merits as a cytologist need no proclama-

tion by me, was the first I believe to put forward such a

notion ; he enunciated the view that the chromosomes were

primitively independent elementary organisms which

live symbiotically with protoplasm, and that the organism

known as the cell arose from a symbiosis between two

kinds of simple organisms, ‘‘Monera.’’!¢

A similar idea lies at the base of the remarkable and

15 Vol. CXLIII, p. 245.

16 Fide Vejdovsky, l. c. I have not had access to the work of Boveri, in

which he is stated to have put forward these ideas

No. 589] THE EVOLUTION OF THE CELL 29

ingenious speculations of Mereschkowsky,'? who assumes

a double origin for living beings from two sorts of proto-

plasm, supposed not only to differ fundamentally in kind

but also to have had origins historically distinct. The

first type of protoplasm he terms mycoplasm,'§ which: is

supposed to have come into existence during what he calls

the third epoch'® of the earth’s history, at a time when

the crust of the earth had cooled sufficiently for water to

be condensed upon it, but when the temperature of the

water was near boiling-point; consequently the waters of

the globe were free from oxygen, while saturated with all

kinds of mineral substances. The second type of proto-

plasm was ameeboplasm, the first origin of which is be-

lieved to have taken place during a fourth terrestrial

epoch when the waters covering the globe were cooled

down below 50° C., and contained dissolved oxygen but

fewer mineral substances. Corresponding with the differ-

ences of the epoch and the conditions under which they

arose, Mereschkowsky’s two types of protoplasm are dis-

tinguished by sharp differences in their nature and con-

stitution.

Mycoplasm, of which typical examples are seen in bac-

teria, in the chromatin-grains of the nucleus and the

chromatophores of plant-cells, is distinguished from

ameeboplasm, which is simply cytoplasm, by the follow-

ing points. (1) Mycoplasm can live without oxygen, and

did so in the beginning at its first appearance when the

temperature of the hydrosphere was too high for it to

have contained dissolved oxygen; only at a later period,

when the temperature became low enough for the water

to contain oxygen in solution, did some of the forms begin

17 Mereschowsky, C., ‘‘ Theorie der zwei Plasmaarten als Grundlage der

Symbiogenesis, einer neuen Lehre von der Entstehung der Organismen,’’

Biol. Centralblatt, XXX, 1910, pp. 278-303, 321-347, 353-367.

18 The term mycoplasm used by Mereschowsky must not be confounded

with the similar word used by Eriksson and other botanists in reference to

the manner in which Rust-Fungi permeate their hosts.

19 In the first epoch the earth was an incandescent mass of vapor; in the

me ee ee ae

of Vater por up

30 THE AMERICAN NATURALIST [Vou L

to adapt themselves to these conditions, and became sec-

ondarily facultative or obligate aerobes. Ameeboplasm,

on the other hand, can not exist without a supply of oxy-

gen. (2) Mycoplasm can support temperatures of 90° C.

or even higher; amceeboplasm can not support a tempera-

ture higher than 45° C. or 50° C. (3) Mycoplasm is

capable of building up albumins and complex organic

substances from inorganic materials; amceboplasm is in-

capable of doing so, but requires organic food. (4) My-

coplasm has restricted powers of locomotion and is in-

capable of amceboid movement, or of forming the con-

iractile vacuoles seen commonly in ameboplasm. (5)

Mycoplasm, in contrast to ameboplasm, is rich in phos-

phorus and nuclein. (6) Mycoplasm is extraordinarily

resistant to poisons and utilizes as food many substances

that are extremely deadly to ameboplasm, such as prussic

acid, strychnine and morphia. (7) Amongst minor dif-

ferences, mycoplasm is characterized by the presence of

iron in the combined state and possesses a far more com-

plicated structure than ameboplasm, a peculiarity which

enables mycoplasmic cell-elements (chromosomes) to

function as the bearers of hereditary qualities.

The course of the evolution of living beings, according

to Mereschkowsky, was as follows. The earliest forms of

life were ‘‘Biococci,’’ minute ultra-microscopic particles

of mycoplasm, without organization, capable of existing

at temperatures near boiling-point and in the absence of

oxygen, possessing the power of building up proteins and

carbohydrates from inorganic materials, and very resist-

ant to strong mineral salts and acids and to various

poisons. From the Biococci arose in the first place the

Bacteria, which for a time were the only living inhabitants

of the earth. Later, when the temperature of the ter-

restrial waters had been lowered below 50° C., and con-

tained abundant organic food in the shape of Bacteria,

-amceboplasm made its appearance in small masses as non-

nucleated Monera which crept in an ameeboid manner on

the floor of the ocean and devoured Bacteria.

No. 589] THE EVOLUTION OF THE CELL 31

The next step in evolution is supposed to have been

that, in some cases, micrococci ingested by the Monera

resisted digestion by them and were enabled to maintain a

symbiotic existence in the ameeboplasm. At first the sym-

biotic micrococci were scattered in the Moneran body, but

later they became concentrated at one spot, surrounded

by a membrane, and gave rise to the cell-nucleus. In this

way, by a ‘‘symbiogenesis’’ or process of symbiosis be-

tween two distinct types of organisms, Mereschkowsky

believes the nucleated cell to have arisen, an immense step

forward in evolution, since the locomotor powers of the

simple and delicate Monera were now supplemented by

the great capability possessed by the Bacteria of produc-

ing ferments of the most varied kinds.

Meanwhile it is supposed that the free Bacteria con-

tinued their natural evolution and gave rise to the Cyano-

phycex, and to the whole group of Fungi. The plant-cell

came into existence by a further process of symbiogenesis,

in that some of the Cyanophycex, red, brown or green in

color, became symbiotic in nucleated cells, for the most

part flagellates, in which they established themselves as

the chromatophores or chlorophyll-corpuscles. In this

way Mereschkowsky believes the vegetable cell to have

come into existence, and the evolution of the Vegetable

Kingdom to have been started, as a double process of sym-

biosis. Those ameboid or flagellated organisms, on the

other hand, which formed no symbiosis with Cyanophy-

cer, continued to live as animals and started the evolu-

tion of the Animal Kingdom.

As a logical deduction from this theory of the evolution

of living beings, Mereschkowsky classifies organisms gen-

erally into three groups or Kingdoms: first the Mycoidea,

comprising Bacteria, Cyanophyceæ, and Fungi, and in

which no symbiosis has taken place; secondly, the Animal

Kingdom, in which true cells have arisen by a simple sym-

biosis of mycoplasm (chromatin) and ameboplasm (cyto-

plasm); thirdly, the Vegetable Kingdom, in which true

32 THE AMERICAN NATURALIST [Vou L

cells have entered upon an additional symbiosis with Cy-

anophyceæ, chromatophores or chlorophyll-corpuscles.

Interesting and suggestive as are the speculations of

Mereschkowsky, they are nevertheless open to criticism

from many points of view. I will not enter here into

criticisms which I regard as beyond my competence. It

is for botanists to pronounce upon the notion that Bac-

teria, Cyanophycee, and Fungi can be classified together

as a group distinct from all other living beings; to decide

whether the protoplasm of the Cyanophycee and Fungi

ean be regarded as consisting of mycoplasm alone, and not

of a combination of nuclei and cytoplasm, such as is found

in true cells and represents, according to Mereschkowsky,

a symbiosis of mycoplasm and ameboplasm. I think I

am right in saying that botanists are agreed in regarding

Fungi as derived from green alge, and as possessing

nuclei similar to those of the higher plants. As a zoolo-

gist the point that strikes me most is the absence of any

evidence that true Monera, organisms consisting of cyto-

plasm alone, exist or could ever have existed. Meresch-

kowsky supposes that when the Monera came into being

they maintained their existence by feeding upon Bacteria.

In order to digest Bacteria, however, the Monera must

have been capable of producing ferments, and therefore

did not acquire this power only as the result of symbiosis

with Bacteria, unless it be assumed that the symbiosis

came about at the instant that amæœboplasm came into

existence. There is, however, no evidence that cytoplasm

by itself can generate ferments. All physiological experi-

ments upon the digestion of Protozoa indicate that the

cytoplasmic body, deprived of the nucleus, can not initiate

the digestive process. Consequently the existence of

purely cytoplasmic organisms would seem to be an im-

possibility.

For my part, I am unable to accept any theory of the

evolution of the earliest forms of living beings which as-

sumes the existence of forms of life composed entirely of |

cytoplasm without chromatin. All the results of modern

No. 589] THE EVOLUTION OF THE CELL 33

investigations into the structure, physiology and behavior

of cells on the one hand, and of the various types of or-

ganisms grouped under the Protista, on the other hand—

the combined results, that is to say, of cytology and pro-

tistology—appear to me to indicate that the chromatin-

elements represent the primary and original living units

or individuals, and that the cytoplasm represents a sec-

ondary product. I will summarize briefly the grounds

that have led me to this conviction, and will attempt to

justify the faith that I hold; but first I wish to discuss

briefly certain preliminary considerations which seem to

me of great importance in this connection.

It is common amongst biologists to speak of ‘‘living

substance,’’ this phrase being preceded by either the defi-

nite or the indefinite article—by either ‘‘the’’ or ‘‘a.’’

If we pause to consider the meaning of the phrase, it is

to be presumed that those who make use of it employ the

term ‘‘substance’’ in the usual sense to denote a form of

matter to which some specific chemical significance can be

attached, which could conceivably be defined more or less

strictly by a chemist, perhaps even reduced to a chemical

formula of some type. But the addition of the adjective

‘‘living’’ negatives any such interpretation of the term

‘‘substance,’’ since it is the fundamental and essential

property of any living being that the material of which it is

composed is in a state of continual molecular change and

that its component substance or substances are inconstant

in molecular constitution from moment to moment. When

the body of a living organism has passed into a state of

fixity of substance, it has ceased, temporarily or perma-

nently, to behave as a living body; its fires are banked or

extinguished. The phrase ‘‘living substance’’ savors,

therefore, of a contradictio in adjecto; if it is ‘‘living’’

it can not be a ‘‘substance,’’ and if it is a ‘‘substance”’ it

can not be ‘‘living.’’

As a matter of fact, the biologist, when dealing with

purely biological problems, knows nothing of a living sub-

stance or substances; he is confronted solely by living in-

34 THE AMERICAN NATURALIST [Vou L

dividuals, which constitute his primary conceptions, and

the terms “‘life’’ and ‘‘living substance’’ are pure ab-

‘stractions. Every living being presents itself to us as a

sharply-limited individual, distinct from other individ-

uals and constituting what may be termed briefly a micro-

cosmic unit, inasmuch as it is a unity which is far from

being uniform in substance or homogeneous in composi-

tion, but which, on the contrary, is characterized by being

made up of an almost infinite multiplicity of heterogene-

ous and mutually interacting parts. We recognize fur-

ther that these living individuals possess invariably spe-

cific characteristics; two given living individuals may be

so much alike that we regard them as of the same kind or

‘*species,’’ or they may differ so sharply that we are

forced to distinguish between them specifically. Living

beings are as much characterized by this peculiarity of

specific individuality as by any other property or faculty

which can be stated to be an attribute of life in general,

and this is true equally of the simplest or the most com-

plex organisms; at least we know of no form of life, how-

ever simple or minute, in which the combined features of

individuality and specificity are not exhibited to the fullest

extent. A living organism may be so minute as to elude

direct detection entirely by our senses, even when aided

by all the resources furnished by modern science; such an

organism will, nevertheless, exhibit specific properties or

activities of an unmistakable kind, betraying its presence

thereby with the utmost certainty. The organisms caus-

ing certain diseases, for example, are ultra-microscopic,

that is to say, they have not been made visible as yet, and

an exact description or definition can not be given of them

at the present time; yet how strongly marked and easily

distinguishable are the specific effects produced by the

organisms causing, respectively, measles and small-pox,

for instance, each, moreover, remaining strictly true and

constant to its specific type of activity; the organism,

whatever its nature may be, which causes measles can not

No. 589] THE EVOLUTION OF THE CELL 35

give rise to small-pox, nor vice versa, but each breeds as

true to type as do lions and leopards.

The essential and distinctive characteristic of a living

body of any kind whatsoever is that it exhibits while it

lives permanence and continuity of individuality or per-

sonality, as manifested in specific behavior, combined with

incessant change and lability of substance; and further,

‘hat in reproducing its kind, it transmits its specific char-

acteristics, with, however, that tendency to variability

which permits of progressive adaptation and gradual evo-

lutionary change. It is the distinctively vital property of

specific individuality combined with ‘‘stuff-change’’ (if I

may be allowed to paraphrase a Teutonic idiom) which

marks the dividing line between biochemistry and biology.

The former science deals with substances which can be

separated from living bodies, and for the chemist specific

properties are associated with fixity of substance; but the

material with which the biologist is occupied consists of

innumerable living unit-individuals exhibiting specific

characteristics without fixity of substance. There is no

reason to suppose that the properties of a given chemical

substance vary in the slightest degree in space or time;

but variability and adaptability are characteristic features

of all living beings. The biochemist renders inestimable

services in elucidating the physico-chemical mechanisms

of living organisms; but the problem of individuality and

specific behavior, as manifested by living things, is beyond

the scope of his science, at least at present. Such prob-

lems are essentially of distinctively vital nature and their

treatment can not be brought satisfactorily into relation at

the present time with the physico-chemical interactions of

the substances composing the living body. It may be

that this is but a temporary limitation of human knowledge

prevailing in a certain historical epoch, and that in the

future the chemist will be able to correlate the individual-

ity of living beings with their chemico-physical proper-

ties, and so explain to us how living beings first came into

existence; how, that is to say, a combination of chemical :

36 THE AMERICAN NATURALIST [Vou. L

substances, each owing its characteristic properties to a

definite molecular composition, can produce a living indi-

vidual in which specific peculiarities are associated with

matter in a state of flux. But it is altogether outside the

scope and aim of this address to discuss whether the boun-

dary between biochemistry and biology can be bridged

over, and if so, in what way. I merely wish to emphasize

strongly that if a biologist wishes to deal with a purely

biological problem, such as evolution or heredity, for

example, in a concrete and objective manner, he must do

so in terms of living specific individual units. It is for

that reason that I shall speak, not of the chromatin-sub-

stance, but of chromatinic elements, particles or units, and

I hope that I shall make clear the importance of this dis-

tinction.

To return now to our chromatin; I regard the chroma-

tinic elements as being those constituents which are of

primary importance in the life and evolution of living or-

ganisms mainly for the following reasons: the experi-

mental evidence of the preponderating physiological role

played by the nucleus in the life of the cell; the extraordi-

nary individualization of the chromatin particles seen uni-

versally in living organisms, and manifested to a degree

which raises the chromatinic units to the rank of living

individuals exhibiting specific behavior, rather than that

of mere substances responsible for certain chemico-phys-

ical reactions in the life of the organism; and last, but by

no means least, the permanence and, if I may use the term,

the immortality of the chromatinic particles in the life-

cycle of organisms generally. I will now deal with these

points in order; my arguments relate, in the first instance,

to those organisms in which the presence of true cell-nuclei

renders the identification of the chromatin-elements cer-

tain, as pointed out above, but if the arguments are valid

in such cases they are almost certainly applicable also to

those simpler types of organisms in which the identifica-

tion of chromatin rests on a less secure foundation.

The results obtained by physiological experiments with

No. 589] THE EVOLUTION OF THE CELL 37

regard to the functions of the nuclear and cytoplasmic con-

stituents of the cell are now well known and are cited in

all the text books. It is not necessary, therefore, that I

should discuss them in detail. I content myself with

quoting a competent and impartial summary of the results

obtained :

A fragment of a cell deprived of its nucleus may live for a consid-

erable time and manifest the power of coordinated movements without

perceptible impairment. Such a mass of protoplasm is, however, devoid

of the powers of assimilation, growth, and repair, and sooner or later

dies. In other words, those functions that involve destructive metabolism

may continue for a time in the absence of the nucleus; those that involve

constructive metabolism cease with its removal. There is, therefore,

strong reason to believe that the nucleus plays an essential part in the

constructive metabolism of the cell, and through this is especially con-

cerned with the formative processes involved in growth and develop-

ment. For these and many other reasons . . . the nucleus is generally

regarded as a controlling centre of cell-activity, and hence a primary

factor in growth, development, and the transmission of specific qualities

from cell to cell, and so from one generation to another.”°

I may add here that the results of the study of life-cycles

of Protozoa are entirely in harmony with this conception

of the relative importance of nuclear—that is chromatinic

—and cytoplasmic cell-constituents, since it is not infre-

quent that in certain phases of the life-cycle, especially in

the microgamete-stages, the cytoplasm is reduced, appar-

ently, to the vanishing point, and the body consists solely

of chromatin, so far as can be made out. In not one

single instance, however, has it been found as yet that any

normal stage in the developmental cycle of organisms con-

sists solely of cytoplasm without any particles of chro-

matin.

While on the subject of physiological experiment, there

is one point to which I may refer. Experiments so far

have been carried on with Protozoa possessing definite

nuclei. It is very desirable that similar experiments

should be conducted with forms possessing chromidia in

addition to nuclei, in order to test the physiological capa-

20 E, B. Wilson, ‘‘The Cell,’’ second edition, 1911, pp. 30 and 31.

38 THE AMERICAN NATURALIST [Vou. L

bilities of chromatin-particles not concentrated or organ-

ized. Arcella would appear to be a very suitable form for

such investigations. This is a point to which my atten-

tion was drawn by my late friend Mr. C. H. Martin, who

has lost his life in his country’s service.

I have mentioned already in my introductory remarks

that the only reliable test of chromatin is its behavior,

and the whole of modern cytological investigation bears

witness to the fact that the chromatinic particles exhibit

the characteristic property of living things generally,

namely, individualization combined with specific behavior.

In every cell-generation in the bodies of ordinary animals

and plants the chromatin-elements make their appearance

in the form of a group of chromosomes, not only constant

in number for each species, but often exhibiting such defi-

nite characteristics of size and form, that particular, indi-

vidual chromosomes can be recognized and identified in

each group throughout the whole life-cycle. Each chromo-

some is to be regarded as an aggregate composed of a

series of minute chromatinic granules or chromioles, a

point which I shall discuss further presently. Most stri-

king examples of the individualization of chromosomes

have been made known recently by Dobell and Jameson?!

in Protozoa. Thus in the Coccidian genus Aggregata six

chromosomes appear at every cell-generation, each differ-

ing constantly in length if in the extended form, or in bulk

if in the contracted form, so that each of the six chromo-

somes can be recognized and denoted by one of the letters

a to f at each appearance, a being the longest and f the

shortest.

(To be continued.)

21 Proc, Roy. Soc. (B), Vol. 89. (In the press.)

THE EUGSTER GYNANDROMORPH BEES

PROFESSOR T. H. MORGAN

CoLUMBIA UNIVERSITY

Azsour fifty years ago von Siebold wrote his classic

paper on ‘‘Zwitterbienen’’ in which he gave an account of

anomalous bees that appeared in considerable numbers in

a hive of a bee breeder, named Eugster, in Constance.*

The particular interest that attached to the case was not

only that a bee might be partly male and partly female,

mixed in all manner of proportions, but that they were

hybrid bees as well, the mother belonging to the race of

Italian bees, while the father or fathers were German

bees. Von Siebold did not state in his paper whether the

male parts of the gynandromorph were like the father, or

were hybrid, or were like the mother. In fact it was not

until 1888 that the importance of such information was

realized. In that year Boveri described a result that he

had obtained with the eggs of the sea urchin, in which as

a result of delayed fertilization (or of some irregularity

in the penetration of the sperm into the egg) the sperm

nucleus fused ‘with one of the two nuclei resulting from

the division of the egg nucleus. In consequence half of

the nuclei were derived from the egg alone, while the other

half of the nuclei arose from the union of the paternal and

a maternal nucleus. If now, as other evidence seemed to

show, one nucleus in the bee produces a male and two

nuclei a female, such a partially fertilized egg should be

male on one side and female on the other side of the body

of the resulting individual. In this way, Boveri pointed

out, the Eugster gynandromorphs might have arisen.

In 1905 I pointed out that the Eugster gynandromorphs

might also be accounted for by means of another hypoth-

esis. If two (or more) spermatozoa should enter the egg,

one of them might unite with the egg nucleus while the

1 Several earlier accounts of gynandromorph bees are extant (See **lit-

erature’’ list). :

40 THE AMERICAN NATURALIST [Vor. L

other might give rise to the nuclei of the rest of the em-

bryo. On this hypothesis the combined nuclei would give

rise to the female parts, while the single nucleus, here

derived from the sperm, would give rise to the male parts.

In support of such a view I pointed out that more than a

single nucleus was known to enter the egg of the bee, and

this condition has more recently been amply confirmed by

Nachtsheim. I also pointed out, for the first time I þe-

lieve, that a decision in favor of one or the other of these

two hypotheses could be obtained if in these gynandro-

morph hybrids the nature of the male and of the female

parts of the adult were known; for on Boveri’s interpre-

tation the male parts (derived from the single egg nu-

cleus) should be maternal while on my view the male parts

(derived from the single sperm nucleus) should be pa-

ternal. In both views the female parts of the gynandro-

morphs should be hybrid and therefore either intermediate

in character or like the dominant strain.

Four years ago Professor Doflein looked through the

collection at Munich, at the request of Boveri, to find out

whether any of the Eugster bees were still preserved

there, and luckily found a jar labelled ‘‘Apis Mellifica,

Zwitterbienen’’ which turned out to be the bees that von

Siebold had obtained. Owing to their long sojourn in

alcohol the color was almost entirely gone and on the color

depended the decision as to the difference between the two

races that combined to produce the gynandromorphs. At

first Boveri despaired of finding out from these alcoholic

specimens whether the male parts were like the father or

like the mother; but on cleaning the parts he found that

he could still determine whether a part was more like the

same part in one or in the other domesticated strain.

Briefly Boveri finds that the male parts of the gynandro-

morphs are maternal, while the female parts are paternal,

which is the dominant character. This conclusion gives a

decisive answer in favor of his hypothesis and sets my

own aside for this case at least.

Boveri’s evidence leaves no reasonable doubt as to the

possibility of determining the nature of the character of

No.589] THE EUGSTER GYNANDROMORPH BEES 41

the gynandromorphs, yet the desirability of having it con-

firmed on living material may be still worth while, since,

as Boveri points out in a postscript, von Engelhardt has

recently (1914) described some hybrid gynandromorphs

from fresh material which lead to the opposite conclusion

from that to which Boveri has arrived. Von Engelhardt’s

bees arose from an Italian queen by a ‘‘domestic’’ drone.

Until it is ascertained what variety was used as the do-

mestic drone the value of the evidence is not entirely

certain.

A student of Boveri’s, Fr. Elsa Mehling, has made a

very careful study of the Eugster gynandromorphs, pay-

ing attention to a number of characters. Her work adds

many details of interest concerning the admixture of male

and female parts, but does not, however, furnish much

additional evidence concerning the origin of these parts.

She arrives at the same conclusion as that reached by

Boveri, viz., that the male parts are maternal.

In this connection it should be recalled that the long

sought for evidence demonstrating that drones inherit the

characters of their mother has at last been found by

Newell. Working at an isolated station forty miles from

Houston, Texas, he mated Italian and Carniolan races of

bees. The Italians are distinctly yellow, while the Carni-

olans are more or less gray.: The stocks used had been

under observation for several generations and were known

to be pure. When virgin Italian queens were mated to

Carniolan drones the workers and queens (both of which

come from fertilized eggs) are like the Italian yellow

stock, which is, therefore, dominant as to color. The

drones from this mating are also yellow, which is expected

if they inherit from their mother, but the cross made this

way is not decisive in regard to the inheritance of the

drones, because the maternal color is here dominant. In

the reciprocal cross the result is decisive. Thus when a

Carniolan queen is mated to an Italian drone the workers

and queens are yellow due to the dominant color of the

father, but the drones are gray like the pure Carniolan

drones. This result proves that the characters of the

42 THE AMERICAN NATURALIST [Vou L

drones come from the mother, which is in accord with

Dzierzon’s theory that the drones arise from unfertilized

eggs. This is further established by the following evi-

dence. The daughters (queens) that come from Italian

queens by Carniolan drones give rise to two kinds of

drones in equal numbers, viz., Italian and Carniolan, which

is the expected result, since such daughters are hybrid and

are expected to produce two kinds of eggs. Reciprocally

also the daughters from Carniolan queens by Italian

drones produce two kinds and only two kinds of drones in

equal numbers. The result also shows that Mendel’s law

applies to the queen bee. Cuénot has recently recorded

the appearance of some drones in hybrid hives that are

intermediate or even like the father, but since the possible

production of drones by hybrid workers was not excluded,

at least so far as the published evidence goes, these spo-

radic cases can not be used to disprove the maternal in-

heritance of the drones.

Boveri has discussed certain cytological possibilities in

relation to the gynandromorph bees that are of interest.

His work, and that of Herbst on sea-urchin embryos, had

shown that haploid nuclei have only half the volume of

diploid nuclei. It might have been anticipated therefore

that the nuclei (and cells) of the drone bee would be half

the size of those of the queen or of the worker bee, but a

study of the cells of drones by Oeninger had already shown

that their nuclei are as large as are those of the workers

which have the diploid number of chromosomes. It is not

possible therefore to determine by microscopic study of

nuclear size whether or not the male parts of gynandro-

morphs come from a single nucleus.

Boveri points out that, since the nucleus of the egg of

the bee, if not fertilized, proceeds to divide, it is improb-

able that the division center is brought in by the sperm, as

appears to be the case in so many other eggs. Nachts-

heim’s observations confirm, he believes, this interpreta-

tion in the bee; for, according to Nachtsheim, three to

seven or more nuclei enter but only one of these fuses

with the egg nucleus. The others move out into the egg,

No.589] THE EUGSTER GYNANDROMORPH BEES 43

their chromosomes are resolved, and a spindle develops.

But these spindles lack centrioles at their poles. The mi-

totic figure that has reached this stage then proceeds to

degenerate. The absence of the centrioles indicates,

Boveri thinks, that the spermatozoa of the bee does not

bring in a division center, hence this cell organ must be

contributed by the egg, and in consequence we can now

easily understand how facultative parthenogenesis is, So

to speak, a normal phenomenon in this egg. Boveri does

not point out however that Nachtsheim’s figures show that

the polar spindles of the bee’s egg also lack centrioles, and

yet mitotic division is accomplished. It seems highly

questionable therefore whether much weight is to be at-

tached to the absence of centrioles in the supernumerary

sperm figures. The chief interest that attaches to Bo-

veri’s argument is his disclaimer that he intended his

striking statement in regard to fertilization, namely, that

the sperm furnishes the dynamic division center for de-

velopment, to be taken as a universal dictum. The incite-

ment of artificial division centers in such eggs as those of

the sea urchin in which the sperm brings in the centriole

(or causes its development in the immediate vicinity of the

sperm nucleus) shows how little importance can be at-

tached to the hypothesis of the genetic continuity of the

centrosome. If in the case of the bee three or more

sperm enter each egg all bees would be gynandromorphs

should all the sperm develop. Obviously, some special

condition must be assumed to be present if these sperms

are to go forward and complete their development which

they begin even under ordinary circumstances. Boveri

himself must also invoke some special condition, such as

retarded fertilization, in order that one of the entering

sperm fuses with one of the products of the first division

of the egg nucleus. It might equally well be postulated

that delay in the fertilization and the consequent impetus

to parthenogenesis might be favorable for the completion

of the division of the supernumerary asters. Ina word

it is doubtful if Boveri’s interpretation gains much from

his cytological argument. If his observations on the dis- .

44 THE AMERICAN NATURALIST [Vou L

tribution of color are well established this further argu-

ment is superfluous.

In 1906 Toyama described a gynandromorph that arose

when two races of silkworm moths were crossed. From

an analysis of the genetic evidence I pointed out that in

this case the male parts of the gynandromorph must have

been paternal and the hybrid parts maternal (dominant).

If the same conditions prevail here as in the bee, viz., one

nucleus producing a male and two producing a female? the

case is in harmony with my hypothesis and not with that

of Boveri. But the evidence for my view is not as strong

as that Boveri’s is now for the bee; yet it may be true,

nevertheless, that in both of these ways gynandromorphs

may arise. A third mode of origin has been shown, from

the genetic evidence, to apply to Drosophila, viz., disloca-

tion during ontogeny of the two sex chromosomes. In fact

we should expect that gynandromorphs would arise in in-

sects whenever certain nuclei come to contain two sex

chromosomes and others only one. The means by which

this segregation takes place may differ under different

conditions.

Goldschmidt has recently explained the remarkable

gynandromorphs that he obtains in crosses between Ly-

mantria dispar and L. japonica in still a different way,

one that involves the relative potencies of the sex factors

in the different races.

LITERATURE CITED

Boveri, Th. 1 Uber partielle Befruchtung. Sitz.-Ber. d. Ges. f.

Morph. u. Phys., Münch., IV.,

Boveri, Th. 1915. Über die Entstehung der Eugsterschen Zwitterbienen.

Arch. f. Entw. Organismen, XLI.

Cuénot, L. 1909. Comp. Rend. Soc. Biol., LXVI.

Dönhoff. 1860a. Ein Bienenzwitter. Bienenzeitung.

2 Whether one is justified in applying to the case of the moth the

hypothesis for the bee may be seriously questioned because in the case of

the moth the male is assumed to be the result of one sex chromosome (z) in

conjunction with the haploid number of autosomes, while in the female

moth one sex chromosome (2) and its mate (w) (which from Doncaster’s

evidence has no sex-determining influence) in conjunction with the diploid

number of autosomes is assumed to stand for the female soma.

No.589] THE EUGSTER GYNANDROMORPH BEES ` 46

Dönhoff. 1860b. Beitr. z. Bienenkunde. I. Uber Zwitterbienen, Bienen-

zeitung.

Goldschmidt, B.: 1912. igen melee an Schmetterlingen, I. Zeit. f.

ind. Abst

rb., i

anet, H. 1861. AEREE Revue et Magas. de Zool., XIII.

Laubender, B. 1801. Einige Bemerkungen über die von Herrn Schul-

m r Lukas neu entdeckten Stacheldrohnen. Ökonom., Hefte XVII.

Lefebure, A. 1835. Description d’un Argus Alexis jauno. . Ann,

soc. entom. France, I F

Menzel, A. 1862a. Abnormitit in der Bildung einer Biene. Bienen-

geitung.

Menzel, A. 1862b. Uber Zwitterbienen. Bienenzeitung.

Menzel, A. 1862c. Hymenopterologische apa I. Über die Ge-

schlechtsverhiltnisse der Biene im allgemeinen und über die Befruch-

tung der a Parthenogenesis dsa Zwitterbildung im beson-

deren. Mitt. 8 entom. Ges., Bd. I, H

Menzel, A. 1864. Toi. Ane Zwittermutter des Eugsterschen Stockes in

Konstanz. Bienenzeitung

Morgan, T. H. 1905. An eiim: interpretation of gynandromorphous

insects. Science, X

Morgan, T. H. 1907. The Cause of Gynandromorphism in Insects. Am.

T. H. 1909. Are the Drone Eggs of the Honey-Bee Fertilized?

XLIII,

Morgan, T. H. 1909. Hybridology and Gynandromorphism. AM. NAT.,

LIH,

Nachtsheim, H. 1913. Cytologische Studien über die Geschlecht tsbestim

mung bei der Honigbiene (Apis mellifica L.). Arch. f. Zellforsch.,

XI.

Newell, W. 1914. Inheritance in the Honey Bee. Science, XLI.

Oehninger, M. 1913. Über Kerngrössen bei Bienen. Verh. d. phys.-med.

Ges. Würzburg, XLII.

Siebold, w Th. 1864 Ueber Zwitterbienen. Zeit. f. wissensch. Zool. XIV.

Siebold, v.C. Th. 1866 Ersatz der abgestorbenen Zwittermutter des Eug-

sterchen Zwitterstockes in Constanz. Bienenzeit. Jahrg. 1866.

Smith, F. 1862. mne Hermaphrodite of Apis mell. from Scotland.

Proc. Entom. Soc. Lon

Smith, F. 1871. Notes on i Tana of Gynandromorphism in Aculeate

Hymenoptera. Trans. Entom. Soc. London.

Toyama. 1906. On Some Silk-worm Crosses, with Special Reference to

Mendel’s Law of Heredity. Bull. of the Coll. of Agr., Tokyo, XII.

Wheeler, W. M. 1903. Some New Gynandromorph Ants. Bull. Am.

Museum Nat. Hist. New York.

Wheeler, W. M. 1910. The Effects of Parasitic and other kinds of Cas-

tration in Insects. Jour. Exp. Zool. VII.

Wittenhagen. 1861. Uber Bienencharakteristik und Bienenzwitter. Bie-

nenzet

SHORTER ARTICLES AND DISCUSSION

PINK-EYED WHITE MICE, CARRYING THE COLOR

AmoneG the many domesticated varieties of the house mouse

(Mus musculus), two sorts with entirely white pelage are known,

—the albino, and the black-eyed white. Numerous experiments

have shown that the albino differs from colored varieties by the

loss of a single factor, the color factor ; for, in crosses with colored

varieties, albinism acts as a recessive allelomorph. The genetic

composition of the black-eyed white is less well known although

several hypotheses have been suggested. Black-eyed whites pos-

sess the color factor as crosses with albinos have shown. They

may be homozygous in the factor for dark eyes. A black-eyed

white male produced 189 dark-eyed offspring in my experiments

when mated to pink-eyed intense brown females. The offspring

of this cross were heterozygous in dark eye (Dd). By mating

them inter se, pink-eyed forms were obtained in the F, genera-

tion, some of which had a pure white coat. In other words, it is

possible to recombine the factors producing the pure white pelage

of the black-eyed whites with the pink-eyed condition. Such pink-

eyed whites resemble true albinos in appearance, but not in zy-

gotic constitution, for they still retain the color factor although

they show no color. To avoid confusion in discussion, I shall

refer to this synthesized form of albino as a pink-eyed white to

distinguish it from the albino lacking the color factor. Predic-

tions often compel subsequent retractions; however, I feel safe in

predicting colored offspring from a cross between the pink-eyed

white and the albino, although externally the mating resembles

a cross between albinos which always breed true. Black-eyed

white strains sometimes show a few colored hairs around the ears,

between the eyes, and in front of the tail. The corresponding

pink-eyed white forms may also show the same characteristic.

The white coat and pink eyes of the albino mouse are due to the

loss of a single factor; but the white coat of the black-eyed

white strains cannot be accounted for in such a simple manner.

Little (713) seemed inclined to the view that the black-eyed white

mouse was a spotted individual in which the spotting was of the

No.589] SHORTER ARTICLES AND DISCUSSION 47

recessive type, in contradistinction to spotting of the dominant

type described by Miss Durham (’08). Through the kindness of

Professor W. E. Castle, a black-eyed white male was received in

the fall of 1914. With this male it-was possible to produce other

black-eyed whites. In such black-eyed whites as I have been

able to test, both dominant and recessive spotting were present.

Furthermore, the recessive spotting always occurred in double

dose. Hence, black-eyed whites were supposed to have the zy-

gotice formula PPss or Ppss, in which P stands for dominant spot-

ting and p for its absence; and s represents the factor for reces-

sive spotting which is allelomorphic to self (S). So far, I have

been able to test sufficiently only two black-eyed white males, botn

of which were clearly of the formula Ppss. When mated to self-

colored females they gave 231 offspring. Since these offspring

showed much variability, they were graded in classes ranging

from self (—9) to black-eyed white (+9) according to the

amount of pigmentation which they showed. A distinct group-

oe around two modes was found as follows:

Chars modes ..| — 9-8 -7—6 —5 —4 —3|—2 ılo ea eae +8/+9

Frequencies. E 24| 8| 16| 33| 26| 8| 4! 6l2| 3| 1 0 0 0

About one-half of the F, offspring was grouped around the lower

mode (126), and the other half (105) grouped around the upper

mode, if we assume the class —7, as the dividing class. Very

few individuals were found in the ‘‘doubtful class.’’ Expressing

the cross of black-eyed white with self-colored in Mendelian

terms, it would

Ppss X ppSS = P, zygotes

Ps+ ps= gametes of black-eyed white P,

pS+ pS= gametes of self-colored P,

PpSs + ppSs =F, zygotes

Spotted + Self

The results conformed to this expectation. The individuals

grouped around the lower mode were self-colored or very nearly

so, as one would expect of individuals heterozygous in self and

recessive spotting, for self is dominant or very nearly dominant

to recessive spotting. Their formula was ppSs. Subsequent

experiments corroborated this, for they produced self and reces-

sive spotted in Mendelian ratios, when mated inter se or to reces-

sive spotted individuals. They never gave black-eyed whites in-

48 THE AMERICAN NATURALIST [Vou L

such matings. Those offspring grouped around the upper mode

were spotted, and had a formula PpSs. When mated inter se or

back to recessive spotted, they gave, besides spotted and selfs,

Rlack-eyed whites; apparently because the combination Ppss

could again be formed. The dominant spotting factor, P, evi-

dently acts more vigorously upon recessive spotting than upon

self. It can not restrict the more extended pigmentation of a

self coat completely. Hence, half of the F, individuals (those

with the formula PpSs) were spotted, or, to describe them more

accurately, spotted with frequent and varying amounts of silver-

ing. The dominant spotting factor, P, can, however, restrict the

limited pigmentation of a recessive spotted coat completely or

almost completely. Hence animals with the formula Ppss were

black-eyed whites.

The origin of our new pink-eyed white forms, which resemble

albinos so closely as to be indistinguishable from them, is evi-

dently due to the substitution of the pink-eye factor for dark-

eye in black-eyed whites, and not due to the loss of the color

factor C. In our cultures, the black-eyed whites have the for-

mula PpssDDCC and the corresponding pink-eyed whites had

the formula PpssddCC where D and d represent dark eye and

pink eye respectively, and C represents the color factor. We

have also produced black-eyed white forms heterozygous in dark

eye, PpssDd. Black and brown are likewise interchangeable in

the dark-eyed whites, for black-eyed whites, heterozygous in

black, have been produced. I see no reason why brown-eyed

whites can not be produced in the usual Mendelian fashion by

mating black-eyed whites to browns, and recovering the white

pelage with brown eyes in the F, generation. Mating the spotted

F, offspring inter se should give, among others, individuals with

the formula PpssbbCC. These would be brown-eyed whites,—

white because of the combined action of P and s, and brown

simply because they lack the differential factor B which changes

brown into black.

The occurrence of pink-eyed whites which resemble albinos may

have some bearing on an anomalous case cited by Bateson (’04)

as follows: ‘‘the production of colored animals by albinos, is not,

so far as I know, illustrated by a single case, with the following

exception. In the later editions of ‘‘Faney Mice’? (Upeott Gill),

Dr. Carter Blake, formerly secretary of the Anthropological In-

stitute commenting on the statement that albino mice of whatever

No.589] SHORTER ARTICLES AND DISCUSSION 49

parentage produce nothing but albinos, writes that a pair of

albinos produced some brown-and-white, some plum, some grey,

and some albinos. If this result occurred under all precautions,

it stands alone.’’ Allen (’04) attempted to account for this case

by postulating an error in recording the true sire, or that the

animals used were not true albinos but black-eyed whites. That

two individuals having white coats and pink eyes can give colored

young is perfectly possible. The pink-eyed whites in my cul-

tures have a white pelage because of the combined effect of the

dominant and recessive spotting, while their pink eye is due to

the loss of the dark-eye factor. They still retain the color factor,

although they show no color. They may be called albinos, if we

define an albino as any pink-eyed white individual; but they

should be carefully distinguished from that type of albinism

which is due to the loss of the color factor. If we mate these two

different types of albinos together, we should obtain colored

young. The cross may be expressed in symbols:

PpssddCC X ppSSDDec ........ P. zygotes

PC pad aein gametes of pink-eyed white

pee +: phe enei gametes of albino

PpSsDdCe + ppSsDdCe........ F, zygotes

Spotted + Selfs

It is interesting to note that the exceptional case, quoted by

Bateson, mentions the occurrence of spotted and selfs in the

cross of two albinos. In plants, as in animals, similar somatic

characters do not necessarily indicate similar germinal con-

stitution.

Our assumption of the interaction of a dominant and recessive

spotting factor to account for the white pelage of pink-eyed and

black-eyed whites is strengthened by the valuable paper of Little

(715). Little has adopted a similar hypothesis for black-eyed

whites in his paper just published, and quite different from the

hypothesis of his earlier paper.(’13). It should be stated that

Little’s experiments furnish even a larger amount of data from

the more convincing type of matings than has been possible in our

own cultures as yet. J. TLEFSEN.

REFERENCES :

Allen, G. M. 1904. Proc, Am. Acad. Arts and Sci., vol. 40, pp. 61-163.

Bateson, W 1903. Proc. Zool. Soe. London, vol. 2, pp. 71-98.

Durham, F. M. 1908. Royal Soc., Rep. Evol. Comm., No. 4, pp. 41-53.

Little, C. C. 1913. Carnegie Inst. Wash., Pub. 179, pp. 11-102. :

Little, C. C. 1915. Am. Nat., vol. 49, pp. "727-740.

50 THE AMERICAN NATURALIST [Vou. L

PARTHENOGENESIS AND SEXUAL REPRODUCTION IN

ROTIFERS. EXPERIMENTAL RESEARCH UPON

BRACHIONUS PALA?

IN a recent number of Bios Miss Lina Moro has presented some

interesting and suggestive results from experiments upon the

rotifer, Brachionus pala. She has subjected the parthenogenetic

females to various chemicals, to changes in nutrition, and to

changes in temperature.

In using FeCl, solutions she has been able to produce male-

producing females in small numbers while in control experiments

in which no FeCl, was used no male-producing females were

produced. Many dilutions of FeCl, were used but M /12,000

seemed to be the optimum dilution. This was added to the cul-

ture water of hay infusion in which the rotifers were living.

Although the number of the experiments were rather small and

the percentage of male-producing females obtained was not

higher than 12 per cent., nevertheless they indicate the possi-

bility of a specific chemical being able to induce the production

of male-producing females.

Not only did FeCl, cause male-producing females to appear

but it also caused the mothers to form the eggs much faster in

their bodies and to extrude them to the outside much faster than

those in the controls. Usually while a female in the control was

producing one egg a female in the FeCl, would produce four

eggs. This rapid formation and production of eggs after it was

once started continued through many subsequent generations

during the three months in which the experiments were carried

on. It might be considered that this new characteristic induced

by a chemical was a case of the formation of a new character

which, after it was once formed, was inherited by the descendants.

It was also determined that the influence of the FeCl, acted

upon the egg while it was yet inside the mother and caused it to

develop into a male-producing female. After the egg was laid its

development could not be altered from a female-producing

female to a male-producing female by the use of FeCl,.

A dilution of HgCl, (M/ 1,200,000,000) was also effective in

causing male-producing females to appear but a smaller number

of offspring were produced than in the FeCl,. The percentage

1‘‘Partenogenesi e Anfigonia nei Rotiferi. Recerche sperimentali sul

Brachionus pala,’’ by Lina Moro, Bios, Vol. 2, Fase. 3, pp. 219-264, 1915.

No.589] SHORTER ARTICLES AND DISCUSSION 51

of male-producing females produced was about 18 per cent. It

also caused an increase in the number and the rate of production

of the eggs by each female as was the case in the FeCl, experi-

ments. KCl (M/12,000) in the very few experiments recorded

caused about 16 per cent. of male-producing females to appear

and CaCl, (M/12,000) caused about 33 per cent. of male-produ-

cing females to appear. In the controls for these experiments no

male-producing females appeared. In all of these chemical ex-

periments.each mother after being transferred from the control

to the culture media containing the various chemicals produced a

family of several daughters but in each family there was never

more than one male-producing daughter. In the FeCl, solutions

each mother produced many daughters and as only one of them

in each family was a male-producer the percentage of male-pro-

ducing females was necessarily lower whereas in the HgCl,, KCl,

CaCl,, solutions each mother produced fewer daughters than in

the FeCl, solutions, and as only one of these in each family was

a male-producer, the percentage of male-producing females was

consequently higher. Various dilutions were used of AICL,,

KCN, NaCl, Na,HAsO,, HCl, and NaOH but none of them caused

male-producing females to appear.

In the nutrition experiments it was found that a constant diet

at a uniform temperature of 15° C.-17° C. or 25° C.-27° C. pro-

duced only female-producing females but in some experiments

in which an abundance of food was used for a time and then was

followed by a period of scanty food or semi-starvation many

male-producing females appeared, especially at the lower

temperature.

In some of the experiments a temperature of 15° C.-17° C. pro-

duced all female-producing females but when the mothers were

put at a temperature of 25° C.—27° C. or at 31° C. as high as 50

per cent. of the daughters were male-producers. When these

same mothers were transferred back to 15° C.-17° C. they again

produced only female-producing daughters. In a few experi-

ments at a constant temperature of 25° C.-27° C. only female-

producing females were produced but when the mothers were put

at a lower temperature they produced many male-producing

daughters. The general conclusion drawn is that whenever the

general cultural conditions are constant and uniform, whether

they refer to nutrition or to temperature, only female-producing

females are produced but when the cultural conditions are sud-

52 THE AMERICAN NATURALIST [ Von. L

denly changed by the disappearance of an abundant diet or by

the rise or fall in the temperature male-producing females are

produced at once. In a few experiments very young females

(1-7 hours after hatching) were put from a high temperature

to a temperature of 9° C.—11° C. and many of them developed °

into male-producers but whether this was due to the tempera-

ture or to some other factor was not known.

Another fact of considerable interest was verified. It con-

cerned the nature of the male-producing females and the sexual

females (the females which produce fertilized eggs). It has

been observed by several investigators that if the small male eggs

of a male-producing female are fertilized, in a species of

Asplancha and Hydatina senta, they develop into the winter or

resting eggs. This was found to be true also in Brachionus

pala.

In all the families of daughters from the various mothers it

was found that the male-producing daughters were among the

earliest ones produced of each family. This was observed in the

families of Hydatina senta by an earlier worker but later it was

found to be due entirely to the method of feeding.

Although, as stated previously, the observations recorded in

this paper are from a rather small number of individuals and

ought to be expanded and verified, nevertheless, they show that

in this rotifer the production of female-producing or male-pro-

ducing females can be regulated by the environment and thus the

results are in a general accord with the observations obtained by

several workers with the rotifers, Asplancha, and Hydatina senta.

D. D. WHITNEY

NOTES AND LITERATURE

AN OUTLINE OF CURRENT PROGRESS IN THE THEORY

OF CORRELATION AND CONTINGENCY

WORKERS in the physical sciences realized long ago that cer-

tain progress depended upon the precision of their instruments

of measurement and the adequacy of their methods of mathe-

matical description and analysis. Biologists, here and there, are

beginning to see the importance of the analytical as well as of the

observational tools. Among the analytical formule none are

of greater usefulness than those for measuring interdependence.

It may not be out of place, therefore, to sketch in simple terms

for the benefit of those who are interested in the methods only as

a means to an end, the progress which is being made in the per-

fection of these instruments of research.

The term current as used in these paragraphs is made more

comprehensive than is conventional; some of the citations are

four or even more years old. The elasticity of the term is justi-

fied in dealing with the literature of a field in which progress is

particularly difficult and in which actual contributions are incor-

porated but slowly into the working technique of the biologist.

Indeed, biologists as a class still think of correlation as synony-

mous with the classical product-moment method. How erroneous

this impression is will appear in the following pages

The purpose of this review is therefore to iadiosto in non-

mathematical terms easily comprehensible to biological readers

the lines of advances in the theory of the measurement of inter-

dependence in order that they may the more easily select for

dealing with their actual data, formulæ of the existence of

which they might otherwise be unaware.

The progress which we have to consider has been along four

different lines:

(a) In the simplification of methods of computation in the

case of familiar formule. (b) In the development of entirely

new formule applicable to data of particular sorts. (c) In the

determination of the corrections to be applied for grouping into

‘‘ broad categories.” (d) In partial correlation, multiple cor-

relation, and the correlation of indices and increments.

54 THE AMERICAN NATURALIST [Vou L

In this review we shall limit ourselves strictly to an outline of

progress which has been made in the theory of the measurement

of the interrelationship of two variates, leaving for considera-

tion at a later time the far more complex subjects of correction

for grouping, partial and multiple correlation, variate differ-

ence correlation and some other topics.

The detailed advances may be most easily understood by con-

sidering the kinds of data with which one has to deal in deter-

mining the degree of interdependence, association or correlation

(to use these terms in a broad sense) between two variates.

An arrangement of the literature according to a key similar to

that familiar in taxonomic works will perhaps be of service to

the biologist who desires to locate at once the literature perti-

nert to the particular kind of data with which he has to deal.

Suppose first of all that the two characters are both suitable

for measurement (or counting) on a quantitative scale and that

for both the measurements form several classes. The choice of

methods for measuring the correlation between them will then

depend upon whether the average values of the y character as-

sociated with serially arranged values of the x character lie in

sensibly a straight line or whether they can best be represented

by some more complex curve. Linearity of regression, as it is

technically called, has therefore a two-fold significance. (a)

Biologically,:it shows that an associated character changes at a

uniform rate (however slight this rate may be) with the varia-

tion of a selected character, (b) Statistically, it justifies the

application of the familiar product-moment method of determin-

ing the correlation coefficient.

Both Characters Measurable on a Quantitative Scale, Regres-

sion Linear—So satisfactory has the product-moment method

proved for data in which both characters are measurable and

regression is sensibly linear, that no fundamental advance has

been made for several years. Boas’s? first formula is, as pointed

out by Pearson,? merely another form of the difference method,

which has been in use for many years.

Several modifications of a purely technical nature which facili-

tate calculation or are useful in special cases have been pub-

1 Boas, F., ‘‘Determination of th

N. S., 29: 823-824, 1909.

2 Pearson, K., ‘‘ Determination

N. S., 30: 23-25, 1909,

e Coefficient of Correlation,’’ Science,

of the Coefficient of Correlation,’’ Science,

No. 589] NOTES AND LITERATURE 55

lished. Pearson® has given a new approximate difference method

which is serviceable in special cases only. Harris* has suggested

a novel difference method for exact work with tables. An alter-

native method of calculating rough moments and product mo-

ments, given by Elderton,® seems to have attracted little atten-

tion, although it has certain advantages for use in adding-machine

computations. A product moment method which possesses

marked advantages for use with machines which allow of simul-

taneous multiplication and summation, and which obtains inci-

dentally the data necessary for testing linearity of regression or

computing the correlation ratio, 7, is now available.® In the

special cases in which the two characters to be centered in the

correlation table are not differentiated, e. g., stature of pairs of

brothers, length of Paramecium, ete., the tables are ordinarily

rendered symmetrical by using each individual once as the « and

once as the y member of the pair. This may be done by actually

forming the symmetrical table, or by using the simple formula

proposed by Jennings.’ If, as is frequently the case, more than

a single pair of individuals are associated, the labor of forming

tables becomes very great. Each individual of a family, each

organ of an individual, or each individual measured from a par-

ticular environment, must then be entered in the table in com-

bination with every other one. Since the number of combina-

tions in each class is n(n—1) and the number of classes must

be at least moderately large, the total number of combinations is

very great. Thus the data for number of nipples in swine

recently published by Parker and Bullard’ require a table of

34,884 combinations to determine the fraternal correlation for

number of nipples. In the case of the Hydra data analyzed by

Lashley,’ tables with from one to nearly two hundred thousand

8 Pearson, K., ‘‘On Further Methods of Determining Correlation,’’

Drapers’ Company Research Mem., Biom. Ser., IV, Dulan and Co., 1907.

4 Harris, J. Arthur, ‘‘A Short Method of Calculating the Coefficient of

Correlation in the Case of Integral Variates,’’ Biometrika, 7: 214-218, 1909.

5Elderton, W. P., ‘‘An Alternative Method of Calculating the Rough

Moments from the Actual Statistics,’’ Biometrika, 4: 374-378, 1905. Also

in his ‘‘ Frequency Curves and Correlation.”’

Harris, J. Arthur, ‘‘The Arithmetic of the Product Moment Method of

Calculating the Coefficient of Correlation,’’ AMER. NAT., 44: 693-699, 1910.

T Jennings, H. S., ‘‘Computing Correlation in Cases Where Symmetrical

Tables are Commonly Used,’’ AMER. NAT., 43: 123-128, 1911.

8 Parker, G. H., and C. Bullard, Proc. Amer. Acad. Arts and Science, 49:

399-426, 1913.

? Lashley, K. S., Jour. Exp. Zool., 19: 210, 1915.

56 THE AMERICAN NATURALIST [Vou L

combinations are given. Methods for the rapid formation of

symmetrical tables from which either correlation or contingency

coefficients may be calculated’? and for the formation of con-

densed tables from which correlation coefficients’! only may be

deduced greatly reduce the necessary labor in such cases. For

the testing of linearity of regression in the case of these intra-

class and inter-class correlations, tables are essential. The use of

such coefficients would, however, be greatly facilitated if calcula-

‘tion could be carried out directly from moments computed from

the classes themselves. Harris’? has given an exhaustive series

of formule by which this can be accomplished, with examples

showing the wide applicability of such coefficients. For example,

these formule fulfil more adequately the purpose of Boas’s

second formula (loc. cit.).

These intra-class correlation formule have been thrown into a

form suitable for measuring substratum heterogeneity in experi-

mental cultures.**

If the x and y character of a pair are differentiated, spurious

values of the correlation coefficient must result from the render-

ing symmetrical of the correlation surface. Pearson many years

ago recognized the difficulty in dealing with groups in which

there is orderly differentiation due, for example, to growth.

Attention has recently been directed” to difficulties arising when

differentiation within the class may exist, but it may be difficult

or impossible to arrange the individuals by any character out-

side of themselves to obtain the constants necessary for deter-

mining the true correlation from the spurious values deduced

10 Harris, J. Arthur, ‘‘On the Formation of Correlation and T Soot

Tables when the Number of Combinations is Large,’’ AMER. NAT., 45:

571, 1911

11 Harris, J. Arthur, ‘‘The Formation of Condensed Correlation Tables

when the Number of Onbicktines is Large,’’ AMER. NAT., 46: 477-486,

1912

12 Haris, J. Arthur, ‘‘On the Calculation of Intra-class and Inter-class

Coefficients of Correlation from Class Moments when the Number of Pos-

noh Combinations is Large,’’ Biometrika, 9: 446—472, 1913.

Harris, J. Arthur, ‘‘On a Criterion of Substratum Homogeneity or

Heterogeneity in Field Experiments,’’ AMER. NAT., 49: 430-454, 1915.

14 Pearson, K., ‘‘On Homotyposis in Homelogoos but Differentiated Or-

ort sai Roy. Soc. Lond., 71: 288-313, 1903.

. Arthur, “On Spurious Values of Intra-Class Correlation

Fastin Arait from Disorderly POA e within the Classes,’’

Biometrika, 10: 412-416, 1914.

No. 589] NOTES AND LITERATURE 57

from the tables. Whether the methods used in such cases by

Harrist: will prove the best available remains to be seen.

Considerable attention has recently been given to the probable

error of the correlation coefficient.

If the number of observations upon which r is based is large

and if it does not approach too closely either of its limiting

values of +1 or — 1, the use of the formula of Pearson and

Filon,

a. oe.

readily evaluated by the use of the tables of 1— r?° given by

Soper?’ used in connection with the z,, of Miss Gibson’s Tables*®

or approximated by the Abac of Heron,” is quite legitimate.

But when either of these conditions is not realized the value of r

found from a single sample will probably not be the true corre-

lation for the population under consideration.

Chemists, agriculturists, physiologists and many others often

must necessarily reason from a relatively small number of obser-

vations. It is therefore of very real importance that some valid

measure of the statistical trustworthiness of such coefficients be

known. Some of the problems concerning the probable error of r

when it approaches its numerical limits or when the number of

cases upon which it is based is small are discussed mathematically

by Soper” as they have been attacked experimentally by ‘‘Stu-

dent.’ Further contributions to the subject are those of Fisher**

and of Pearson,?* who summarizes the series of studies and gives a

table to facilitate the interpretation of correlation coefficients

based on small samples. He says:

16 Harris, J. Arthur, ‘‘On the Significance of Variety Tests,’’ Science,

. 5., 36: 318-320, 1912, and Biometrika, l. c.

17 Soper, H. E., In ‘‘ Tables for Statisticians and Biometricians.”’

18 Biometrika, 4: 385-392, 1906. Also in Pearson’s Tables

1° Heron, D., ‘An Abac for Determining the Probable Errors of Corre-

lation Coefficients, ’? Biometrika, 7: 411, 1910. Also in Pearson’s Tables.

20 Soper, H. E., ‘On the Probable Error of a Correlation Coefficient to a

Second Approximation,’? Biometrika, 9: 91-115, 1913.

*1“*Student,’’ ‘Probable Error of a Correlation Coefficient,’’ Biometrika,

6: 302-310, 1908.

= Fisher, R. A., ‘‘Frequency Distribution of the Values of the Correla-

tion Coefficient in Samples from an Indefinitely Large Population,’’ Bio-

metrika, 10: 507-521, 1915.

?3 Pearson, K., “On the Distribution of Small Samples’’; Appendix I to

Papers by ‘Student’? and R. A. Fisher, Biometrika, 10: 522-529, 1915.

58 THE AMERICAN NATURALIST hotels

We think it must be concluded that for samples of 50 the usual theory

of the probable error of the standard deviation holds satisfactorily, and

that to apply it for the case of n= 25 would not lead to any error which

would be of importance in the majority of statistical problems.

The original papers should be read by those who are dealing

with coefficients lying near the limits of the range of correlation,

or who must work with small samples. Those who can by extra

labor obtain larger series of data should do so, for no knowledge

of the theory of the probable error can ever take the place of

widened series of data, although it may be essential to the inter-

pretation of constants based of necessity on a limited number of

observations,

Both Variates Measurable on a Quantitative Scale ; Regression

Non-Linear.—For cases in which the rate of change in the y char-

acter can not be deseribed by a straight line, the proper measure

of interdependence is Pearson’s** correlation ratio, n. The value

of the correlation ratio is two-fold. (a) It furnishes a measure

of the interdependence of two variates in cases in which the use

of the correlation coefficient is not fully justified. (b) It affords

a means of testing, by the use of Blakeman’s criterion,” for

linearity of regression. Thus in deciding between the correla-

tion coefficient and the correlation ratio, the calculation of each

of the constants may, in critical cases, be necessary.

A further test of the goodness of fit of regression curves has

also been given by Slutsky.2* This method, which involves the

well-known x? of Pearson’s test for goodness of fit, should have

wide usefulness. An illustration of its application has recently

been given by Pearl.?7

One Variate Describable in Multiple Categories, the other

Measurable on a Quantitative Scale.—Such cases are occasionally

met with in many fields of work. For example, one may desire

to know in fractions of a scale ranging from 0 to 1 the relation-

ship between any deseribable but not measurable environmental

24 Pearson, K., ‘‘On the General Theory of Skew Correlation and Non-

— Regression,’’ Drapers’ Co. Res. Mem., Biom. Ser., II, Dulan and Co.,

25 Blakeman, J., ‘‘On Tests for Linearity of Regression in Frequency

Distributions, ’’ Biometrika, 4: 332-350, 1905.

26 Slutsky, E., ‘‘On the Criterion of the G

Lines and on the Best Method

Stat, Soc., 77: 78-84, 1914,

** Pearl, R., ‘An Important Contribution to Statistical Theory,’’? AMER,

oodness of Fit of Regression

of Fitting them to the Data,’’ Jour. Roy.

NAT., 48: 505-507, 1914

De r eee © cow hn ena, Tare SOE Si

No. 589] NOTES AND LITERATURE 59

factor and any measurable characteristic of the organisms sub-

jected to its influence. Or in testing the assertions of such wri-

ters on criminology as Lombroso and Havelock Ellis against the

results of actual measurements of criminals, one may find it

desirable to correlate between the kind of crime and any cephalic

measurement.

For the analysis of such data the correlation ratio may be of

great service.

One Character Alternative, the other Measurable on a Quanti-

tative Scale.—Suppose now that one of the correlation ratio tables

of the kind discussed in the foregoing paragraph were reduced,

as far as the qualitatively appreciable but not measurable char-

acter is concerned, to two classes only, while the measured variate

remained as before. Such tables actually oceur in practise with

great frequency. For example, one may wish to correlate be-

tween the form of a dimorphic crustacean and physical measure-

ments. Or it may be desirable to ascertain the correlation be-

tween type (tubular or ligulate) of a composite flower and the

number of divisions in the corolla. Or one may wish to measure

the relationship between type and time required for germination

in the seeds of a dimorphic plant species. Or a series of individu-

als may be classified by the social worker or prison warden as

alcoholic and non-alcoholic and the investigator desires to corre-

late between alcoholism (which is really a graduated character,

although classified in the available records into the two alterna-

tive classes only) and any physical measurement or the extent of

criminality as measured by number of convictions or months

spent in prison.

In this reduced form the data can no longer be treated by the

correlation ratio method, but must bé attached by a recent for-

mula due to Pearson,’ and known as the Bi-serial correlation

coefficient.

Soper” has continued his work on the probable error by deter-

mining the standard deviation of constants caleulated by this

formula.

Both Characters Classified in Multiple Categories.—If instead

28 Pearson, K., ‘‘On a New Method of Determining Correlation Between

Ct |

the Correlation Ben ah Biometrika, 10: am Toe

60 THE AMERICAN NATURALIST (Vou. L

of both characters being measurable on a quantitative scale, or

one character recorded in a number of categories and the other

measurable on a quantitative scale, both characters are not quan-

titatively measurable, but describable in a number of classes only,

neither the correlation coefficient nor the correlation ratio can be

used. In such cases, which in practical work are very frequent,

Pearson’s contingency methods®® must be used. These have been

too long in use to require discussion or illustration here. Certain

corrections to be applied will be considered at another time.

The probable error of the contingency coefficient presents con-

siderable difficulty. Those who have to deal with it should con-

sult papers by Blakeman and Pearson** and by Pearson.*?

One Variate Classified in Alternative, the Other in Multiple

Categories.—Consider a contingency table reduced to a two-fold

grouping for one of the characters, but retaining the multiple

division for the other. Such a table is comparable with the con-

densation of the correlation ratio table discussed above. It must

be analyzed by a special method.*?

The formula has not as yet had extensive practical application.

It has been used to determine the relationship between alcoholism

as an alternative character and type of crime classed in multiple

categories, and between alcoholism in the parent and health of

the children. It may prove especially valuable in dealing with ©

the interrelationship of various teratological conditions in mor-

phological work.

Both Characters Classified in Alternative Categories Only.—As

the extreme case we may think of a contingency table reduced

to a two-fold grouping for each of the characters. This is then

the four-fold table for alternative characters, i. e., (A) and (not

-A), (B) and (not -B).

In the past, two methods have been chiefly employed for obtain-

ing constants from such tables, Pearson’s four-fold correlation

coefficient and Yule’s coefficient of association.

30 Pearson, K., ‘On the Theory of Contingency and its Relation to Asso-

ciation and Normal Correlation,’’ Drapers’ Co. Res, Mem., Biom. Ser., I.

Dulan & Co., 1904

31 Blakeman, John, and K. Pearson, ‘‘On the Probable Error of Mean

Square Contingency,’’ Biometrika, 5: 191-197, 1906.

82 Pearson, K., ‘On the Probable Error of a Coefficient of Mean Square

Contingency,’’ Biometrika, 10: 570-573, 1915.

33 Pearson, K., ‘‘On a New Method of Determining Correlation when One

- Variable is Given in Alternative and the Other in Multiple Categories,’’

Biometrika, 7: 248-257, 1909.

No. 589] NOTES AND LITERATURE 61

For several years critical workers have realized that very little

reliance is to be placed upon Yule’s very simple coefficient of

association. This coefficient and another measure of correlation

‘‘the theoretical value of r°’ proposed in his ‘‘Introduction to

the Theory of Statistics’? have been discussed by Heron.**

Pearson and Heron** and Pearson** have gone into these methods

and others proposed by Yule’ in a masterly way. To discuss

this memoir alone would require far more than the space avail-

able for this general index of the correlation methods. Their

treatment can leave no doubt—if any existed in the minds of

those who have tried to use these formule in serious statistical

work—that except in very special cases all these association and

colligation formule are likely to work harm rather than to be of

service in the hands of the biologist.

This demonstration of the untrustworthiness of the various

substitutes for the correlation coefficient practically throws us

back upon the old four-fold method of Pearson, and upon another

novel method to be discussed ina moment. The difficulty of com-

putation has been one of the greatest obstacles in the way of the

more general application of this method and has frequently

resulted in the substitution of the less reliable coefficient of as-

sociation. The necessary labor of calculation has been much

reduced by two series of tables by Everitt.**

The determination of the probable error of the coefficient of

correlation calculated from the four-fold grouping has always

een excessively laborious. While four-fold correlations have

been calculated in hundreds of cases, the determination of the

probable error has been made for less than a hundred of the

coefficients. Pearson®® has now given tables to facilitate the cal-

84 Heron, D., ‘‘ The Danger of Certain Formule Suggested as Substitutes

oo Coefficient,’’ Biometrika, 8: 109-122, 1911.

ce = . . .

Teh eA and D. Heron, ‘‘On Theories of Association, ’”’ Biometrika,

mies a = the Surface of Constant Association,’’ Bio-

a e, G. U., ‘On the Methods of Measuring Association between Two

ae. na Soc., 75: 579-641, 1912.

Nit heel 1 òk beeta mae Tetrachoric Functions for Four-fold Cor-

i trika, 7: 437-451, 1909; ‘‘Supplementary Tables

metrika, 8- in ‘Ta

nh Oia 7: 885-806, 1912. Also in ‘‘Tables for Statisticians and Bio-

2 .

Pearson, K., ‘‘On the Probable Error of a Coefficient of Correlation as

62 THE AMERICAN NATURALIST [Vou L

culation of approximate probable errors which are sufficiently

exact for all practical purposes.

Finally, the most important recent development in the theory

of correlation is probably Pearson’s novel method of dealing with

variates classed in alternate categories only.*°

The fundamental conception of this method is exceedingly

simple. Given the table,

A; Ao Totals

Do are es Oa a b a+b

Pe ey ue c d c+d

Potala. siran a+c b+d N

where the large letters represent any alternative (e. g., Men-

delian) characteristic of an individual, and the small letters

denote the frequency of occurrence of the several possible combina-

tions, it is clear that

ate b+d a+b c+d

YN We

give the independent probabilities of the two pairs of character-

istics. The four pertinent products of these ratios give the

chances on the assumption of the independence of the two char-

acters A and B, of the four possible combinations. Then if there

be no correlation, within the limits of the errors of random

sampling

a WISE aT)

a N (S52 x24?) =o,

and so on. The squares of the four differences between the ob-

served frequencies, a, b, c, d, and those which would be expected

if the two characters were really independent, gives the familiar

x’ of Pearson’s test for goodness of fit. The significance of this

test may be determined from Palin Elderton’s tables,“ and this

is, in the case in hand, a measure of correlation. It has been

Found from a Four-fold Table,’’ Biometrika, 9: 22-27, 1913. Also in

Tables for Statisticians and Biometricians,

40 Pearson, K., ‘‘On a Novel Method of Regarding the Association of

Two Variates Classed Solely in Alternate Categories,’’ Drapers?’ Co. Res.

Mem., Biom. Ser., VIT, Dulan and Co., 1912,

41 Elderton, W. P., ‘‘Tables for Testing the Goodness of Fit of ‘Theory

to Observation,’’ Biometrika, 1: 155-163, 1901. Also reprinted in Pear-

son’s volume of tables.

No. 589] NOTES AND LITERATURE 63

used as such during the past several years by some of us in prac-

tical problems in which we found it impossible to place reliance

upon Yule’s coefficients and did not feel warranted, because of

underlying assumptions, in depending solely upon the classical

four-fold method. But it is a measure given in terms utterly in-

comprehensible to the ordinary mind, which is quite incapable of

thinking in millions or in multiples of millions.

What Pearson has done with such brilliancy is to furnish a

means in mathematical theory and working tables of passing from

the incomprehensible scale of pure probability to the familiar and

usable and widely comparable scale of correlation.

As yet it is too soon to be able to state the results of extensive

practical application of the new coefficients, but they should have

wide usefulness.

Both Characters Classified by Rank in Series.—In some cases,

neither measurements nor classification of the individuals dealt

with in categories are given in the data, but merely their position

or rank in the series.

Rank may be numerically expressed, and the suggestion has

been made that the correlation of grades or ranks is a quite legiti-

mate measure of interdependence in such cases. Pearson** has,

however, pointed out the very real difficulties encountered in

such work. Those who are tempted to use these methods should

acquaint themselves with the dangers as pointed out in this

memoir.

One Variate Given by Rank in Series, the Other Measured on

a Quantitative Scale-—Such cases are not likely to occur with

great frequency in biological work. Possible instances are those

in which one wishes to correlate between position in an intensity

of pigmentation series and size or fertility—both quantitatively

measurable characters. The formule have been given by

Pearson.**

One Variate Given by Multiple or Broad Categories, the Other

by Rank in Series——Practical applications in biology should be

rare. For formule see the paper by Pearson just cited.

From the foregoing outline it must be clear that of recent

years the conception of correlation has been greatly extended and

the possibilities of the practical usefulness of correlation methods |

42 Pearson, K., ‘‘On Further Methods of Determining Correlation,’’

Draper’s Co. Res. Mem. , Biom. Ser., IV, Dulan and Co., 1907. :

-or Ranks,’? Biometrika, 10: 416-418, 1914.

64 THE AMERICAN NATURALIST [ Vou. L

vastly increased by the deduction of formule suitable for dealing

with data of the most diverse sorts.

The most valuable feature of a summary such as the present

may possibly not lie in the fact that it exhibits to biologists the

wide array of statistical tools which are now available for dealing

with the most diverse sorts of data which they may collect, and

shows where directions for their use may be found, but in the

suggested warning that the hasty application of the first learned

or the most easily calculated formula may lead to constants of

little value. Most biologists can use a scalpel or a beaker with

great success, but many at least would hesitate to try to handle

without special training all the instruments which are to be seen

in the surgeon’s case or to use all the glassware on the organic

chemist’s shelves. Each kind of tools require their special train-

ing. Notwithstanding popular conceptions to the contrary, this

is also true of the biometric tools.

J. ARTHUR HARRIS

VOL. L, NO. 590 FEBRUARY, 1916

EHE

AMERICAN

NATURALIST

A MONTHLY JOURNAL =

Devoted to the Advancement of the Biological Sciences with =

The American Naturalist

MSS intended for publication and bocks, eto., , intended for review should be

sent to the rgi of THE AMERICAN NATURALIS rrison-on-Hudson, New York,

Sho icles containing summaries of ‘tesa earch work bearing on the

problems Steg fakar evolution are especially welcome, and will be given preference

"po

ne hundred red reprints of contributions are supplied to authors free of charge.

Puvtiar a reprints will = supplied at cos

a Oae tenet y giaet should oie a the pirs. 1

8 on oi ‘our ars a year. F age

Canadian postage twenty-five cents additional. The charge for single copies is

forty cents. The Divort sin rates are Four Dollars for a page.

THE SCIENCE PRESS

Lancaster, Pa. Garrison, N. Y.

-NEW YORK: Sub-Station 84

Entered tter, April 2, 1908, at the Post Office at Lancaster, Pa., under the Act ot

Congress of March 3, 1879.

: BO O KS F OR OR SALE TAKAN NATURAL STONY SPEARENS

Condition and Low

Specialty: Bird Skins, Oology, Entomology, Marine

NATURAL HISTORY Animals and others. Catalogue free. Correspond-

solicited.

_ JOHN D. SHERMAN, |

a Jr. i T. FUKAI, eaten

. ‘£08 Satata Avinas VERNON, N. Y. i, Saitama, Japet

o | : ‘PUBLISHED BY THE 1 CAMBRIDGE UNIVERSITY PRESS

_ | MIMICRY IN BUTTERFLIES

THE

AMERICAN NATURALIST

Vou. L. February, 1916 No. 590

A FURTHER ANALYSIS OF THE HEREDITARY

TRANSMISSION OF DEGENERACY AND

DEFORMITIES BY THE DESCENDANTS

OF ALCOHOLIZED MAMMALS

CHARLES R. STOCKARD ann GEORGE PAPANICOLAOU

DEPARTMEHT OF ANATOMY, CORNELL UNIVERSITY MEDICAL SCHOOL,

i New York City

Introduction.

Material and methods.

Direct effects of the aleohol treatment on the animals.

Influence of the treatment on the descendants of alcoholized animals.

The influence of internal and external factors on the quality of the

offspring,

Relative conditions of the male and female descendants from paternal

and from maternal alecoholized ancestors.

General considerations.

Summary and conclusions.

Literature cited.

INTRODUCTION

A LITTLE more than two years ago the senior author

(Stockard, °13) recorded in this journal experiments

which had then been running for three years and seemed

to show a definite injury of the germ cells by treating

mammals with the fumes of alcohol. This injury of the

male germ cells is of such a nature that an alcoholized

male guinea pig almost invariably begets defective off-

Spring even when mated with a vigorous normal female. —

At that time it was also shown that F, animals, the off-

Spring of treated parents, though themselves not treated,

65 : eo

66 THE AMERICAN NATURALIST [Von. L

had the power to transmit the defective condition to their

young, and such F, young were equally if not more de-

fective than the immediate offspring of the treated

animals.

In 1914 in a short abstract Stockard showed further

that the offspring from F, individuals were apparently

more defective than their parents and were often grossly

deformed. One case was recorded of the occurrence of

a litter of two F, animals, both of which were extremely

weak and neurotic, showing a condition suggesting paral-

ysis agitans, and further than this the two animals were

typical anophthalmic monsters. The eyes were com-

pletely absent, no optic nerve or optic chiasma or visible

optic tracts along the tuber cinereum could be found on

a careful gross examination of the brain. The two ani-

mals were produced by parents (F,) that had never been

treated with alcohol, the four grandparents (F,) had also

not been treated, while the three great grandfathers had

been alcoholized and the three great grandmothers were

normal untreated individuals.

Defective eyes and absence of one eye or both eyes have

been frequently met with in the experiment, as well as the

peculiar nervous condition, and these symptoms are to be

considered indicative of the injury or change induced in

the male germ cells by the experimental treatment, which

in the above case was transmitted through three genera-

tions. No question could remain as to the action on the

germ cells, as only male ancestors had been treated;

every female of the line was an untreated animal.

This abstract called attention to the fact that there was

a tendency for the results to differ in subsequent genera-

tions from treated males as compared with the descend-

ants of treated females—not enough data were then pres-

ent to offer any explanation of these differences and a

consideration of them will be undertaken in the present

paper.

At that stage of the experiment it was also difficult to

offer an exact analysis of the mode of transmission of

No. 590] TRANSMISSION OF DEGENERACY 67

thè defects and the type of injury induced by the alcohol

treatment, since the total numbers were not large and

the F, animals had only a few matings, while further gen-

erations had not become available for breeding.

The same experiments have now been continued for

more than five years and a number of animals have been

used, over 700, which cover the behavior of four genera-

tions and supply data of sufficient extent to allow a more

thorough analytical consideration of the heredity prob-

lem concerned.

Experiments of this nature on mammals are fraught

with many difficulties, slowness of breeding, small size of

litters, difficulty of handling, ete. Yet such material of-

fers one very great advantage in that the quality of the

offspring and generations studied is of such a complex

that one is enabled to detect indications of rather slight

injuries or changes in the material carriers of heredity

which would not become evident on lower forms with less

diversity in their methods of behavior and structural ap-

pearance. In other words, we take it that such condi-

tions as are spoken of as racial degeneracy in man and

Mammals are often very difficult or even at times im-

possible to detect in lower forms.

These conditions are for many reasons thought to be

inherited. If so their inheritance must be due to a path-

ological condition of the material carriers of heredity,

the chromosomes, or what not, since they are not normal

states and, like diseases, are constantly arising in normal

families on account of one or another form of intoxica-

tion. Is it possible then to produce such a racial degen-

eracy artificially by treating only one generation of the

animals and by so doing observe a pathological behavior

of the carriers of heredity? Arguing from analogy there

must be pathological heredity due to diseased or altered

chromosomes in the germ cells just as truly as there is a

known pathological behavior of every other organ and

tissue of the animal body. ae

It becomes then a problem to study the possible meth-

68 THE AMERICAN NATURALIST [ Von. L

ods of modifying the chromosomes or carriers of the in-

herited qualities of organisms in order to further analyze

their normal physiological behavior; in the same way

that experimental embryology has been able to supply so

many valuable clues to the normal processes of devel-

opment.

In the following pages we believe the facts indicate

that individual guinea pigs are now living in this experi-

ment that have had the carriers of hereditary qualities,

the chromatin, of their germ cells injured for a longer

time than four years. And during this time they have

given rise to offspring of more or less degenerate or de-

formed type, and in some cases these offspring have

passed this modified chromatin on through three genera-

tions, all of which contain pathological chromatin and

show somatic defects and deformities as an index of their

tainted chromatic ancestry. Modified chromatin has

been living in the experiment for more than four years

in five different generations of animals as a result of the

treatment on the one original, P,, parent generation.

We have tried to regulate every controllable source of

error and there can be no doubt that the conditions are

brought about in the way described. Could the degen-

eracy which is so pronounced have previously existed in

the stock? This question has been controlled in the first

place by the use of two entirely different stocks from dif-

ferent sources and obtained one and one half years apart,

the first in the fall of 1910 and the other in the early win-

ter of 1912. The responses of the two stocks to the ex-

perimental treatment have been identical. As a second

method of control every animal has been tested by one or

more normal matings before being introduced into the

experiment, and only those giving normally strong off-

spring have been used. A further crucial control is the

constant mating of normal untreated animals from both

stocks under identical cage conditions with the experi-

mental individuals. These animals continue to breed

normally until very old, when they gradually become

No. 590] TRANSMISSION OF DEGENERACY 69

sterile. But none have ever given rise to a defective or

deformed individual, and the rate of mortality of the

young indicates the average healthy condition found in

normal guinea-pig breeding. There is a striking contrast

between the records of these normal young and the mor-

tality record, the frequency of easily recognized nervous

symptoms of degeneracy, and the prevalence of gross de-

formities in the experimental races.

The external as well as internal factors are to be consid-

ered not only in individual or embryonic development, but

also in heredity. And the present experiments now dem-

onstrate for mammals that either the spermatozoon or the

ovum may be experimentally injured or modified in such

a manner as not only to give rise to (abnormal) sub-

normal development in the resulting embryo, but the

effects of the injury may be transmitted from generation

to generation, until an affected line actually fades out

through degeneracy and sterility as a result of the trans-

mitted condition.

MATERIAL AND METHODS

The animals used in the experiments have been ordi-

nary vigorous guinea pigs of large size, particular care

being taken to select animals less than one year old to

begin with and good breeders.

At the beginning of the experiments alcohol was given

along with the food, but the animals ate less and the food

usually disagreed with them. It was then administered

in diluted form by a stomach tube; this method was even

more unst ful, disturbing digestion and seeming to

upset the animals considerably. It is certain that alcohol

given to animals through the stomach deranges their ap-

petite and digestion to such an extent that the experi-

menter is unable to determine whether the resulting ef-

fects are due to the aleohol, as such, or to the generally

deranged metabolism of the animal. When given in

drinking water they take little or none of the water and

the treatment is insufficient. For these reasons an inha-

70 THE AMERICAN NATURALIST [Vou. L

lation method of treatment was resorted to early in the

study, and, as far as experience goes, it has no serious

disadvantages and does not complicate the conditions of

the experiment.

This method may be merely described in brief for the

convenience of the reader, since it has been fully recorded

with illustrations of the fume tanks in previous publica-

tions. A fume tank of copper is made of sufficient size

to supply breathing space for four or five guinea pigs at

one time. The tank has four outlets, so that a definite

amount of fumes may be passed through in a given time

and the ventilation controlled. In this way each animal

could be given a definite measured dose. The individ-

uals, however, differ so much in their resistance to the

treatment that it has been found better to treat all to

about the same degree of intoxication. Such a physio-

logical index is more reliable, since every animal may be

affected to the same degree each day. For this purpose

the animals are placed in the fume tank on a wire screen,

and absorbent cotton soaked with alcohol is placed be-

neath the screen, so that they inhale the alcohol fumes

arising from the cotton to saturate the atmosphere of the

tank.

Ether was given in a similar manner. The animals

are much more readily overcome by these fumes and

must be carefully watched while inhaling even the most

dilute doses.

To avoid handling the females during pregnancy, spe-

cial treating cages are devised. An ordinary box-run

with a covered nest in which the animal lives is connected

by a drop-door with a metal-lined tank, having a similar

screen arrangement, etc., to that of the general treatment

tank. The pregnant animal may be driven daily into the

tank and thus treated with alcohol fumes throughout her

pregnancy without being handled in any way that might

disturb the developing fetus.

Particular care is necessary in mating the animals in

such an experiment, as the females are often slow to con-

No. 590] TRANSMISSION OF DEGENERACY 71

ceive and some of the F, and F, individuals of both sexes

are not very prolific and in many cases are almost or

quite sterile. Each female is kept in a separate run and

the male is placed in with her just before the time of the

expected heat period, ovulation, and he remains in her

cage for from two to three weeks so as to be present at

the second ovulation, provided the female was not made

pregnant by the first mating. The ovarian cycle of the

guinea pig as worked out by L. Loeb seems to correspond

closely to what is found in mating experiences.

After mating, the male is removed from the cage and

the female remains alone until the young are born.

These are left with the mother for about fifteen days, then

separated, and the female mated again. In this way the

normal females may sometimes give as many as four lit-

ters per year, but the experimental animals breed much

slower and it is difficult to get even three litters per year.

Direct EFFECTS or THE ALCOHOL TREATMENT ON THE

ANIMALS

Several of the guinea pigs have now been treated with

the fumes of alcohol almost to the point of intoxication

for six days per week for a period of five years. This is

a considerable space in the life of a guinea pig, which

probably would not often extend beyond six or seven

years.

The animals are affected by the alcohol fumes in vari-

ous ways; some of them are stupefied and become drowsy,

while others become stimulated and excited and some-

times even vicious, constantly fighting and biting at the

other animals in the fume tank. The fumes inhaled into

the lungs pass directly into the circulation, so that the

animals show signs of intoxication very soon after being

put into the tank, yet the intake of alcohol is so gradual

that they may remain for one hour or more without be-

_ Coming totally anesthetized. | oe

The mucosa of the respiratory tract is considerably

12 THE AMERICAN NATURALIST [Vou. L

irritated during the early stages of the treatment, but

develops a resistance so that later little effect can be no-

ticed. The cornea of the eye is greatly irritated, often

becoming milky white and opaque during the first few

months. In some cases this later clears and the animal

is again able to see, though some of the animals treated

for several years have remained entirely blind. The gen-

eral condition of the animals under the fume treatment

is very good. They all continue to grow if the treatment

is begun before reaching their full size, and become fat

and vigorous, taking plenty of food and behaving in a

typically normal manner.

Some of the treated animals have died and others have

been killed at different times during the progress of the

experiment and their organs and tissues examined care-

fully and then studied microscopically. All have seemed

practically normal. Tissues from several animals treated

as long as three years have been examined and the heart,

stomach, lungs, kidneys, and other organs present no no-

ticeable conditions that might not be found in normal

individuals. Aleoholized animals are usually fat, but

there is no fatty accumulation in the parenchyma of any

of the organs.

Several of the animals, both males and females, have

been partially castrated during the experiments and the

ovaries and testes have been found to be in a healthy con-

dition, though certain possible changes may be present

which are now being closely studied cytologically and

experimentally.

The treated animals are, therefore, little changed or in-

jured so far as their behavior and structure goes. Nev-

ertheless, the effects of the treatment are most emphati-

cally shown by the type of offspring to which the alcohol-

ized individuals give rise, whether they be mated together

or with normal individuals. The further significance of

the nature of the effects is indicated by the quality of the

subsequent generations descended from such an ancestry.

No. 590] ` TRANSMISSION OF DEGENERACY 73

INFLUENCE OF THE TREATMENT ON THE DESCENDANTS OF

ALCOHOLIZED ANIMALS

It may be well in the first place to consider the results

of the experiments from a general standpoint and then

to undertake an analysis of the reactions and conditions

presented in the several generations and from the several

lineal combinations. The records of the matings of the

aleoholized animals in various pairs, the control or

normal matings, and the matings of the F, and F, gen-

erations, the children and grandchildren of the alcohol-

ized individuals are summarized in the general Table I.

This table gives a record of all the matings of the kinds

indicated up to July 1, 1915. A similar table was pub-

lished two years ago, when the number of animals con-

sidered was much smaller and the actual indications from’

the results were less certain than now. On comparing

this table with the former one, however, it will be seen

that the continuation of the experiments has fully sub-

stantiated the results as previously recorded. The table

now shows the records of 571 matings which produced

682 full-term young and 189 early abortions or negative

results. These numbers are now of considerable magni-

tude in spite of the fact that the experiment is conducted

on mammals which produce only small litters and breed

slowly as compared with lower animal forms.

In the first horizontal line the record of pairing alco-

holized male guinea pigs with normal females is given.

This combination could only produce defective or sub-

normal young as a result of the injured male germ cells,

Since the ova are normal and develop in a normal un-

treated mother. This then is the definite test of the in-

fluence of the alcohol treatment on the germ cells.

Ninety such matings have in 37 cases given negative

results; that is, failures to conceive, or early abortions.

_ -Aus 41 per cent. of the matings of such males were non-

Productive, while less than 25 per cent. of normal mat-

Mgs under the same breeding conditions failed to produce

full-term litters. Ten stillborn litters, each consisting of : a

74 THE AMERICAN NATURALIST [Vou. L

two young, twenty stillborn young, resulted from the 90

matings. While the 90 control matings gave only two still-

born litters, and in both cases these were unusually large

litters of four individuals each, and they were probably

dead on account of the fact that the mother could not give

normal birth to so many offspring. The stillborn litters

by the aleoholized fathers were all ordinary-size litters of

two young. Thus, while 11 per cent. of the matings of

aleoholized males resulted in stillborn litters, only 2 per

cent. stillborn litters occurred from normal matings.

Forty-three living litters were produced or a little less

than 48 per cent. of the matings gave full-term living

young, while 73 per cent. of the normal matings give liv-

ing litters of young.

The 43 litters from alcoholic fathers contained in all

82 young, and 35, or almost 43 per cent., of these died soon

after birth, while 66 similar litters from the control lost

only 19 young, or 16 per cent., out of 118 individuals.

Finally, then, from the 90 matings of alcoholic males with

normal mates only 43 full-term litters resulted, consisting

in all of 102 young; 55 of these, or 54 per cent., died at

birth or soon after, and only 47 individuals, or 46 per

cent., survived. Only about half as good record as the

78.5 per cent. surviving young from the matings of normal

animals. Almost all of the offspring were very excitable,

nervous animals and three of them showed gross deformi-

ties of the eyes, while no such conditions were found

among any of the offspring of normal animals bred under

identical conditions.

These records leave no doubt that the alcoholized male

guinea pig is injured in such a way as to induce a decid-

edly bad effect upon the quality and mortality of his off-

spring when compared with the records from normal

animals.

The second horizontal line of Table I shows the results

obtained when alcoholized female guinea pigs are paired

with normal males. In this case there is a double chance

to injure the offspring. First through the influence of

No. 590] TRANSMISSION OF DEGENERACY TO

the treatment on the oocytes or the unfertilized ova-

rian egg, a direct effect on the germ cells comparable

to the injury of the germ cells in the case of the treated

males considered above. While in the second place,

the developing embryo in the uterus of an alcoholized

female may be directly affected by the strange sub-

stances contained in the blood and body fluids of the

mother. Thus a defective individual may be produced

as a result of development in an unfavorable environment

or as a result of being derived from an injured or de-

fective egg cell.

Thirty-three matings of alcoholized females with nor-

mal males have in seven cases, 21 per cent., given nega-

tive results or early abortions; this compares very favor-

ably with the records of the control animals. Four

stillborn litters consisting of three individuals each were

produced. This is a record of 12 per cent. stillborn lit-

ters against only 2 per cent. from normal matings. The

aleoholized females gave birth to 22 living litters con-

taining 44 young, and 23, or 52 per cent., of these died,

only 48 per cent. surviving against 84 per cent. survivals

among the young of similar control litters. The records

of the matings of aleoholized females compare very unfa-

vorably with the record of the control matings. Yet the

behavior of the aleoholized females is very little, if any,

worse than the records shown by the alcoholized males in

spite of the double chance the female has to injure her

young.

The third horizontal line of the table indicates the re-

sults obtained when alcoholized males are paired with

aleoholized females. Here there is every chance for the

treatment to show its effect. The percentage of early

abortions or negative results is very high, about 50 per

cent. more than double that of the control matings. Ten

per cent. of the matings produced stillborn litters each

Consisting of two young. Only 17 living litters were born

out of 41 matings, about 41 per cent., against 73 per cent.

living litters from 90 control matings. The 17 living lit-

76 THE AMERICAN NATURALIST (Von: L

ters contained only 26 young, and 12 of these, or 46 per

cent., died soon after birth, while but 16 per cent., or one

third as many, of the control offspring died out of a total

of 118 individuals. The data from the double alcoholic

matings is, therefore, extremely bad in the light of normal

matings from the same animal stocks bred under exactly

the same cage and food conditions.

The fourth horizontal line summarizes the records of

all the matings of directly alcoholized animals. In all

164 such matings have been made; 64 of these, or almost

40 per cent., gave negative results or early abortions.

Highteen stillborn litters occurred, consisting of 40 indi-

viduals against only two questionable stillborn litters

from 90 control matings. Eighty-two, or only 61 per

cent., living litters were born, consisting of 152 individ-

uals, 82, or 54 per cent., of which survived and 70, or 46

per cent., died soon after birth; in all 110 full-term young

died, while only 82, or 42 per cent., of the total 192 full-

term young resulting from the 164 alcoholic matings sur-

vived. On the other hand, out of a total of 126 full-term

young from only 90 control matings, 99, or 78.5 per cent.,

survived. The control matings were far more prolific

than those of the alcoholized animals and the condition

of the young as indicated by the mortality record was far

superior to that of the alcoholic offspring.

The fifth line records the outcome of 90 control mat-

ings which have been scattered through the entire prog-

ress of the experiment under exactly the same conditions

and from the same animal stocks as the experimental

matings. Highty-four per cent. of the young in the 66

living litters resulting from the matings of normal ani-

mals have survived and all are strong, healthy individ-

uals; in not one instance do they show an indication of

nervous degeneracy or any type of recognizable struc-

tural deformity, while such degeneracy as well as de-

formities are extremely prevalent among the offspring

and descendants of the aleoholized animals. One other

point to be mentioned in considering the records of the

TRANSMISSION OF DEGENERACY

No. 590]

R RE E TI De a

pi SI 9 L 9 Z 6 SI

om ‘puoves X 'ıəuə3 py} ‘

Tis eit 0 0 I OU ease "* ooyooys x ‘sous PAL

ii st g or 8 ¥ £ mooo t-t te qewoux '1auo3 payg,

EA pe ‘aro 7 ‘aed [ ‘OT yI 8 P Cl £g rrer t te Iouog3 puooasxX ‘sous PIL .

ssojoAo : :

€I| SE g “opo wd ‘6l | 12 830] "Pp T ‘FT L oz gt oe ou PIX ‘10u03 PIL

ssojoAe

q “a əuo T “O'P § ‘ZS| 69 ‘p 9 “wd g ‘ep | 69 9I $ 6z g6 |°''' ' ‘‘1eueB puooes xX “10003 puoveg

83| 68 opez] og op Tt ‘Lt 8 9T Moor SS ‘++ ‘oToyooys X ‘toes puooeg

‘eo E ‘GZ| LE ‘red g ‘Og | Eg "0°90 9 ‘g € OL OF Aone ae uopeIouo3 puooeg

4| f I 4 0 0 0 et ee *Aouvusord Jump poywos, ġ

66| Lz 61 | 99 8 Z gZ C6 i "+ § 'WIOUX O ‘WIU 1014000

Z8! OTI OL | g8 OF SI ¥9 POT ee

FI| 02 CL i dE 8 j4 0z Poo oy 3 oroyooex O onogoory

1Z| ge ez | 3g Zí ¥ yi S R ` ooyooye x P *uLION,

90g ‘LF 29 owi ‘oe | &F 0Z oT Le Ls as Seca a ates 6 eens © oTOyoory

aontodV |ssupen

sunox Suppan pud yg 194V 81979 T Bunog BOVE ALOE Jo sreupuy oy} Jo uOPIpUCD

A PADR T80L uoog 3uUMKqa sunox ZUMI uIoqrng JO Jequin N u10q {Is Jo soe saquinNn :

8A} B39 N

SIVWINY GALVANI 40 SLNVINAOSA FHL NO IOHOITY AO SLOT

I Wiavi

78 THE AMERICAN NATURALIST [Von. L

control matings is the fact that from 90 matings only two

stillborn litters were produced and, as mentioned above,

both of these litters were of so large a size that the

mothers seemed unable to successfully deliver them and

one of the mothers failed to recover from the process and

died a few days later. These two cases make the con-

trol records appear worse than they actually should, but

in spite of this the control matings have given data

equally as good as those generally obtained by careful

breeding experiments with vigorous normal stocks. The

stock in these experiments is unquestionably good, as the

control matings very readily show.

Four normal females were mated and then treated with

alcohol throughout their periods of pregnancy and, as the

sixth horizontal line of the table indicates, such a treat-

ment was not at all injurious in these particular cases.

It actually happened that some of these young were un-

usually vigorous. The numbers are very small, but this

is a direct test, and if such a treatment were really decid-

edly effective in its action on the embryo or fetus in utero

these eight young animals should have at least shown

some response. It is very possible that after the treat-

ment has been continued for a long time, a year or more,

that the mother then presents a uterine environment un-

favorable for normal development, since the offspring of

such individuals are almost always subnormal. In these

cases, however, the inferior quality of the offspring may

be due to the action of the alcoholic treatment on the

ovarian germ cells rather than the direct environmental

effect on the developing embryo or fetus, there is no way

at such a stage to separate the two possible effects.

The next three horizontal lines, seventh, eighth and

ninth, give the data resulting from the matings in vari-

ous combinations of the F, animals, that is, offspring

from alcoholic parentage, but which are not themselves

treated with aleohol. The records of these non-treated

F, individuals are most instructive for an understanding

of the actual influences of the alcoholic treatments.

No. 590] TRANSMISSION OF DEGENERACY 79

When such F, animals are paired with normal indi-

viduals the seventh line shows that almost 22 per cent.

of the matings failed, which is not a bad record. The

proportion of stillborn litters, however, from the F, by

normal combination was three times as great as from

normal matings and 75 per cent. of the stillborn young

produced showed gross defects of the eyes, having opaque

lenses or typical cataract conditions, while not one of 126

young from normal matings has shown this or any other

noticeably abnormal structure. Thirty-three living lit-

ters were produced containing in all 54 individuals, 29,

or 54 per cent., of which died soon after birth, while 25

survived. Two of those dying soon after birth were par-

alyzed and unable to walk, while three of the 25 survivors

have defective opaque eyes, cases similar to that illus-

trated by Fig. 1, and many show different nervous symp-

toms. Thus of 62 full-term young produced by F, ani-

mals with normal mates, only 25, or 40 per cent., survived

for more than a short time after birth, and 12 per cent.

of these have gross defects and more than half of them

are nervous, excitable individuals, which when mated with

normal animals or in any other combination always give

very poor quality offspring, if any at all.

The eighth line shows the records of 53 matings be-

tween F, animals and alcoholics. This combination again

gives data comparing most unfavorably with the control

and in some ways even worse than the records of matings

between two alcoholic animals. Fifteen per cent. of such

matings produced stillborn litters! Only one combina-

tion gives a worse record of stillborn that is, matings

among F, animals. Almost half of the young in the liv-

ing litters died and here again some were deformed. De-

formities are strikingly more abundant among the off-

Spring from F, and F, parents than from the directly

alcoholized animals.

The record of 95 inter se matings of F, animals is

Shown in the ninth line. Thirty per cent. of such mat-

= 8S gave negative results or early abortions, over 7 per

80 THE AMERICAN NATURALIST [ Vo.

No. 590] TRANSMISSION OF DEGENERACY

a Fia. 2, A and B. 307 Fz Ọ (one in litter). Inbred, A brother anā sister

The eye of one si = perea , the other

offspring of alcoholic mal -

e (AN) ( AN). e A

eye ball apparent tly absent on living examination. A typical er monoph- —

thalmicum asymmetricum. This animal now’21 months old “fs completely sterile. —

82 THE AMERICAN NATURALIST [Von. L

cent. stillborn litters and 62 per cent. living litters. Little

less than half of the living young died soon after birth,

in all 43, nine of which, or more than one in five, 21 per

cent., were paralyzed or deformed; the figures in Plates

I and II illustrate the paralytic conditions. Fifty-two of

the offspring survived, three with deformed eyes, one

with one eyeball completely absent, monster monophthal-

micum asymmetricum (Fig. 2, 307 2), and almost all of

the 52 are very nervous, excitable animals which when

bred give rise to deformed or highly degenerate offspring.

The offspring from the F, animals mated in any com-

bination are generally far below the normal in power to

survive and in quality of structure. When compared

with the offspring from directly aleoholized parents, the

offspring from the F, combinations show an equally bad

mortality record and a very much higher proportion of

paralyzed and deformed individuals. The 95 matings

inter se of F, animals demonstrate conclusively that such

individuals carry defective or abnormal germ cells which

give rise to defective developmental products. These de-

generate F, offspring owe their subnormal condition to

the effects of the action of the alcohol treatment upon the

germ cells of their grandparents which have been trans-

mitted to them through their parents. In other words,

the carriers of hereditary qualities have been modified

in the first parental generation, and the effects of this

modification are expressed in their offspring F,, and also

in their grandchildren, the F, generation.

The next line of the table, the tenth, indicates further

how the effects of the original modification are trans-

mitted to the great grandchildren or through three gen-

erations since the injury. Forty-eight inter se matings

of F, animals gave the results here shown. Almost 42

per cent. of the matings gave negative results or early

abortions, the poorest record in this respect shown in the

entire table. About 15 per cent. of such matings gave

stillborn litters, 7 in 48 matings, which is remarkably

high when compared with any of the above combinations.

No. 590] TRANSMISSION OF DEGENERACY 83

The hind legs of one of the stillborn young were deformed

in the peculiar manner illustrated in Figs. 4 and 5.

Twenty-one living litters were produced, containing in

all 32 young; 19 of these, almost 60 per cent., died: soon

Fie Two in litter, both same — three normal grea t-grand-

mothers. ua ae alcoholic great-grandfath The n ents were single first

cousins, Both animals completely eyeless, si with para agitans, one died

the second and the other the third day after birth, haste anophthalimia me

brain no indication of optic nerve, the other slight proc

_ after birth, and only 13 survived. One of ie 19 that died

= Was paralyzed and unable to stand, while 8 of them, a

Strikingly high proportion, were “et daformed. a - a 2

84 THE AMERICAN NATURALIST [Vou. L

had one or both eyes deformed (Figs. 1 and 2), and two

were anophthalmic monsters, being completely without

eyeballs, optic nerves, optic chiasma or any gross signs

of optic tracts (Fig. 3). The brains are now being

studied in sections. Figs. 1 to 3 illustrate animals show-

ing the different eye conditions—asymmetrical eyes, mon-

strum monophthalmicum, and anophthalmic monsters.

Figs. 5 and 6 of Plate III illustrate the brains of a normal

and an anophthalmic specimen for a comparison of the

condition of the optic nerves, ete.

Forty-six full-term young were produced by the F,

matings, but only 13 of these, or just 28 per cent., were

able to survive, while about three times this proportion,

or 78.5 per cent. of the full-term young from control mat-

ings, survived as vigorous healthy individuals. The 13

living F, animals are all rather weak and degenerate and

almost completely sterile according to a considerable

number of careful matings with strong, fertile guinea

pigs. The alcoholic race seems at this stage of the ex-

periment about to fade out in the fourth generation, while

normal control lines from the same original stocks have

passed far beyond this generation, continuing to breed

normally and showing no signs of degeneracy, and never

in any case giving rise to a grossly deformed animal.

The eleventh line indicates again the very decided ef-

fects transmitted by the descendants of animals which

had suffered a modification of their germ plasm by the

alcoholization of their tissues. In 33 cases F, and F,

animals were paired together. Fifteen of these matings

gave negative results or early abortions, while about 12

per cent. of the matings resulted in stillborn litters of two

young each. Only 14 living litters were produced by the

33 matings; these contained in all 23 young, only 7 of

which survived. Thus from a total of 31 full-term young

only 7, or about 22 per cent., were capable of surviving.

All of these young animals are nervous and weak and

several offspring from these bi

When F, animals are mated with normal individuals

No. 590] TRANSMISSION OF DEGENERACY 85

the results are very little if any improved over the two

above combinations. Seventeen such matings gave only

three failures or early abortions, but a high proportion,

23 per cent., of stillborn litters arose, while 10 living lit-

ters, consisting of only 12 individuals, were born. In all

20 full-term young were born and only about one in three

of them survived. In this experiment, although one mate

was a normal animal, the F, mate carried germ cells of

so inferior a quality that the output of the combination,

admitting the numbers are small, leaves no doubt of the

transmission, through three generations, of defective con-

ditions induced by alcoholizing the great grandparents of

the offspring on only one side of the family, or in only

one of the parental lines.

The last line of the table gives the records of mixed

combinations of F, and F, individuals, and here the data

are closely similar to those obtained from other combina-

tions of these animals; only about 25 per cent. of the full-

term young born are capable of surviving, while 78.5 per

cent. of the control young are living.

Briefly, then, 571 matings tabulated in Table I, the rec-

ords to July 1, 1915, have given rise to 682 full-term

young, as well as a large number of premature abortions.

A careful study of all these young animals extending

over a period of five years has afforded data which con-

vincingly show that the treatment of either the male or the

female guinea pig with fumes of alcohol affects the qual-

ity of the offspring to which these animals give rise even

when paired with normal mates. And further, the changed

quality of the offspring is subsequently transmitted

through succeeding generations with even more severe

marks of degeneration and deformity than those exhib-

ited by the offspring of the directly treated animals.

_ Other combinations and back crosses are now in prog-

Tess which are fully in line with the above, but which have

Not yet afforded sufficient analytical data to record. »

____ The defects caused by the alcohol treatment seem to be-

-~ largely confined to the central nervous system and organs

86 THE AMERICAN NATURALIST [Vou. L

of special sense. Paralysis agitans is very common

among the F, F, and F, animals. Paralyzed limbs are

often observed, the animals being unable to stand or

walk (Plates I and II). The eye is also a peculiarly sen-

sitive indicator and presents in the various descendants

of ‘aleoholized individuals all degrees of degeneration—

Fic. 4. Hind feet of No. 488 Fo, 3 Q. All great-grandparents were alcoholic

as well as the maternal grandparents. Inbred from mother by son. This animal

was one of a litter of two stillborn. The left hind fect, C, had only one toe an

the right, D, one toe and a stump,.A and B, normal right and left hind feet.

opaque cornea, cataract or opaque lenses, small defective

eyes, complete absence of one eye and finally complete

absence of both eyeballs—anophthalmie monsters. In

the latter case the extrinsic eye muscles, the third, fourth

and sixth nerves, the lachrymal glands and other struc-

No. 590] TRANSMISSION OF DEGENERACY 87

7 Mia Fig. 4

s% have only one metatarsal and toe, the r and a tarsal jones.”

right foot, B, has the ove oe — and the first metatarsal with the inane

almost complete. © shows rmal skeleton of a right hind limb with the

three toes and seven tarsal rane

88 THE AMERICAN NATURALIST (Vou. L

tures of the orbit are present, though the eyeball is com-

pletely wanting.

Not only are the above congenital eye defects present,

but in several instances members of the alcoholic lines

have become blind during the first year or year and a half

after birth, whereas in our control this has never oc-

curred.

The several illustrations referred to above show speci-

mens exhibiting these various defects. Figs. 1, 2 and 4

of Plates I and II are photographs of animals of indi-

cated lineage which show paralytic conditions. Figs. 4

and 5 illustrate defective extremities. Figs. 1 to 3 show

various degrees of defective eyes and absence of eyeball.

It is peculiarly interesting to find these particular eye

conditions exhibited by the descendants of alcoholized

animals, since, as Stockard (’10) has previously shown,

closely similar eye conditions are obtained in great num-

bers by directly treating the eggs of fish with solutions

of aleohol; and like conditions were also obtained, though

not so Sinataten ty: by treating hens’ eggs (714) with al-

cohol fumes either before or during incubation.

The table just considered gives only a general idea of

the experiment and is in no way analytical. We shall

now attempt to analyze these data in such a manner as to

determine the influence of internal factors, as, for exam-

ple, inbreeding on the results. The influence of the size of

the litter on the quality of the offspring. The behavior of

F, and F, individuals derived from different lines, and

whether there is a difference in the effects on male and

female animals, and the manner of transmission of these

effects.

(To be continued.)

FECUNDITY IN THE DOMESTIC FOWL AND THE

SELECTION PROBLEM?

DR. RAYMOND PEARL

In the December number of the American NATURALIST

Professor W. E. Castle? directs a vigorous attack against

the present writer’s work on fecundity. Any one reading

Professor Castle’s article could scarcely fail, I think, to

carry away the impression that the whole of the writer’s

studies of the past eight years on fecundity in the domes-

tic fowl are to be regarded as essentially valueless. I

assume that it was not the intention to convey this impres-

sion. The fact, however, appears to be as here stated.

With such a conclusion I can scarcely be expected to agree.

I shall therefore attempt, in the following pages, in the

first place, to call attention to some points regarding my

own work which Professor Castle appears to have over-

looked, and which seem calculated to give it at least some

slight degree of significance, and in the second place, to

set forth very briefly my reasons for venturing, in the

present state of knowledge, to hold a different opinion

from his in regard to some phases of the selection problem.

The general plan of Professor Castle’s paper appears

to be to make a comparison between his selection experi-

ments with rats, and my selection experiments with poul-

try, to the very great disadvantage of the latter. To

this general comparison no general comment on my part

can be made, except assent to Castle’s conclusion that his

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

Periment Station, No. 94

2 Castle, W. E., ‘Some Experiments in Mass Selection,” AMER. NAT- - o

_ URALIS?, Vol. XLIX, pp. 713-737, 1915.

l 89

90 THE AMERICAN NATURALIST [Vou. L

work on the selection problem is vastly superior to my

own. Since the subject of such comparison has been

opened it gives me great pleasure to pay tribute, in all

sincerity, to Professor Castle’s splendid series of experi-

ments on selection in rats. In respect of the numbers of

animals involved and their superior adaptability for such

an experiment, his work with rats altogether transcends

anything which has been done with fowls. These selec-

tion experiments constitute an achievement of which their

author may well be proud. I have ventured to disagree

with Professor Castle’s interpretation of the results for

reasons which will presently be stated. But this differ-

ence of opinion, I would most strongly emphasize, con-

cerns only the interpretation. We are at one in our high

admiration of the factual basis afforded by the rat experi-

ments.

Granting all this, however, it seems to me that possibly

the case against my studies of fecundity in toto is not quite

so bad as Castle makes it out to be. Let us examine his

points seriatim. In the first place the strictures upon the

character egg production on p. 714 seem to me to overdo

the matter a bit. It is of course true that it is a character

confined in its expression to one sex, though that it is also

a character which is transmitted by the other sex even

Castle somewhat grudgingly admits (p. 715). It also isa

character which comes to expression only in the adult.

Of this Castle makes a great point throughout his paper,

emphasizing that this means that only a small proportion

of all offspring born can take part in selection experi-

ments. From the standpoint of methodology this point

has nothing like the significance which Castle attributes

to it, for the very simple reason that in all breeding ex-

periments, his own included, there is a vast amount of

random sampling between the population of parental

genes and the population of offspring somata. When

Professor Castle breeds a pair of rats only a very few

No.590] FECUNDITY IN THE DOMESTIC FOWL 91

sperm and ova out of the vast hordes the parents produce

take part in the production of the resulting litter. He

operates, of course, upon the basis that those germ cells

which do take part in the formation of the litter constitute

a random sample of the whole population. When I put

pullets into the house to test their egg production I oper-

ate on precisely the same basis, viz., that I have a random

sample of the family from which they are taken. As a

matter of fact, I have been at great pains to ensure that

the sampling should be random. In all of my studies on

the inheritance of fecundity I have regarded this as a

point of paramount importance, and have never made

use (except occasionally for confirmation of points already

made out on other material) of families in which I had

not either tested all the daughters as to egg production

or a sufficiently large random sample to be fairly represen-

tative of the family. Further I have repeatedly made

careful ad hoc investigations of the adequacy and random-

ness of my sampling.

Castle’s next point is, as the matter stands, apparently

well taken. He quotes (p. 715) a statement which I some-

what rashly made to the effect ‘‘that phenotypic varia-

tion of the character fecundity in fowls, markedly tran-

scends, in extent and degree, genotypic variation.’’ Pro-

fessor Castle’s treatment of this statement is perfectly

legitimate. If it were true, as stated, it would admit of

being turned around as it is in Castle’s next sentence, and

then it surely would be silly to talk about either selection

for this character or about its Mendelian inheritance.

What I should have said when I wrote that unfortunate

sentence, but did not, was that phenotypic variation may

transcend genotypic in fecundity, not that it always or

regularly does. Because it may I wanted to point out the

need for great care in respect of environmental conditions

in interpreting results with this character. The real point

1s this: Long experience in working with winter egg pro-

_ duction in poultry has convinced me that under properly

= Controlled environmental conditions this character is as

92 THE AMERICAN NATURALIST [Vou. L

definitely and regularly controlled by hereditary factors

as is the plumage color and pattern. On the other hand,

it is a character which is rather particularly sensitive to

environmental influences in one direction, namely, down-

ward. I can breed a flock of birds which I know will be

high winter layers if properly fed, housed and managed.

But if these birds are starved, housed in a damp cold

place and otherwise maltreated they will lay but a few if

any eggs. Under such conditions the genotypic condi-

tion would be swamped by the environment. It was this

sort of thing I had in mind when I made the statement

that Castle quotes. It should be particularly noted, how-

ever, that this is a somewhat one-sided matter. I can

(because I have done so) breed a flock of pullets lacking

totally the factors for winter production. With such

birds nothing can be done in the way of feeding or man-

agement which will make them lay before some time in

February or March when the spring cycle begins.

Now all my work on fecundity has been done in a public

institution. Egg production is a commercially valuable

thing. We have had to submit the results of our breeding

operations, in the shape of the birds themselves, to the

practical test of farmers, poultrymen, ete. In doing this

there has always been vividly before my mind the fact that

unless the birds were given proper feed and care, no

matter what the genes they carried, they would not lay

many eggs.

On the other hand the degree of expression of the char-

acter in birds carrying the factors for high fecundity may

be favorably influenced by exceptionally favorable circum-

stances, though the possible effects in this direction are,

according to my observations, much smaller in amount

than in the opposite direction.

While Castle’s comments on the unfortunate sentence

under discussion are technically perfectly legitimate, I do

not think he is quite fair to the essential underlying point

of genetic epistemology, namely, the impossibility of judg-

ing the genetic constitution by the somatic appearance.

No. 590] FECUNDITY IN THE DOMESTIC FOWL 93

This of course is the reason for the progeny test. I do

not think I am in any sense exaggerating if I say that it is

one of the chief results of the Mendelian method of study-

ing inheritance to show that in many cases and for many

characters it is impossible, in the absence of a progeny

test, to be sure of the genetic constitution of the individual

from an examination of the soma alone. I fancy that if I

cared to be fussily nasty in my controversial methods I

could cite page after page from Professor Castle’s Men-

delian writings where even he, in order to be quite sure

about the genetic constitution of an individual, has had to

breed it. This is all I mean by the progeny test. Why

am I and my fowls held up to scorn and ridicule because

I say that it is frequently impossible to tell the genetic

constitution of a fowl with respect to fecundity without

breeding it? Surely fecundity in poultry and coat color

in rats only differ in this respect in degree, if they differ

at all, not in kind. TI think if any one will read pp. 604 and

605 of my last Narurazisr paper, which is the immediate

objective of Professor Castle’s attack, he will have to

admit that the interpretation which I give of the earlier

results is not entirely senseless, and might indeed explain

them. In any case, it is in thorough accord, methodolog-

ically considered, with the very best current Mendelian

usage, including that of Professor Castle himself.

This brings us to the most serious phase of Castle’s at-

tack, namely that in which he denies the validity of my

conclusions respecting the inheritance of the character

fecundity in fowls.

On the top of p. 716 he asserts that I ‘‘assume’’ that

two Mendelizing factors are concerned in the inheritance

of fecundity, ‘‘but without any sufficient published evi-

dence for either conclusion.’? As I have published? many

pages of evidence in demonstration of my conclusions on

this point, one can only infer from this statement of

In particular in the Journal of Experimental Zoology, Vol, 13, 1912. oe

94 THE AMERICAN NATURALIST [Von. L

Castle’s that he regards that evidence as totally worth-

less. It has not so appealed to other workers. Further-

more I think it can be shown that methodologically my

treatment of the problem of inheritance of fecundity

stands on precisely the same plane as Mendelian work in

general, and Professor Castle’s Mendelian work in par-

ticular. This I shall now try to do.

The essence of a test of a Mendelian hypothesis lies in

this: the genetic constitution of the parents of an array

of offspring necessitates that the individual offspring

bearing different segregating characters, or different

segregating categories of the same character, shall occur

in definite numerical proportions. If the observed nu-

merical proportions of the offspring agree, within the

limits of error due to random sampling, with the propor-

tions expected from the Mendelian hypothesis, then this

fact constitutes valid evidence in support of the hypoth-

esis. If no exceptions to this rule appear and a sufficient

number of agreeing cases are adduced the hypothesis is

regarded as demonstrated. The number of cases neces-

sary to constitute a proof is a purely individual matter.

What one person will consider sufficient to establish proof

another will not.

Now in the case of fecundity in fowls, Pearl and Sur-

face® first established that the Barred Plymouth Rock

stock at the Maine Experiment Station was not homo-

zygous in respect of winter egg production, but that it

contained, with frequent occurrence, individuals of high

fecundity, and also individuals of low fecundity. The

race not being homozygous with respect to fecundity, it

was possible to test the Mendelian inheritance of this

+ Cf., for example, Morgan, T. H., ‘‘ Heredity and Sex,’’ New York, 1913,

Doncaster, L., ‘‘The Determination of Sex,’’ Cambridge, 1914, and Jo-

hannsen, W., ‘‘Elemente der exakten Erblichkeitslehre,’? Zweite Ausgabe,

Jena, 1913, Plate L, Vererbungslehre, Leipzig, 1913, Brown, E., ‘‘ Poultry

Husbandry,’’ London, 1915, Sturges, T. W., ‘‘The Poultry Manual,’’ 3d

edition, London 1915.

5 Pearl and Surface, ‘‘Data.on the Inheritance of Fecundity Obtained

from the Records of Egg Production of the Daughters of ‘200-Egg’ Hens,’’

Me. Agr. Exp. Sta. Bull. 166, 1909,

No. 590] FECUNDITY IN THE DOMESTIC FOWL 95

character within the race, without crossing, by the above

scheme.

The next step was the definition of the. categories of

the character winter egg production. From long study

of the character I concluded that the natural categories

in this strain were (a) zero winter production, (b) winter

production between zero and 30 eggs, and (c) a winter

production of over 30 eggs., These were chosen as work-

ing categories. If any one will turn to p. 719 of Profes-

sor Castle’s paper and examine Fig. 1, which is there

printed, they will find that even he chooses categories of

the character with which he is working. Nowhere have

these ever been quantitatively defined; nowhere has he

ever presented any evidence that the step from his rat

grade+1 (for example) to his grade + 2 represents a

more or less inclusive category than a difference in winter

production of from 0 to 30 eggs. Professor Castle reads

us a beautiful little homily about Mendel’s peas. But I

am not clear that either Mendel or Castle has shown that

the amount of variation within the category ‘‘yellow”’ is

less than the amount of variation within my fecundity

category of ‘‘under 30.’’ From the only study which has

ever been made of the matter, Weldon’s,® I should cer-

tainly conclude that the category ‘‘under 30’’ in winter

egg production carries within itself distinctly less varia-

tion than the category ‘‘yellowness’’ in peas. Castle’s

assertion about my fecundity categories ill becomes one

whose work in genetics has dealt almost without a single

exception with non-quantitatively defined Mendelian cate-

gories. Of course, as a matter of fact, he knows, I know,

and everybody knows that the variations within the Men-

delizing category are of no significance so far as the Men-

delian result is concerned. I happen to have observed,

for example, that there are at least four genetically dis-

tinct rose combs in poultry. Yet they are all rose; any

Of them crossed with single gives a3: 1 ratio in F}.

ê Weldon, W. F, R., ‘‘Mendel’s Laws of Alternative Tahoritanes in

oa Biometrika, Vol. I, pp. 228-265, 1902.

96 THE AMERICAN NATURALIST [Vou. L +

Having chosen these categories of the character fecun-

dity because they appeared to represent natural divisions,

I proceeded to show for hundreds of matings the distribu-

tion of the progeny when individual females whose per-

formance fell into one or another of the categories were

mated to particular males. This was done both for the

pure bred Barred Rocks and for crosses. The results

at once showed that definite ratios were appearing with

regularity and constancy. Further analysis showed that

a Mendelian hypothesis which postulated two factors, one

sex-linked and the other not, accounted for all the facts.

If all this does not conform to the classic canons of Men-

delian experimentation, I am sure I do not know what does.

Castle charges me with suppressing datą. There are

just two things which I wish to say regarding this charge.

The first is that I shall publish the complete raw data of

my work on the inheritance of fecundity when I have fin-

ished my own study of these, and not sooner. I am using

this material for the study of various problems. There

appears to be no reason why I should make valuable orig-

inal records public property until such time as I have fin-

ished my own analysis of them. If Professor Castle will

examine my published papers he will find that in lines of

work which I am finishing and leaving, complete raw data

are published (cf. for example ‘‘A Biometrical Study of

Egg Production in the Domestic Fowl,’’ Parts I to III).

In the second place I wish to say that so far as any

question of concealment is concerned Professor Castle, or

any of his students, will be very welcome to come to the

laboratory at any time, for as long as they like, and make

any examination of the original record books in connec-

tion with published results and conclusions.

There is one further point which needs consideration

concerning the charge of suppression of pertinent facts.

An important reason, I think, why Professor Castle’s own

interpretation of his rat selection experiment has not been

No.590] FECUNDITY IN THE DOMESTIC FOWL 97

freely and universally accepted by workers in genetics

lies in the fact that he has never presented his results in

such a form that any other interpretation of the data could

by any chance be tested. There is, from the methodolog-

ical standpoint, only one way in which an adequate test

can be made as to whether any observed change in the

composition of a population is the result of a sorting, or

of true germinal change, or an adequate idea gained of

how the change came about. This is the method of indi-

vidual pedigree analysis. Only one extensive mass selec-

tion experiment has ever been analyzed in this way, and

that is in Surface’s’ discussion of the Illinois corn results.

The Hagedoorns® called Castle’s attention two years ago

to the necessity of individual pedigrees before any just

opinion could be formed as to the meaning of the data.

To paraphase Castle’s damning indictment of the present

writer I may be permitted to call attention to the fact that,

so far as concerns the individual pedigree of his rats, ‘‘1

formation is denied us’’ by Castle.

In bringing to a close this part of the discussion I wish

to emphasize that, in spite of Castle’s assertion to the con-

trary, any unprejudiced person who will take the trouble

to examine the facts will find that, so far as concerns

methods of dealing with the data and presenting them for

publication, the method of their Mendelian analysis, the

method of presenting the results of selection experiments

by a series of averages, and other matters of method, my

work with fecundity in fowls exactly parallels at every

point Castle’s work with hooded rats, and is in every way,

so far as I am able to judge, exactly as critical as his.

His experiments are more extensive in scope than mine,

and the character fecundity is a more difficult one to deal

with, but so far as methodology is concerned the two re-

Searches stand on precisely the same footing. I have not

T Surface, F. M., ‘‘The Result of Selecting Fluctuating Variations. ’’

Data from the Tlinois Corn Breeding Experiments. IV° Conf. int. de Gen.,

PP ne ks

A bL and A. ©., ‘Studies on Variation and Selection,’’

os Zetec rs ind, Abst.—und Vererbungslehre, Bd. XI. pp. 145-83, 1914.

98 THE AMERICAN NATURALIST [Vou. L

lumped the data any more, nor have I ‘‘suppressed’’ data

any more than he has. On the contrary I have published

a great deal of exact data, in a series of papers from this

laboratory, regarding the character fecundity, its normal

variation, ete.

The next point which Castle makes is that the changes

which occurred in mean flock production during the six-

teen years, for which figures were given in the paper which

he criticizes, were probably due to environmental, or at

least to non-genetic effects. In making this point he

calmly disregards all that I have ever published about the

experiments, the means taken to be sure that environ-

mental effects were not mistaken for genetic, ete., and

proceeds in his discussion as though all my work on the

subject had been absolutely uncritical and that I had never

given a thought to checking the correctness of the results.

In the first place he notes the changes in the numbers of

birds on which the average in different years are based,

and points out that these numbers change in a roughly in-

verse direction to the means. He then says:

Has not the better environment and lessened competition of small

numbers possibly something to do with the result?

They have not. Had Professor Castle been less eager to

demolish these fecundity results he might have noted that

I have repeatedly stated that since 1908 all birds in these

experiments have been kept in flocks of the same size,

namely 125 birds per flock. The number of such flocks

has at times varied, but not the number in each flock® ex-

cept by very small numbers, such as resulted from losses

by death, the necessity occasionally of putting a few extra

birds in a pen for a brief period and similar very minor

® To prevent any mental strain in reconciling the above statement with the

third column of Table I, p. 599, in my NATURALIST paper, let me hasten to

say that the pens were filled out, if the number of Barred Rocks in the

selection experiments did not just equal multiples of 125, with birds from

other experiments.

No.590] FECUNDITY IN THE DOMESTIC FOWL 99

fluctuations. In the first four years (1899-1900, 1900-

1901, 1901-1902, 1902-1903) of the old experiment the

birds were kept in 50-bird flocks. During the five years

following (i. e., to 1908-1909) they were kept in 50, 100,

and 150 bird flocks. Just precisely how much (or really

how little) difference the size of flock made in average egg

production has been fully and minutely analyzed bio-

metrically and published by Pearl and Surface!’ some six

years ago. It seems reasonable to suggest that before

indulging in fast and loose criticism on such a simple

point of fact as this it would become Professor Castle to

read the literature respecting the work he is attacking.

Since this material seems to have been forgotten it may

be well to repeat here that the results showed (Pearl and

Surface loc. cit, p. 115) that in general there was no sig-

nificant difference in winter production between 50, 100,

and 150 bird flocks. In later months of the laying year

differences appeared but only in the last month of the win-

ter period (February) was there any significant excess of

even 50-bird flocks over the others. Furthermore, besides

the material which has already been published regarding

the possible influence of environmental factors on the re-

sults of these experiments, I have carried out a number of

special investigations on different phases of this general

question which have not yet been published. For example,

Thave minutely analyzed the data regarding date of hatch-

ing to see whether that might not enter as a significant

factor in the interpretation of the results. The data on

this question are being prepared for publication now, but

1t may be said in advance that the results show that date

of hatching can not possibly have had anything to do with

the rise in average flock production which has occurred

between 1908 and 1915.

10 Pearl and Surface, ‘‘A Biometrical Study of Egg Production in the

Domestic Fowl. Part I. Variation in Annual Production. U.S. Dept. of ee

Agr., Bu. A. I. Bull. 110, pp. 1-80, 1909. Also the effect of flock size upos —

SNe production is specifically discussed in detail in Part II, of the same

Study,’’ pp. 113-117, 1911. - , Ce ee

100 THE AMERICAN NATURALIST [ Vou. L

Vil

Turning now to the general problem of selection there

are certain fundamental matters which it seems to me are

in danger of being lost sight of in the rapid shiftings of

view point which are an essential part of any general con-

troversial campaign, such as Professor Castle’s writings

of the last few years would indicate that he engaged in.

These are:

1. The pure-line concept has certainly been one of the

most useful working tools in the practical breeding of

plants and animals that has ever appeared. Particularly

in plant breeding the pioneer work at Svalöf, which has

been repeated and duplicated on a most extensive scale in

plant breeding laboratories all over the world, demon-

strates in the most complete manner that, whatever may be

happening in the germ-plasm of rats, certainly the germ-

plasm of our common cereal crops is in such a state or

condition that selection within the pure line is without

effect. This is a fact, real and definite. It lies definitely

at the basis of very extensive commercial seed breeding

operations in various different countries. To any one

familiar with the extent and stability of the practical ap-

plications of the pure-line concept in cereal breeding op-

erations, some of our current discussions of the selection

problem seem very academic indeed. Even the justly

celebrated magnitude of Castle’s rat experiments is

scarcely of the same order as the combined and accordant

experience of expert cereal breeders throughout the world.

Before any one makes up his mind finally about the prob-

lem of the efficiency of selection within the pure line it

will be well to remember that besides Johannsen’s famous,

if now in certain quarters somewhat distrusted, beans,

there are all the Svalöf oats, wheats, ete., to be reckoned

with. :

2. No one has ever disputed the power of systematic

selection to alter populations, which were not pure-lines.

Such alteration may extend the range of variation very

greatly beyond what it was in the original population.

No.590] FECUNDITY IN THE DOMESTIC FOWL 101

From a methodological standpoint, however, it is neces-

sary to have a very different sort of evidence from that

afforded by changing general population means, such as

Castle gives for his rats, and I for fecundity, to prove

that the process of selection has been the cause of a change

in the absolute somatic equivalent of a particular gene or

hereditary determinant.

3. It is just in connection with this last point that there

seems to me to be a good deal of unclear thinking and ar-

guing at cross-purposes about the selection problem. Let

us examine the logic of the matter symbolically.

Let there be a character A, whose somatic variation in

the general population is given by a frequency distribu-

tion of area Z 34" A, where Z is the frequency of occur-

rence of the somatic state or condition A,, and so on to Zn

and An. Now suppose that selection is practised for the

somatic condition A,,, but that in the original population

Áss is the most extreme variation in that direction found

to exist. Then for A,,, Z,,—0, and for Ags, Zss is very

small. Let it be further supposed that the somatic differ-

ence between the 4,, and A, condition may be of any de-

terminate magnitude R. It makes no difference to the

logic of the case whether R is large or is extremely minute.

Now suppose, as a limiting case, that we assume a gamete-

soma correlation of 1, i. e., perfect. Then in the gonads

of an individual somatically As, all the germ cells will

bear the factor ass. If two such individuals are bred to-

gether the progeny will be somatically A,s.11 Suppose

that for m generations the matings are of Ass X Azs. This

is continued selection. Then suppose in the m + 1th

generation, ew Ass X Áss parents, appears an As, indi-

vidual.

Concretely this represents a step in advance in the di-

rection of selection. Let us analyze the possible ways in

which this may have happened.

11 This is precisely the condition which prevails in a pure line of oats,

al for purely phenotypic vartion, BEET by environmental

102 THE AMERICAN NATURALIST [Vou. L

(a) First we may assume that 4,,, and A;., the par-

ents of this Ass,

gametes. This would correspond to what is called a mu-

tation. The gamete-soma correlation has been broken by

the appearance of a new kind of gamete different from

the parental gametes. There has been a sudden definite

change in the germ plasm, such that an azs germ plasm

has changed to an dz, germ plasm.

(b) Or, we may assume that Az»,,,, was produced by

the union of two as gametes, but that these gametes de-

develop a 39 soma instead of a38 soma. This assumption

leads logically straight to genetic indeterminism, a con-

clusion which, I think, is repugnant to all that is known

regarding the physiology of the hereditary process.

Embracing alternative (a) then, we may next inquire as

to the possible cause of this sudden change of the germ

plasm, by an amount of which the somatic equivalent is R,

from ds, tO a,» If we say that this change has been

caused by a selection, we can only conclude that the fact

that Áss, Ags, . . - Ags have been placed in particular

cages or apartments to breed, for this is the only physical

thing that selection means in this case, has been the cause

of the germinal change. For by hypothesis there has

been no mixing of germ-plasms. We have been prac-

tising straight selection of the most extreme somatic in-

dividuals, all by hypothesis As, and each homozygous.

It seems to me a misuse of terms to say in such a case as

that postulated that selection has caused the appearance

of the variation which it selects, unless we are prepared to

say that the physical act of the selection of the individuals

for mating physiologically effects the germ plasm. Such

an assumption we are all agreed would be nonsense.

What has happened in the postulated case is precisely

this: a new heritable variation in the direction of selection

has appeared while selection was in progress. If we say

any more than this we are going beyond our facts. If

the selectionist would state his results in this form, and

instead of having a,, gametes had azs

No.590] FECUNDITY IN THE DOMESTIC FOWL 103

not incessantly harp on the string that ‘‘selection caused”

his results, he would be on logically solid ground and

would receive a more respectful hearing from those who

place a high value upon clear thinking and sound logic in

scientific matters.

Now up to this point in the argument there has been no

biological point involved, so far as I can see, to which any-

body, whether of the pure line or the selection faith can

take exception. Certainly I am perfectly willing to ad-

mit that germ-plasm changes do sometimes occur, of all

magnitudes from the most minute up. Further no one,

I take it, will deny that, having appeared, these variations

may be seized upon and preserved by selection. I do de-

sire to emphasize, however, that there is no evidence, as

yet, that the selection causes the variations.

It may be objected that the postulated case is too simple

and leaves out of account too many factors. All this,

however, will not affect the logic of the case. General-

ized, that logic is as follows: A heritable difference be-

tween two individuals or races implies a difference in the

germ plasm. The difference in the germ plasm must have

made its initial appearance at a definite point of time.

At that time the germ plasm changed from its previous

condition. The cause of that change can not be conceived

to be the selection for breeding purposes of the parents

bearing the unchanged germ plasm. To assert that the

new variation is a result of amphimixis due to mating un-

like parents would be, in the present state of genetic

knowledge, a ridiculous begging of the question, because,

in the first place, by hypothesis in any selection experi-

ment individuals genetically as nearly alike as possible.

are always mated together, and in the second place, as se-

lection continues homozygosity automatically increases.

The whole fact of the matter is that the assertion that

Selection per se causes changes in the germ plasm, is a

wholly new addition to the classic Darwinian selection

theory, tacked on quite inadvertently, I believe, by some

of the modern exponents of that theory. Darwin never —

104 THE AMERICAN NATURALIST [Vou. L

supposed that selection was a cause of favorable varia-

tion. Instead he repeatedly pointed out that the funda-

mental problem behind natural selection was that of the

cause of the variations which selection preserved. That

problem remains to-day practically in the same condition

that it was left by Darwin. We are no nearer, essen-

tially, now than we were then to knowing the cause of new

variations. The assertion that new variations are caused

by selection is the rankest kind of mysticism plus bad

logic.

But if selection of the parents can not be supposed

the cause of new variations in the individual, then clearly

what selection does, and all it can do, is to change the

germinal constitution of a race or population by preserv-

ing those individuals in which new variations have ap-

peared, and multiplying them. This is exactly what has

been done in the hooded rat experiment, it seems to me on

Castle’s interpretation of the case. In that experiment

every favorable variation in the many thousands of rats

has been preserved and the individuals bearing it have

been multiplied. Others have been thrown away. The

range of the character in the direction of selection has

been extended far beyond the original range. But would

it have been so extended, or could it have been, if the favor-

able variations had not appeared for selection, or if, hav-

ing appeared, they had not been heritable? Suppose one

started such an experiment with a character which was in

a stable condition and not varying. Take, for example,

the single comb of fowls, and attempt by selection from

a pure single-combed race to produce a stable rose-combed

race by selection alone. Prophecy is dangerous business,

but I do fancy one would be a very long time on that

job! Characters, so far as I can see, will be altered fol-

lowing selection just in proportion as they are varying

genotypically. The cause of the alteration is to be sought

in the cause of the variations, not in the selection only.

I have for some time felt that probably the differences

in opinion between the selectionists, as represented by

No.590] FECUNDITY IN THE DOMESTIC FOWL 105

Castle, and the advocates of the pure-line concept, reduces

itself finally very largely to a dispute over the use of

words, if both are discussing the same objective facts or

experiments. It is repugnant to the logical faculties of

the pure-linists to be told that selection is a cause of new

variations. On the other hand, I suspect that this par-

ticular use of words, which is offensive to our,camp, would

not be deemed absolutely essential to the making of their

case by Castle and his followers. Castle’s special béte

noir appears to be that the pure-linists seem to him to

deny the possibility of germinal variation, except it be

large in amount (a proper De Vriesian mutation). Now

I am in no wise authorized to speak for the pure-line ad-

vocates, but I can say for myself, and I venture to think

others would agree, that this contention forms no part of

the real, genuine pure-line body of doctrine. The fol-

lowers of the pure-line merely have observed in fact that

it is not so easy to change all things by a process of selec-

tive breeding as it has been to change the pattern of

Castle’s rats, or the egg production of my fowls. Many

characters, and many organisms, when got into a homo-

zygous condition exhibit any germinal variation so rarely

as to make any change by the selection of such variation

Impossible within the limits of finite experimentation.

Neither Johannsen nor any followers of his, so far as I

am aware, have ever attempted to set any limitations on

how big or how little a germinal variation could be.

THE EVOLUTION OF THE CELL. II

By tHe Lare PROFESSOR E. A. MINCHIN, F.R.S.

Even more remarkable than the relation of the chromo-

somes to cell-reproduction is their behavior in relation to

sexual phenomena. In the life-cycles of Metazoa the

sexual act consists of thé fusion of male and female pro-

nuclei, each containing a definite and specific number of

chromosomes, the same number usually, though not al-

ways, in each pronucleus. It has been established in

many cases, and it is perhaps universally true, that in the

act of fertilization the male and female chromosomes re-

main perfectly distinct and separate in the synkaryon or

nucleus formed by the union of the two pronuclei, and,

moreover, that they continue to maintain and to propa-

gate their distinct individuality in every subsequent cell-

generation of the multicellular organism produced as a

result of the sexual act. In this way, every cell of the

body contains in its nucleus distinct chromatinic elements

which are derived from both male and female parents and

which maintain unimpaired their distinct and specific in-

dividuality through the entire life-cycle. This distinct-

ness is apparent at least in the germ-cell-cycle of the or-

ganism, but may be obscured by secondary changes in the

nuclei of the specialized tissue-cells.

Only in the very last stage of the life-cycle do the group

of male and female chromosomes modify their behavior in

a most striking manner. In the final generation of oogo-

nia or spermatogonia, from which arise the oocytes and

spermatocytes which in their turn produce the gamete-

cells, it is observed that the male and female chromosomes

make a last appearance in their full number, and then

fuse in pairs, so as to reduce the number of chromosomes

to half that previously present.

106

No. 590] THE EVOLUTION OF THE CELL 107

In Aggregata also Dobell and Jameson have shown that

the union of the pronuclei in fertilization brings together

two sets each of six chromosomes, and that these then fuse

with one another in pairs according to type, that is to

say a with a, b with b, c with c, and so on. Analogous

phenomena have been demonstrated also in the gregarine

Diplospora. We have here a difference in detail, as com-

pared with the Metazoa, in that the fusion takes place at

the fertilization and not as the first step in the maturation

of the germ-cells; but in both cases alike the fusion of

chromatin-elements individually distinct and exhibiting

specific characteristics is to be regarded as the final con-

summation of the sexual act, though long deferred in the

Metazoan life-cycle.

As Vejdovsky has pointed out, there can be no more

striking evidence of the specific individuality of the

chromosomes than their fusion or copulation in relation

to the sexual act. Is there any other constant element or

constituent of living organisms exhibiting to anything like

the same degree the essentially vital characteristics of in-

dividuality manifested in specific behavior? If there is,

it remains to be discovered.

I come now to the question of the permanence and im-

mortality, in the biological sense of the word, of the

chromatinic particles, which may be summarily stated as

follows: the chromatinie particles are the only constit-

uents of the cell which maintain persistently and uninter-

ruptedly their existence throughout the whole life-cycle of

living organisms universally.

I hope I shall not be misunderstood when I enunciate

this apparently sweeping and breathless generalization.

I am perfectly aware that in the life-cycle of any given

Species of organism there may be many cell-constituents

besides the chromatin-particles that are propagated con-

tinuously through the whole life-cycle; but cell-elements

which appear as constant parts of the organization of the

cell throughout the life-cycle in one type of organism may

be wanting altogether in other types. With the exception

108 THE AMERICAN NATURALIST [Vou. L

of the chromatin-particles there is no cell-constituent that

can be claimed to persist throughout the life-cycles of or-

ganisms universally. To take some concrete examples;

the cytoplasmic grains known as mitochondria or chron-

driosomes have been asserted to be persistent elements

throughout the germ-cycle of Metazoa, and the function

of being the bearers of hereditary tendencies has been as-

scribed to them. But Vejdovsky?? flatly denies the al-

leged continuity in cases investigated by him, and though

chrondriosomes have been described in some Protozoa,

there is no evidence whatever that they are of universal

occurrence in Protista. Centrosomes, intranuclear or ex-

tranuclear, have been stated to be constant cell-compo-

nents in some organisms; whether that is true or not it

seems quite certain that in many organisms the cells are

entirely without centrosomic bodies of any kind, as for

example in the whole group of Phanerogams. So it is

with any other cell-constituent that can be named. It may

be that this is only the result of our incomplete knowledge

at the present time. I am prepared, however, to chal-

lenge anyone to name or to discover any cell-constituent,

other than the chromatinic particles, which are present

throughout the life-cycle, not merely of some particular

organism, but of organisms universally.

In this feature of continuity the chromatin-constituents

of the cell present a remarkable analogy with the germ-

plasm of Metazoa. Just as the germ-cells of Metazoa go

on in an uninterrupted, potentially everlasting series of

cell-generations, throwing off, as it were, at each sexual

crisis a soma which is doomed to but a limited lease of

life, during which it furnishes a nutritive environment for

further generations of germ-cells; so in the series of cell-

generations themselves, whether in the germ-cell-cycles

of Metazoa or in the life-ceycles of Protista the chromatin-

particles maintain an uninterrupted propagative series

within a cell-body of which the various parts have a limited

duration of existence, making their appearance, flourish-

22 L. c., pp. 77-89.

No. 590] THE EVOLUTION OF THE CELL 109

ing for a time, and disappearing again. This analogy be-

tween the chromatin of cells and the germ-plasm of multi-

cellular organisms becomes still more marked when we

find that in many Protozoa the chromatin may undergo a

“necialization into generative and trophic chromatin, the

former destined to persist from one life-cycle to another,

the latter destined to control cell-activities merely during

one cycle, without persisting into the next. The differen-

tiation of generative and trophic chromatin is now well

known to occur in many Protozoa, and in its most extreme

form, as seen in the Infusoria, it is expressed in occur-

rence of two distinct nuclei in the cell-body.

To recapitulate my argument in the briefest form; the

chromatinic constituents of the cell contrast with all the

other constituents in at least three points: physiological

predominance, especially in constructive metabolism ; spe-

cific individualization; and permanence in the sense of

potential biological immortality. Any of these threé

points, taken by itself, is sufficient to confer a peculiar dis-

tinction to say the least, on the chromatin-bodies; but

taken in combination they appear to me to furnish over-

whelming evidence for regarding the chromatin-elements

as the primary and essential constituents of living or-

ganisms, and as representing that part of a living body

of any kind which can be followed by the imagination, in

the reverse direction of the propagative series, back to the

very starting-point of the evolution of living beings.

In the attempt to form an idea as to what the earliest

type of living being was like, in the first place, and as to

how the earliest steps in its evolution and differentiation

came about, in the second place, we have to exercise the

constructive faculty of the imagination guided by such few

data as we possess. It is not to be expected, therefore,

that agreement can be hoped for in such speculations; it

would indeed be very undesirable, in the interests of

science, that there should be no conflict of opinion in

theories which, by their very nature, are beyond any pos-

sibility of direct verification at the present time. The

110 THE AMERICAN NATURALIST [ Von. L

views put forward by any man do but represent the visions

conjured up by his imagination, based upon the slender

foundation of his personal knowledge, more or less lim-

ited, or intuition, more or less fallacious, of an infinite

world of natural phenomena. Consequently such views

may be expected to diverge as widely as do temperaments.

If, therefore, I venture upon such speculations, I do so

with a sense of personal responsibility and as one wishing

to stimulate discussion rather than to lay down dogma.

To me, therefore, the train of argument that I have set

forth with regard to the nature of the chromatinic constit-

uents of living organisms appears to lead to the conclusion

that the earliest living beings were minute, possibly ultra-

microscopic particles which were of the nature of chroma-

tin. How far the application of the term chromatin to the

hypothetical primordial form of life is justified from the

point of view of substance, that is to say in a biochemical

sense, must be left uncertain. In using the term chro-

matin I must be understood to do so in a strictly biological

Sense, meaning thereby that these earliest living things

were biological units or individuals which were the ances-

tors, in a continuous propagative series, of the chromatinic

grains and particles known to us at the present day as

universally-occurring constituents of living organisms.

Such a conception postulates no fixity of chemical nature;

on the contrary, it implies that as substance the primitive

chromatin was highly inconstant, infinitely variable, and

capable of specific differentiation in many divergent di-

rections.

For these hypothetical primitive organisms we may use

Mereschkowsky’s term biococci. They must have been

free-living organisms capable of building up their living

bodies by synthesis of simple chemical compounds. We

have as yet no evidence of the existence of biococci at the

present time as free-living organisms; the nearest ap-

proach to any such type of living being seems to be fur-

nished by the organisms known collectively as Chlamy-

dozoa, which up to the present have been found to occur

No. 590] THE EVOLUTION OF THE CELL 111

only as pathogenic parasites. In view, however, of the

minuteness and invisibility of these organisms, it is clear

that they could attract attention only by the effects they

produce in their environment. Consequently the human

mind is most likely to become aware in the first instance

of those forms which are the cause of disturbance in the

human body. If free-living forms of biococci exist, as is

very possible and even probable, it is evident that very

delicate and accurate methods of investigation would be

required to detect their presence.

I am well aware that the nature and even the existence

of the so-called Chlamydozoa is uncertain at the present

time, and I desire to exercise great caution in basing any

arguments upon them. In the descriptions given of them,

however, there are some points which, if correctly stated,

seem to me of great importance. They are alleged to ap-

pear as minute dots, on the borderline of microscopic visi-

bility or beyond it; they are capable of growth, so that a

given species may be larger or smaller at different times;

their bodies stain with the ordinary chromatin-stains; and

they are stated to reproduce themselves by a process of

binary fission in which the body becomes dumbbell-shaped,

appearing as two dots connected by a slender thread,

which is drawn out until it snaps across and then the

broken halves of the thread are retracted into the daugh-

ter-bodies. This mode of division, strongly reminiscent

of that seen in centrioles, appears to me to permit of cer-

tain important conclusions with regard to the nature of

these bodies; namely, that the minute dot of substance has

no firm limiting membrane on the surface and that it is of

a viscid or semi-fluid consistence. |

If it be permissible to draw conclusions with regard to

the nature of the hypothetical biococci from the somewhat

dubious, but concrete data funished by the Chlamydozoa,

the following tentative statements may be postulated con-

cerning them. They were (or are) minute organisms, —

each a speck or globule of a substance similar in its reac-

= tions to chromatin. Their substance could be described

112 THE AMERICAN NATURALIST [Vou L

as homogeneous with greater approach to accuracy than in

the case of any other living organism, but it is clear that

no living body that is carrying on constructive and de-

structive metabolism could remain for a moment perfectly

homogeneous or constant in chemical composition. Their

bodies were not limited by a rigid envelope or capsule.

Reproduction was affected by binary fission, the body di-

viding into two with a dumbbell-shaped figure. Their

mode of life was vegetative, that is to say, they reacted

upon their environmental medium by means of ferments

secreted by their own body-substance. The earliest forms

must have possessed the power of building up their pro-

tein-molecules from the simplest inorganic compounds;

but different types of biococci, characterized each by spe-

cific reactions and idiosyncrasies, must have become dif-

ferentiated very rapidly in the process of evolution and

adaptation to divergent conditions of life.

Consideration of the existing types and forms of living

organisms shows that from the primitive biococcal type

the evolution of living things must have diverged in at

least two principal directions. Two new types of or-

ganisms arose, one of which continued to specialize fur-

ther in the vegetative mode of life, in all its innumerable

variations, characteristic of the biococci, while the other

type developed an entirely new habit of life, namely a

predatory existence. I will consider these two types sep-

arately.

(1) In the vegetative type the first step was that the

body became surrounded by a rigid envelope. Thus came

into existence the bacterial type of organism, the simplest

form of which would be a Micrococcus, a minute globule

of chromatin surrounded by a firm envelope. From this

familiar type an infinity of forms arises by processes of

divergent evolution and adaptation. With increase in

size of the body the number of chromatin-grains within

the envelope increase in number, and are then seen to be

imbedded in a ground-substance which is similar to cyto-

plasm, apparently, and may contain non-chromatinic en-

No.590] ` THE EVOLUTION OF THE CELL 113

closures. With still further increase of size the chroma-

tin-grains also increase in number and may take on vari-

ous types of arrangement in clumps, spherical masses,

rodlets, filaments straight or twisted in various ways, or

even irregular strands and networks,?* and the cytoplasmic

matrix, if it is correct to call it so, becomes correspond-

ingly increased in quantity. I will not attempt, however,

to follow up the evolution of the bacterial type further,

nor to discuss what other types of living organisms may be

affiliated with it, as I have no claims to an expert knowl-

edge of these organisms. I prefer to leave to competent

bacteriologists and botanists the problem of the relation-

ships and phylogeny of the Cyanophycee, Spirochetes,

etc., which have been regarded as having affinities with

Bacteria.

(2) In the evolution from the biococeus of the pred-

atory type of organism, the data at our disposal appear

to me to indicate very clearly the nature of the changes

that took place, as well as the final result of these changes,

but leave us in the dark with regard to some of the actual

details of the process. The chief event was the forma-

tion, round the biococci of an enveloping matrix of proto-

plasm for which the term periplasm (Lankester) is most

suitable. The periplasm was an extension of the living

substance which was distinct in its constitution and prop-

erties from the original chromatinic substance of the bio-

coccus. The newly-formed matrix was probably from the

first a semi-fluid substance of alveolar structure and pos-

sessed two important capabilities as the result of its phys-

ical structure; it could perform streaming movements of

various kinds, more especially ameboid movement; and

it was able to form vacuoles internally. The final result

23 See especially Dobell, ‘‘Contributions to the Cytology of the Bac- -

teria,’’ Quart. Journ. Mier. Science, LVI (1911), pp. 461, 462. I can not

follow Dobell in applying the term ‘ ‘nuclei’’ to these various arrangements

of the chromatin-grains in Bacteria. Vejdovsky compares them with chro-

= Mosomes; but there is no evidence that they play the part in the division

and distribution of the chromatin-grains which is the special function of

chromosomes, as will be discussed-in more detail presently.

114 THE AMERICAN NATURALIST [ Vou. L

of these changes was a new type of organism which, com-

pared with the original biococci, was of considerable size,

and consisted of a droplet of alveolar, amceboid periplasm

in which were imbedded a number of biococci. "Whether

this periplasm made its first appearance around single

individual biococci, or whether it was from the first asso-

ciated with the formation of zooglea-like colonies of bio-

cocci, must be left an open question.

Thus arose in the beginning the brand of Cain, the

prototype of the animal, that is to say, a class of organism,

which was no longer able to build up its substance from

inorganic materials in the former peaceful manner, but

which nourished itself by capturing, devouring, and di-

gesting other living organisms. The streaming move-

ments of the periplasm enabled it to flow round and en-

gulf other creatures; the vacuole-formation in the peri-

plasm enabled it to digest and absorb the substance of its

prey by the help of ferments secreted by the biococci.

By means of these ferments the ingested organisms were

killed and utilized as food, their substance being first

broken down into simpler chemical constituents and then

built up again into the protein-substances composing the

body of the captor.

A stage of evolution is now reached which I propose to

call the pseudo-moneral or cytodal stage, since the place

of these organisms in the general evolution of life corre-

sponds very nearly to Haeckel’s conception of the Monera

as a stage in the evolution of organisms, though not at all

to his notions with regard to their composition and struc-

ture. The bodies of these organisms did not consist of

a homogeneous albuminous ‘‘plasson,’’ but of a periplasm

corresponding to the cytoplasm of the cell, containing a

number of biococci or chromatin-grains. Thus their

composition corresponded more clearly to that of plasson

as conceived by Van Beneden, when he wrote: ‘Si un noyau

vient à disparaître dans une cellule, si la cellule redevient

un cytode, les éléments chimiques du noyau et du nucléole

s’étant repandus dans le protoplasme, le plasson se trouve

No. 590] THE EVOLUTION OF THE CELL 115

de nouveau constituté.’ If we delete from this sentence

the word ‘‘chimiques’’ and also the words ‘‘et du nu-

eléole,’? and substitute for the notion of the chemical so-

lution of the chromatin-substance that of scattered chro-

matin-grains in the periplasm, we have the picture of the

eytodal stage of evolution such as I have imagined it. It

should be borne in mind that the ultimate granules of

chromatin are probably in many cases ultra-microscopic ;

consequently they might appear to be dissolved in this

cytoplasm when really existing as discrete particles.

In the life-cycles of Protozoa, especially of Rhizopods,

it is not at all infrequent to find developmental phases

which reproduce exactly the picture of the pseudo-mon-

eral stage of evolution, phases in which the nucleus or

nuclei have disappeared, having broken up into a number

of chromatin-grains or chromidia scattered through the

cytoplasm. We do not know as yet of any Protozoa, how-

ever, which remain permanently in the cytodal stage, that

is to say, in which the chromatin-grains remain perma-

nently in the scattered chromidial condition, without ever

being concentrated and organized into true nuclei; but it

is quite possible that some of the primitive organisms

known as Proteomyxa will be found to exhibit this condi-

tion and to represent persistent Pseudo-monera or

eytodes.

The next stage in evolution was the organization of the

chromatin-grains (biococci) into a definite cell-nucleus.

This is a process which can be observed actually taking

place in many Protozoa in which ‘‘secondary”’ nuclei

arise from chromidia. It seems not unreasonable to sup-

pose that a detailed study of the manner in which second-

ary nuclei are formed in Protozoa will furnish us with a

picture, or rather series of pictures, of the method in |

which the cell-nucleus arose in phylogeny. To judge from

the data supplied by actual observation, the evolution of

the nucleus, though uniform in principle, was sufficiently —

- diversified in the details of the process. As one extreme

We have the formation of a dense clump of small, separate : : -

116 THE AMERICAN NATURALIST [VoL b

chromatin-grains, producing a granular nucleus of the

type seen in Dinoflagellates, in Hæmogregarines, and in

Diatoms. Amongst the chromatin-grains there may be

present also one or more grains or masses of plastin

forming true nucleoli. At the opposite extreme a clump

of chromatin-grains becomes firmly welded together into

a single mass in which the individual grains can no longer

be distinguished, forming a so-called karyosome, consist-

ing of a basis of plastin cementing or imbedding the

chromatin-grains into a mass of homogeneous appear-

ance. Whatever the type of nucleus formed, the concen-

tration of the chromidia into nuclei does not necessarily

involve all the chromidia, many of which may remain free

in the cytoplasm.

In the chromidial condition the chromatin-grains scat-

tered in the cytoplasm are lodged at the nodes of the alvė-

olar framework.24 Consequently a supporting framework

of cytoplasmic origin, the foundation of the linin-frame-

work, was probably a primary constituent of the cell-

nucleus from the first. In many nuclei of the karyoso-

matic type it is very difficult to make out anything of the

nature of a framework, which, however, in other cases is

seen clearly as delicate strands radiating from the kary-

osome to the wall of the vacuole in which the karyosome

is suspended. Probably such a framework is present in

all cases, and each supporting strand is to be interpreted

as the optical section of the partition between two proto-

plasmic alveoli.

With the formation of the nucleus the cytode or pseudo-

moneral stage has become a true cell of the simplest type,

for which I propose the term protocyte. It is now the

starting-point of an infinite series of further complica-

_tions and elaborations in many directions. It is clearly

24 Cf. Dobell, ‘‘ Observations on the Life-History of Cienkowski’s Arach-

nula,’’ Arch. Protistemkunde, XXXI (1913), p. 322. The author finds that

in Arachnula each nucleus arises from a single chromatin-grain, which grows

to form a vesicular nucleus. Since the fully-formed nucleus contains nu-

merous grains of chromatin, the original chromidiosome must multiply in

this process.

No. 590] THE EVOLUTION OF THE CELL 117

impossible that I should do more than attempt to indicate

in the most summary manner the various modifications

of the cell-type of organism, since to deal with them con-

scientiously would require a treatise rather than an ad-

dress, and, moreover, many such treatises exist already.

The most conspicuous modifications of cell-structure are

those affecting the periplasm, or, as we may now term it,

the cytoplasm. In the first place, the cell as a whole takes

various forms; primitively a little naked mass of pro-

toplasm tending to assume a spherical form under the

action of surface-tension when at rest, the cell-body may

acquire the most diverse specific forms maintained either

by the production of envelopes or various kinds of exo-

skeletal formations on the exterior of the protoplasmic

body, or of supporting endoskeletal structures formed in

the interior. The simple ameboid streaming movements

become highly modified in various ways or replaced by

special locomotor mechanisms or organs, flagella, cilia,

ete., of various kinds. The internal alveolar cytoplasm

develops fibrille and other structures of the most varied

nature and function, contractile, skeletal, nervous, and so

forth. The vacuole-system may be amplified and differ-

entiated in various ways and the cytoplasm acquires man-

ifold powers of internal or external secretion. And finally

the cytoplasm contains enclosures of the most varied kind,

Some of them metaplastic products of the anabolic or

catabolic activity essential to the maintenance of life,

others of the nature of special cell-organs performing

definite functions, such as centrosomes, plastids, chro-

matophores, etc., of various kinds.

With all the diverse modifications of the cytoplasmic

cell-body the nucleus remains comparatively uniform. It

may indeed vary infinitely in details of structure, but in

Principle it remains a concentration or aggregation of nu-

merous grains of chromatin supported on some sort of

framework over which the grains are scattered or

= ¢lumped in various ways, supplemented usually by plastin =

_ nucleolar substance either as a cementing ground-sub-

118 THE AMERICAN NATURALIST [Vou. L

stance or as discrete grains, and the whole marked off

sharply from the surrounding cytoplasm, with or without

a definite limiting membrane. There is, however, one

point in which the nucleus exhibits a progressive evolu-

tion of the most important kind. I refer to the gradual

elaboration and perfection of the reproductive mechan-

ism, the process whereby, when the cell reproduces itself

by fission, the chromatin-elements are distributed between

the two daughter-cells.

The chromatin-constituents of the cell are regarded, on

the view maintained here, as a number of minute gran-

ules, each representing a primitive independent living

individual or biococeus. To each such granule must be

attributed the fundamental properties of living organisms

in general; in the first place metabolism, expressed in con-

tinual molecular change, in assimilation and in growth,

with consequent reproduction; in the second place specific

individuality. As the result of the first of these proper-

ties the chromatin-granules, often perhaps ultra-micro-

scopic, may be larger or smaller at different times, and

they multiply by dividing each into two daughter-gran-

ules. As a result of the second property, chromatin-

granules in one and the same cell may exhibit qualitative

differences and may diverge widely from one another in

their reactions and effects on the vital activities of the

cell. The chromatin-granules may be either in the form

of scattered chromidia or lodged in a definite nucleus.

When in the former condition, I have proposed the term

chromidiosome*® for the ultimate chromatinic individual

unit; on the other hand, the term chromiole is commonly

in use for the minute chromatin-grains of the nucleus.

The terms chromidiosome and chromiole distinguish

merely between the situation in the cell, extranuclear or

intranuclear, of the individual chromatin-grain or bio-

coccus.

25 ‘Introduction of the Study of the Protozoa,’’ Arnold, 1912, p. 65.

SHORTER ARTICLES AND DISCUSSION

INHERITANCE OF CONGENITAL CATARACT

CaTARACcT is the opacity of the eye caused by a faulty forma-

tion of the lens. Certain forms of cataract are congenital and

hereditary. Other forms which appear later in life may either

be hereditary or due to pathological causes. ,

In the normal eye the delicate fibers which go to make up the

lens are glued together along their sides and at their ends where

they unite in lines radiating from the poles of the lens to form

a completely transparent body. Anything which prevents the

perfect conjunction of these fibers causes a defect in the trans-

parency of the lens. This imperfection has been compared by

Harman (2) to the white spots in the finger nail, caused by slight

injuries to the nail bed, and he has shown it to be correlated with

faulty formation of the dental enamel.

There are various causes for the inhibition of proper lens de-

velopment, and these give rise to different forms of cataract.

Only those forms have been considered here which are congenital.

The most common form of congenital cataract is the lamellar,

perinuclear or zonular cataract. This manifests itself as a dark

circular disk with the density increasing from the center to the

perimeter, forming characteristic zones. These zones are flecked

by small wedge-shaped dashes arranged regularly in a spoke-

fashion about the disk. The disk is located between the nucleus

of the lens and the cortex; and is caused by a thickening of the

layers at that place.

Discoid cataract is a slight form of the lamellar, less than 4

mm. in diameter, and located at the posterior pole of the lens.

The opacity is uniform throughout, but is not easily visible.

(It is sometimes confused with anterior polar cataract, of which

the origin is not definitely known, but which is not congenital.)

Coralliform or axial cataract, cataracta fusiformis, is an

Opaque line running through the lens from anterior to posterior

pole with a spindle-shaped swelling towards the center of the

lens

Anterior and posterior cortical cataract, cataracta corticallis, —

Where the opacity takes a more or less —— gntiine, ye

119 a

120 THE AMERICAN NATURALIST [Von. L

cataracta punctata, formed by minute white dots scattered uni-

formly through the lens or grouped in the anterior cortical

layers, and other forms of circumscribed, stationary, lenticular

opacities which though rare are known to be congenital (3), have

also been used in the compilations given here.

Senile cataract, which also seems to be hereditary, and those

forms of cataract arising from lesions, diseases of the eyeball, and

certain general diseases such as cholera and tetany, have been

omitted.

Although congenital and other forms of cataract have long

been considered by the medical profession to be influenced by

heredity, no definite analysis was made until 1905, when the first

paper by Nettleship (1) appeared. Nettleship’s data have been

the basis of Bateson’s (9) conclusion that the abnormality is in-

herited as a dominant character. Bateson acknowledges that

normal parents have produced abnormal children, but these

cases he explains as either origin de novo, or due to faulty clas-

sification of the parents, who in reality may have been slightly.

affected with cataract.

Davenport (3) has followed Bateson’s conclusion in regard

to the inheritance of cataract, and makes the eugenie recommen-

dation that unaffected parents from affected stock may marry

without fear of producing abnormal children.

In the ‘Treasury of Human Inheritance’? Harman (2) gives

one hundred genealogical tables dealing with congenital cat-

aract. Each table represents two or more generations with a

detailed account of the condition of each individual in regard to

congenital defects of the eye. The data used in this paper have

been taken from these tables. Only those families are used in

which there is no doubt as to the condition of the parents or

the children in respect to the abnormality, and where there is

no question as to the total number of children in each family.

After discarding all the doubtful cases, and picking a sibship

with its parents from the table as a family, there is left a total

of one hundred and twenty-five families which are classified into

three different categories, as follows: (A) both parents normal

with at least one abnormal child; (B) one parent normal, the

other affected with some form of congenital cataract, with at

least one abnormal child; (C) both parents abnormal, giving

only abnormal children.

There are 31 families with both parents normal which give

some abnormal children. In a total of 153 children from these

No.590] SHORTER ARTICLES AND DISCUSSION 121

families, 61 are affected with cataract. This suggests that the

character is more likely to be inherited as a recessive than as a

dominant. Surely it is not possible to explain so many cases as

origin de novo or as due to faulty classification of the parents.

In the second category given above (B), where one of the

parents is affected and the other normal, the number of defec-

- tive children would be expected to be approximately the same

whether the character was inherited as a dominant or a recessive.

If the abnormality is considered as a recessive character, the

ratio of 61 affected to 92 unaffected, already spoken of as having

been obtained in the first category of families, shows an excess

of recessives over the simple Mendelian expectancy for a mono-

hybrid. This is to be expected since the criterion for including

any family in the tabulation is the production of at least one

abnormal child. In families with a small number of children

it is probable that in some cases only normal children are pro-

duced in matings of heterozygote by heterozygote which should

give, on the average, one fourth recessive. The observed re-

sults must then be compared to a modified Mendelian ratio which

will allow for the omission of all-normal progenies. Such ratios

have been calculated by Apert (4) and by Wright (5). The

expected proportions given here are calculated according to the

method given by the latter. na

The proportion of recessives varies according to the number

of children in the family and ranges, for a three-to-one ratio,

from 100 per cent. in families with one child to very nearly 25

per cent. in families with fifteen children. The proportion is

calculated from the formula

<= 4 — Gy’

where N is the number of children. Since the criterion for in-

cluding any family is the production of one abnormal child,

all families with one child must have 100 per cent. abnormal chil-

en. The proportion decreases, according to the law of chance,

as the number of children in the family increases, finally reach-

ing 25 per cent. as the number of children becomes large.

Table I compares the results obtained with the theoretical ex-

pectancy, worked out according to this method. — |

The method used for testing the agreement of the observed

result with the theoretical is the one given by Pearson (6) and

= Elderton (7). It was originally used to test various series of bio-

122 THE AMERICAN NATURALIST [Vot L

TABLE I

NORMAL X NORMAL (BOTH HETEROzYGOUS— NN X NN)

| | |

| | | L C d Ob; ed

un TL Propria | Seah a (hanes of ono O

N | Families | on Children | se mea wees ee

1 2 2 1.0000 | 2.00 2 060

2 4 8 5714 4,57 6 1.43 447

3 6 18 4324 7.78 13 5,22 3.502

4 5 20 3657 Tol 8 69 0

5 0 0 3278 0 0 000

6 7 42 3041 yargi 19 6.23 3.039

7o 1 T .2885 ve — .02 000+

8 2 16 2778 44 5 56 .070

9 1 9 2703 2.43 1 —1.43 841

10 2 20 2649 5.30 3 2.30 .998

11 ] 1] 2610 2.87 2 — .87 .263

31 153 51.49 61 | 9.225

Eeee £1

Fea 4s

Per Cent.

Normal Abnormal Abnormal

Observed (io 75.2 oe, 92 61 40.

Calenited 250) en na 102 51 33

Diferenta 5 hoe. “10 10 T

logical measurements. Attention has been called to the applica-

tion of this method of testing theoretical Mendelian ratios by

Harris (8).

By calculation from the data given in Table I the measure of

agreement, or “‘P,”” is .418. ‘‘P’’ is a value ranging from 0 to

1, proportional to the closeness with which the observed facts

agree with the theoretical. In this case in four times out of ten,

random samplings of similar data would give results deviating

more widely from the theoretical. A possible explanation for

this rather wide discrepancy will be given later.

The see from matings of normal by abnormal are given

in Table I

A total r 448 children from 90 families is used. 232 children

out of the 448 are found to be defective, whereas 238 are ex-

pected according to the modified Mendelian ratio.

As before, the criterion for including any family in the tabu-

lation is the production of at least one affected child. Only in

this way can the matings of heterozygous -na Nn (nor-

mal), with homozygous recessives, nn (ab 1), be distin-

No.590] SHORTER ARTICLES AND DISCUSSION 123

TABLE II

NORMAL X ABNORMAL (NN X NN).

Family | of No. of Proportion Number of | Number of 0-—C ood

N Families | Children Recessive Recessives | Recessives C

| | x C 0

Dee 9 | 1.0000 9.00 9 00 | .000

2i S| 22 .6667 14.67 20 5.33 | 1.936

ets 10 30 .5714 17.14 19 1.86 .201

4 12 48 | .5333 25.60 24 —1.60 .100

5 10 50 .5161 25.81 27 1.19 054

6 8 48 508 24.38 23 —1.38 078

7 11 ai 5039 38.80 40 20 037

8 9 72 5019 36.14 31 5.14 731

9 3 27 5009 13.52 15 1.4 162

10 | 3 30 500 15.02 9 —6.02 | 2.412

oo 2 22 .5002 11.00 10 —1.00

12 0 0 .5001 .00 0

13 1 13 5001 6.50 5 —1.50 | .346

| 90 448 237.58 232 6.147

Noo ie.

= .862

g Per Cent.

Normal Abnormal Abnormal

ecrvel .. 2.40.2 ae 216 232 52

Cetsateted: <5... aa 210 238 53

Differences 2.2.60... 52 6 a 1

guished from the matings of homozygous dominants, NN, with

recessives. In the former matings approximately 50 per cent.

of the children are expected to be abnormal. In the second

case only normal children are expected, who should all be hetero-

zygous for the abnormality. It is entirely possible in small

families that matings which should give part abnormal offspring

might give only normal children; but all these matings are ex-

cluded, for no distinction can be made between them and the

more usual matings of abnormal with homozygous normal, also

giving only normal children. :

Having thus excluded part of the data, the modified ratio 1s

calculated as before, except that in this case it applies to a one-

to-one ratio, instead of a one-to-three. It ranges from 100 p

cent. in families with one child, to very nearly 50 per cent. in

families with fifteen children, and is calculated from the formula

X= 9 = 0T ee.

The agreement of observation with expectancy is eit "e o

124 THE AMERICAN NATURALIST [ Vou. L

“P” having a value .862 means that, in nearly nine cases out of

ten, other random samplings would deviate more widely from

the theoretical.

The critical test as to whether or not congenital cataract can

be considered as a simple recessive character lies in the matings

of abnormal by abnormal. Families of this kind should have

only abnormal children. Only three such matings are avail-

able. In two of these, five and two children, respectively, are

the total numbers produced; and these are all abnormal. The

other case is a doubtful one. Both parents are classified as hav-

ing discoid cataract; one is given as seriously affected, the other

only slightly so. Seven children are given for this mating: two

are abnormal, and the others apparently normal. Three of these

five died in infancy, so that their classification is doubtful, but

there is no question as to the others. Assuming, as Bateson does

in his cases, that there has been a faulty classification of the

parents, that the parent given as slightly affected is not con-

genitally defective, but adventitiously so early in life; then this

one discrepancy might be conveniently overlooked.

Another explanation may be found, however, in the fact that

heterozygous individuals sometimes show the recessive character.

Cases are known where a small percentage of heterozygous indi-

viduals show the recessive character, although the homozygous

and heterozygous dominants are generally indistinguishable.

If such is the case here, the slightly affected parent is hetero-

zygous, and the occurrence of normal children is expected.

This assumption also helps to explain the deviation of observed

results above the theoretical in Table I, and their deviation

below the theoretical in Table II. If heterozygous individuals

are sometimes classified as recessives, it would affect the classi-

fication of both the parents and the children. A few matings

of abnormal (heterozygote showing the recessive character) by

normal would be included in the category B which rightfully

belong in category A. Thus matings giving a one-to-three ratio

would be included among matings giving a one-to-one ratio. This

would tend to reduce the observed results below the expected.

A faulty classification of the children would tend to raise the

results, but this would not be as strong a deviating factor as

when influencing the parents and therefore a number of chil-

dren. Thus the balance would tend slightly toward a decrease

in the regular expectancy ; a result which fits well with that ob-

tained in Table IT.

No.590] SHORTER ARTICLES AND DISCUSSION 125

In the matings of normal by normal, there is, of course, no

opportunity for this error to influence the classification of the

parents, since abnormals, whether heterozygous or not, would

not come here. But with the children, the number with the re-

cessive character would be raised above the regular expectancy ;

a result which coincides with that in Table I.

If with more matings of abnormal by abnormal it is found

that, with a few exceptions, only abnormal children are given,

the evidence that cataract is a recessive character rather than a

dominant will be fairly conclusive. It seems rather strange

that congenital cataract manifesting itself, as it does, in such

different ways, should be determined by a single unit factor.

These things, however, must be explained in the simplest pos-

sible manner ; an attempt to work it out with two or more factors

would introduce great complications, and be practically im-

possible with the data as they have been gathered heretofore.

The fact that a recessive character may not be recognized, for

it occurs in mass data in a greater proportion than would be

expected at first, should be noticed. Finally the approxima-

tion of the results obtained with those expected from the single

unit factor form the best reason for its acceptance.

That certain geneticists should have laid down eugenic rules

based on the inheritance of this character as a dominant is, at

the very least, unfortunate. It is not only because of a mistake

in the method of inheritance, but such rules should never be

made until the exact hereditary processes are positively known,

since such practises are likely, not only to bring discredit upon

the science, but to injure people who endeavor to follow them in

the regulations of their lives.

D. F. Jones,

S. L. Mason.

BUSSEY INSTITUTION,

HARVARD UNIVERSITY.

BIBLIOGRAPHY

1l. Nettleship, E. On Heredity in the Various Forms of Cataract. Rep.

Memoirs, XI, Part IV, Section XIIIa. Dulau & Co. 1910.

3. Davenport, C. B. Heredity in Relation to Eugenics. Holt, 1911.

4. Apert, E. The Laws of Naudin-Mendel. The Journal of Heredity,

Nov., 1914, 492-497. Translation from Eugenique, II, 5, 129 fi.,

Mai, 1914.

126 THE AMERICAN NATURALIST [Vou. L

Wright, Sewall. 1914. Unpublished MS.

Pearson, Karl. Phil. Magazine, July, 1900, Vol. L, 157-175.

Elderton W. P. Biometrika, 1901, Vol. 1, 155-163.

Harris, J. Arthur. American Naturalist, Dec., 1912, 741-745.

Bateson, Wm. Mendelian Principles of Heredity, University Press,

1909.

Be TAS

rar

. Fuchs, Ernst. Text-Book of Ophthalmology, translation by A. Duane,

4th edition.

HUXLEY AS A MUTATIONIST

ELSEWHERE I have pointed out that Galton’ held with equal

firmness to continuity and discontinuity in variation, and that

the American horticulturist and botanist, Thomas Meehan,’ held

clear mutationist conceptions which he supported by accurate ob-

servations of variations in many plants. It seems worth while

to add a note on the attitude of Huxley with regard to this ques-

tion

Whenever Huxley expressed himself on this matter he usually

took occasion to say explicitly that he could see no reason why

variations should not be discontinuous as well as continuous, and

one of the few points on which he differed from Darwin was in

ascribing greater significance to such marked changes. Several

statements of his position in this matter are found in his volume

of essays entitled Darwiniana.

Thus he says (p. 77) :

Mr. Darwin’s position might, we think, have been even stronger than

it is if he had not embarrassed himself with the aphorism “ natura non

facit saltum,” which turns up so often in his pages. We believe, as we

have said above, that Nature does make jumps now and then, and a

recognition of the fact is of no small importance in disposing of many

minor objections to the doctrine of transmutation.

Elsewhere (pp. 34, 404) Huxley refers to the well-known Ancon

sheep, which originated from a single ram in the flock of a Mas-

sachusetts farmer named Seth Wight. The story of this breed

of sheep is told in a letter from Col. David Humphreys to Sir

Joseph Banks, then President of the Royal Society.? The farmer

kept a flock of 15 ewes and one ram on the banks of the Charles

River, at Dover, Mass., 16 miles from Boston. In 1791 a ram

1**Galton and Discontinuity in Variation,’?’ Amer. NAT., 48: 697-699,

1914,

2‘*An Anticipatory Mutationist,’’ AMER. NAT., 49: 645-648,

2 Humphreys, D., 1813, ‘‘On a New Variety in the Breeds of sae? ” Phil.

Trans. Roy. Soc., 1813: 88-95.

No.590] SHORTER ARTICLES AND DISCUSSION 127

lamb was born having a short length of back and short bandy

legs. Seeing an advantage in such an animal owing to its in-

ability to jump fences, it was bred to the flock, the original ram

being killed. The first year thereafter two lambs had the pecu-

liarities of their father, and in following years a number more

Ancon lambs were produced. The latter when bred together

always, with one questionable exception, produced Ancons.

Hence the character was evidently a recessive, having orig-

inated from the normal through a negative variation or mutation,

presumably in one germ cell. This being the case, the variation

must have been carried in a latent or recessive condition for a

certain number of generations until inbreeding brought it out in

a homozygous form. The original ram which was killed must

have been heterozygous for this character, also one at least of the

ewes and probably more; for one such heterozygous ewe was

necessary to produce the original Ancon ram, and the two Ancons

which appeared next year in the back-cross not improbably came

from different mothers. It is therefore impossible to say just how

long this condition may have been handed on in a ‘‘latent’’ con-

dition before inbreeding brought it out.

With few exceptions, the Ancons showed alternative inherit-

ance when crossed with normal sheep, and (l. ce., p. 90).

Frequent instances have happened where common ewes have had

twins by Ancon rams, when one exhibited the complete marks of ru

tures of the ewe; the other of the ram.

Incidentally this shows that such twins came from separate ova.

In a fiock the Ancon sheep tended to keep together and separate

from the normal members of the flock. The breed seems to have

attained some popularity, but their flabby subscapular muscles,

infirm construction, loose joints, crooked forelegs and awkward

gait, while preventing them from jumping fences made them

difficult to drive to market. Butchers also found the carcasses

smaller and less saleable, so that they were soon supplanted after

the introduction of the Merino. They were already scarce in 1813

and afterwards became extinct.

uxley remarks regarding this case:

Varieties then arise we know not why; and it is more than probable

that the majority of varieties have arisen in this “spontaneous ” man-

ner, though we are, of course, far from denying that they may be traced,

in some cases, to distinct external influences. . . . But however they

o _ May have arisen, what especially interests us at pressit is, to remark

128 THE AMERICAN NATURALIST [ Vou. L

that, once in existence, many varieties obey the fundamental law of

reproduction that like tends to produce like; and their offspring ex-

emplify it by tending to exhibit the same deviation from the parental

stock as themselves.

After further discussing the case, Huxley remarks (Op. cit.

p. 39) :

Here, then, is a remarkable and well-established instance, not only of

a very distinet race being established per saltum, but of that race breed-

ing “true” at once, and showing no mixed forms, even when crossed

with another breed.

Réaumur’s case of a Maltese couple having a hexadactylous son,

three of whose four children were again hexadactylous, also

‘comes in for Huxley’s comment (p. 35 ff.). The following dicta

on the subject of variation, from the same volume, are also worth

quoting :

Indeed we have always thought that Mr. Darwin unnecessarily ham-

pered himself by adhering so strictly to his favourite “ Nature non facit

saltum.” We greatly suspect that she does make considerable jumps in

the way of variation now and then, and that these saltations give rise

to some of the gaps which appear to exist in the series of known forms

(p. 97).

I apprehend that the foundation of the theory of natural selection is

the fact that living bodies tend incessantly to vary. This variation is

neither indefinite, nor fortuituous, nor does it take place in all directions,

in the strict sense of these words. . . . A whale does not tend to vary in

the direction of producing feathers, nor a bird in the direction of de-

veloping whalebone (p. 181).

The importance of natural selection will not be impaired even if fur-

ther inquiries should prove that variability is definite, and is determined

in certain directions rather than in others, by conditions inherent in that

which varies. It is quite conceivable that every species tends to pro-

duce varieties of a limited number and kind, and that the effect of

natural selection is to favour the development of some of these, while it

opposes the development of others along their predetermined lines of

modification (p. 223).

From these and similar statements it appears evident that were

Huxley living to-day he could scarcely escape being classed as a

mutationist.

R. RUGGLES Gates

UNIVERSITY OP CALIFORNIA

1916

L, NO, 591

VOL

The American Naturalist

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

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

Sho icles containing summaries of research work bearing on the

[a publica of organic evolution are especially welcome, and will be given preference

e hundrea reprints of bape jane gs are supplied to authors free of charge.

F Ey pote ae my aad supplied at co

Subscri e arira should be sent to the publishers. The

subse soe I is “fou r dollars a year. Foreign postage is fifty cents and

Canadian postage twenty-five cents additional. The charge for single copies is

forty cents. The advertising ye are Four Dollars for a page.

THE SCIENCE PRESS

Lancaster, Pa. Garrison, N. Y.

_ NEW YORK: Sub-Station 84

En d tter, April 2, 1908, at the Post Offi t Lancaster, Pa., under the Act of

Congress of Mareh 3, 1879.

BO OKS KS FOR S ALE | JAPAN NATURAL HISTORY SPECIMENS

Perfect Condition and Lowest Prices.

Specialty: Bird Skins, Oology, Entomology, Marine

NATURAL HISTORY Animals and others. Catalogue free. Correspond-

Catalogue on application ence solicited.

JOHN D. SHERMAN, Jr. | T. FUKAI, Naturalist,

403 Seneca Avenue MOUNT VERNON, N. Y. Konosu, Saitama, Japan

. PUBLISHED BY THE CAMBRIDGE UNIVERSITY PRESS

MIMICRY IN BUTTERFLIES

- - Bound in 1 Buckram. “Royal svo, Pp. viii+188. With 16 E (12 of which

are colored). Price $4.50.

o w Reoreato ORUNDALL Punwerr, F.R.8., Fellow of Gonville and Caius Goias

nE ; Sek eaa hen f Genetics in the University of Cambridge.

aS Re ~ EXTRACT FROM THE PREFACE

a “This little book has been written in the it to several

classes ipia itten Tor mny appa 4

Sn AT r have been asked by friends of

in life to

THE

AMERICAN NATURALIST

VoL. L. March, 1916 No. 591

HT ESIRN AND THE RATE OF EVOLUTION

IN ANGIOSPERMS

Proressor EDWARD C. JEFFREY

HARVARD UNIVERSITY

In responding to an invitation to contribute to the

morning program of the American Society of Naturalists,

it has seemed to me that a statement emphasizing some

of the morphological features of the greatest of all bio-

~ logical problems, the modus operandi of the process of

evolution, would be of interest to my fellow biologists.

The most distinguished as well as the most profound in-

vestigator, which our science has yet produced, Charles

Darwin, has unequivocally expressed the opinion in the

‘‘Origin of Species,’’ that morphology is the soul of 1 Eo

ural history. As Iam addressing a body of men who call

themselves naturalists, my theme will, I ee not apg ear

unimportant.

The rate of evolution has not been

130 THE AMERICAN NATURALIST [Vou. L

have unquestionably been the progressive cooling of our

earth’s surface, as well as those recently recognized sec-

ular periodic twilights of the sun god, known as glacial

periods. The latter have worked in an exterminating

manner and have wiped out well nigh completely whole

types of plants and have left the way clear for the unre-

stricted development of better adapted forms. For ex-

ample, at the end of the Paleozoic, in the late Permian,

we find world-wide evidence of glaciation, which resulted

in the virtual extinction of the great cryptogamic forests,

which contributed the raw materials of our most abundant

coals. With the passing of the arboreal Cryptogams, the

Gymnosperms became the predominant element of the

world forests in the Mesozoic. At the end of the Creta-

ceous there was another age of extinction, which wiped

out the mass of Gymnosperms and particularly the Coni-

fers. The naked seeded plants, which prevailed in the

medieval period of our earth’s history, have in the vege-

tation of to-day been reduced in the number of species to

the merest fraction of seed-producing plants; which in

the present age are overwhelmingly angiospermous.

From the present standpoint, however, the progressive

but not spasmodic cooling of our earth is of even greater

importance. Investigations initiated in my laboratories

have made it clear that herbaceous Angiosperms have

been derived from woody ones as a response to the in-

creasing coldness of terrestrial climates. Plants of this

organization are of such efficiency that they are able to go

from seed to seed in a few weeks and thus pass through

the inclement winter season in a resting stage. The orig-

inal researches in this direction were undertaken by Pro-

fessor Eames. The theme in the past two years has

undergone a profitable exploitation by other former stu-

dents in both botanical and geological publications. The

origin of the herbaceous type in the Angiosperms has in

itself added a notable impetus to the rate of evolution in

the group. Whatever hypothesis one adopts as to the

mode of the origin of species, it is quite clear that the

No. 591] HYBRIDISM IN ANGIOSPERMS 131

multiplication of generations as well as of individuals,

rendered possible by the appearance of the herbaceous

type of small size and short reproductive cycle, will con-

tribute to the acceleration of evolutionary processes.

A noteworthy feature, which distinguishes the huge

aggregation of Angiosperms now inhabiting the surface

of the globe (in the neighborhood of one hundred and

forty thousand species) from the saved remnant of the

Gymnosperms, is their inherent variability. This high

degree of variability has naturally made the Angiosperms

a very difficult group from the systematic standpoint and

has likewise put them in the foreground in connection with

discussions as to the origin of species. Two of the oldest

tribes of the coniferous Gymnosperms are the pines and

and the araucarians. I have had the good fortune to be

able to make a careful comparison of structure extending

to all important details, between living representatives

of these tribes and their predecessors in the Cretaceous

of the eastern United States. It is quite clear from these

studies that the genus Pinus and the genus Araucaria in

the remote times of the Age of Chalk, differed only in the

smallest particulars from their living descendants. The

conclusion inevitably follows that the course of evolution

here has been very slow. The actual situation corre-

sponds accurately with the data derived from the past.

A white pine, compared with an evening primrose or a

rose, is relatively constant and invariable. :

The remarkable variability of the Angiosperms, as fre-

quently expressed in terms of the difficulty of systematic

identification, brings us naturally to the much debated

question of the origin of variability. Darwin, as is well

known, simply accepted this phenomenon as a fact and

did not, after the first, at any rate, attempt to explain the

condition in terms of other phenomena. It is interesting,

however, to note that in the beginning he was disposed to

accept hybridization as the cause of the variability of

species and apparently abandoned this belief only because

he could find no evidence for its occurrence on a

132 THE AMERICAN NATURALIST [Vou. L

ciently extensive scale. Quite recently the view that

heterozygosis is responsible for the mutability of species

has again been advanced by Lotsy in an interesting article

published in the Archives Néerlandaises. This author

very definitely takes the position that variability in gen-

eral is due to hybridization, and that true species (not

necessarily those of Linnæus and other systematists) are

invariable. With this view I am personally in agree-

ment, with the limitation that the statement goes much

too far.

It is one of the commonplaces of breeding that the off-

spring resulting from hybridization is extremely variable

and may be characterized by a greater or less degree of

sterility. Taking the particular case of the Angiosperms,

it is found that when species of lilies, irises, honeysuckles,

ete., are crossed, the result is a highly mutable progeny

with a greater or less degree of sexual sterility, the latter

condition most easily recognized in the microspores or

pollen. The main purpose of the present statement is to

make it clear to my fellow naturalists that in nature a

high degree of variability often exists in the case of the

Angiosperms, expressed either in terms of difficulty of

systematic determination in view of intergrading forms,

or often in the less obtrusive form of multiplication of

species in a given genus. This extreme degree of vari-

ability is very largely accompanied by the highly signifi-

cant phenomenon of pollen sterility.

A family of Angiosperms much in the foreground in

recent years is the Onagracee or Evening Primrose

family. In the case of the genus Œnothera remarkable

conditions have been discovered by De Vries. The plants

of O. lamarckiana, when grown in large numbers, show a

number of individuals, sometimes as high as one twentieth

of the total number, markedly different in character from

typical O. lamarckiana. This phenomenon was at first

thought to be peculiar to this species of Œnothera and a

great deal of importance was consequently attached to

clearing up its somewhat dubious systematic position.

~ No. 591] HYBRIDISM IN ANGIOSPERMS 133

Fortunately we are relieved from the uncertainties neces-

sarily connected with this kind of investigation, by the

discovery in more recent years that other and perhaps all

species of the genus possess the same features. The ac-

tivity of systematic botanists in recent years in making

new species of Œnothera is highly significant in the pres-

ent connection. The exceptional individuals which grow

up in cultures of species of Œnothera have been termed

by De Vries and his disciples ‘‘elementary species.” The

biological world has been asked to believe that in the

appearance of these new forms in cultures of @nothera,

we have the phenomenon of mutation or the origin of

species at a leap. This view of the matter is, however,

open to serious question. The species of (nothera, as

well as their so-called mutants, are distinguished by a

degree of pollen sterility often extreme. This condition

has convinced so accomplished a geneticist as Professor

Bateson that the so-called elementary species of Œno-

thera are segregates resulting from previous hybridiza-

tion. This view of the matter is supported by the fact

that the products of hybridization are often relatively

fixed forms, as indeed has been noted by Brainerd in his

extremely interesting observations on hybrid wild violets.

Obviously the question of possible mutation in the

genus Œnothera entered into a new and biologically more

advantageous phase when other species than O. lamarck-

iana came into the discussion. Clearly a still wider view

should even more clarify the situation. Two years ago

Miss Ruth Holden, who is at present living in Cambridge,

England, made the interesting discovery that the common

fireweed, Epilobium angustifolium, growing wild near

Cambridge and also cultivated in the Cambridge botanic

garden, was characterized by a large degree of sterility

of pollen. She at once generously communicated her dis-

covery to me and at the same time suggested a reason

for the condition of pollen found in the English specimens

of Epilobium angustifolium. I must here remind you

that under the genus Epilobium are included two distinct

134 THE AMERICAN NATURALIST [Vou. L

subgenera, namely Chamenerion, distinguished, among

other features, by its distinct pollen grains; and Epi-

lobium proper having its pollen grains in groups of four.

E. angustifolium belongs to the section Chamenerion, and

in the southern part of Canada and the Northern States

has no allied species except in mountainous regions (e. g.,

mountainous Quebec and Colorado). Acting on the sug-

gestion supplied by Miss Holden’s discovery, Mr. C. A.

Fig. 1. Pollen of hybrid Iris germanica,

Forsaith, one of my graduate students, has investigated

the conditions of sterility found in species of Epilobium

belonging to the section of Chamenerion. Through the

kindness of the Gray Herbarium of Harvard University

he has been able to study some two hundred specimens

from various geographic regions. The conditions in Epi-

lobium (Chamenerion) angustifolium in the northern

part of its range, where it coincides in distribution with

its allied species, Æ. latifolium, are most interesting.

Nearly nine tenths of the specimens showed the pollen to

be imperfect. In contrast, the material from the southern

limits, where E. angustifolium does not coincide in distri-

No. 591] HYBRIDISM IN ANGIOSPERMS 135

bution with E. latifolium, are almost uniformly character-

ized by a high degree of perfection. To be specific, speci-

mens from Ontario, western Quebec, and New Hampshire

and Massachusetts show pollen perfectly developed or at

most with a few grains disorganized. Mr. Forsaith ex-

tended his investigation, again through the courtesy of

the Gray Herbarium, to the other genera and species of

the Onagracez, with similar results. The investigation asa

whole will be described elsewhere, but it will be necessary

Fic, 2, Pollen of Zauschneria californica, a monotypic representative of the

On :

to consider a few more illustrations in the present connec-

tion. There is one quite monotypic species in the order,

namely Zauschneria. It was found that in this the pollen

is practically perfect and the same state of affairs is pres-

ent in the two geographically limited species of Gongylo-

carpus, one occurring in Vera Cruz and the other on the

opposite side of the continent in Lower California. The

general situation in the case of the Onagraceæ, a family

much in the foreground at the present time by reason of

the investigations of De Vries and his disciples, is that

monotypic species or those geographically isolated have

136 THE AMERICAN NATURALIST [Vou. L

perfect pollen and are little characterized by variability ;

while where the species are numerous and coincident in

their range both variability and pollen sterility are con-

spicuous.

We may now consider another highly variable group,

which has not infrequently been called a hybrid family,

namely the Rosaceew. The genera Rosa, Rubus and Ora-

tegus are notable for the extreme difficulty they have

offered from the systematic point of view. Three of my

Fic. 3. Pollen of Chamaenerion (Epilobium) angustifolium from Massachusetts.

graduate students have investigated these genera and the

results may be conveniently summarized by reference to

the genus Rubus. In the case of Rubus, in regions where

it has been exhaustively studied, there is almost no end to

the species which may be set up. In Europe, in fact, the

species have mounted into the thousands. The situation

may for the sake of brevity be considered under three

heads. First, there are species which range together and

have flowering periods which overlap—a condition com-

mon to the mass of our ordinary Rubi. In Rubus villosus,

No. 591] HYBRIDISM IN ANGIOSPERMS 137

the blackbriar, and R. strigosus, the wild red raspberry,

both very variable species, the pollen is extremely bad.

Where these species occur on islands, however, the pollen

is generally much more perfect, probably as the result of

isolation. I have noticed, for example, that R. villosus

and R. strigosus from Cape Breton Island have consider-

ably better pollen than that found in the case of conti-

nental material of the same two species. What is true of

these particular species holds more or less well for a

Fig. 4. Pollen from Chamaenerion (Epilobium) angustifolium from the vicinity

of Cambridge, England, showing abortive grains.

large number of others of similar range and flowering

periods. Next may be considered a species of limited ge-

ographic range, namely R. deliciosus from the Rocky

Mountains. Here the pollen is practically entirely per-

fect, a few defective elements being occasionally found.

Last may be described R. odoratus, the so-called flower-

ing raspberry, which blossoms after the mass of other

species have shed their flowers. Here, as one might ex-

pect, the pollen is highly perfect and practically un-

mingled with shrivelled grains. A general study of the

Rosacæ, which can not even be summarized in the brief

138 THE AMERICAN NATURALIST [Von L

time at my disposal, shows clearly that propinquity, geo-

graphical or phenological, is to a large extent correlated

with pollen imperfection in the group.

Limitations of time make it necessary to proceed sum-

marily with other illustrations. Next may be cited the

Betulacee and Fagacee. Each of these orders has one

strikingly polytypic species, Betula in the one case, and

Quercus in the other. Interestingly enough, it is in these

two genera that variability and gametic sterility coincide.

Fic. 5. Pollen of Rubus deliciosus from the Rocky Mountains, showing well

developed grains.

One might continue at length through the Dicotyledons,

but one other example must suffice for this division of the

Angiosperms. The Solanacew have one huge genus, Sola-

num itself, in which there are nine hundred species. In

this genus not only is there extreme variability, but also

a large degree of pollen sterility. In the monocotyledo-

nous division we may start with the grasses. Monotypic

grasses have perfect pollen, as is illustrated, for example,

by the wild rice, Zizania aquatica. In the genus Alope-

curus, with numerous and propinquitous species, on the

contrary the pollen conditions frequently indicate gen-

No. 591] HYBRIDISM IN ANGIOSPERMS 139

etical contamination. Proceeding to aquatics, in the Po-

tamogetonacex, the monotypic Zannichellia and Zostera

have perfectly developed microspores;while Potamo-

geton, with its numerous species, is often distinguished

by a large degree of pollen imperfection. Similar state-

ments hold in a like sense in regard to members of the

Alismaceæ, Sparganiacee, etc.

Fic. 6. Pollen of Rubus villosus (Blackbriar), showing high degree of im-

rfection.

The pressure of time compels a summing up of the situ-

ation without further references to detailed facts, which

will be supplied by publications soon to appear. The gen-

eral condition in the Angiosperms in contrast to the Gym-

nosperms is a large degree of variability in the species.

Where the species are highly inconstant and cause great

difficulty to the systematist, as, for example, in the Ona-

graceæ, Rosaceæ, Solanaceæ, Birches, Oaks, ete., there is

often a large degree of pollen sterility. Where isolation,

geographical, phenological or specific, is present the con-

tents of the anther sacs are strikingly perfect in their

development. In other words, where interspecific cross-

ing is possible, there is often clear evidence of its presence

140 THE AMERICAN NATURALIST [Vou L

in the form of a high degree of variability, accompanying

a considerable manifestation of sterility in the gametic

cells, particularly the pollen. In the numerous species of

Rosa or @inothera, we find in regard to both variability

and the phenomenon of sterility, a marked contrast to the

also numerous species of the very old genus Pinus. In

Pinus there is practically no imperfection in the develop-

Fig. T. Pollen of Zizania aquatica, a genus with few isolated species.

ment of the microspores, even in exotic species, and the

species are very clearly marked and constant.

If associated variability and gametic sterility are reli-

able indications of hybridization, then it becomes clear

that the Angiosperms, unlike the Gymnosperms and the

mass of the vascular Cryptogams, are often character-

ized by heterozygosis. It has been recently suggested

that pollen imperfection is not so much an evidence of

hybridization as of mutability. This criticism appears to

fail for various reasons. First, for nearly a hundred

years practically all students of hybridization in plants

have noted pollen sterility and imperfect development of

the seed as peculiar characteristics of hybrids. Secondly,

No. 591] HYBRIDISM IN ANGIOSPERMS 141

in genera with often highly sterile species, such as Rubus,

the species which are isolated for any reason from the

rest have either perfect pollen or manifest a much less

marked degree of sterility. An objection urged by De

Vries to gametic degeneracy as a criterion of hybridism

needs apparently only to be stated to supply its own refu-

tation. The distinguished plant physiologist of Amster-

dam, in a recent article in which he criticizes the writer’s

Fie. 8. Pollen of in tenia pratensis, showing high degree of imperfection

hich may occur in a polytypic genus.

attitude in regard to the intimate relation between de-

fective pollen, hybridization and so-called mutation, some-

what superciliously, states that the degeneracy of spores

in connection with the development of the megaspores of

the heterosporous vascular Cryptogams (and one might

add the seed plants as well) might with equal validity be

regarded as evidence of hybridism in the megasporic

sporangia. One has only to carry De Vries’s argument

to its logical conclusion to prove its entire fallacy. Since

microsporic sporangia (in which there is no spore degen-

eration apart from hybridism) and megasporic sporangia

occur ordinarily or at least primitively on the same plant,

142 THE AMERICAN NATURALIST [Vou. L

it follows that so far as the phenomenon of spore degen-

eration is concerned some sporangia (the megasporangia

or seeds) are of hybrid origin and others (the micro-

sporangia or anthers) are not. The logical absurdity of

this conclusion will be clear to every one.

There seems to be no question on the basis of the well-

established criteria of hybridism, that many Angiosperms

present clear indications that their species are of hetero-

zygous origin. Since one of the most efficient methods of

inducing variability in connection with the development

of improved varieties of ‘plants is hybridization, often on

a very large scale, it seems not unreasonable to regard

spontaneous hybridization in the Angiosperms (the evi-

dences for which are so numerous and so impressive) as

having an incaleulably large effect on their rate of evo-

lution. There is, however, apparently no reason for

assuming a similar condition in the Gymnosperms and the

vascular Cryptogams. The great and indeed overwhelm-

ing advantage which the Angiosperms have secured in the

struggle for existence over the lower groups of vascular

plants is apparently connected in an intimate way with

hybridism on the one hand and the development of her-

baceous types (in response to progressive climatic re-

frigeration) on the other. If this conclusion is correct

we must reject the assumption of universal hybridism as.

the sole cause of variation put forward by Lotsy as much

toosweeping. Small variations unquestionably character-

ize the Gymnosperms, and in the course of long geological

time have availed in the absence of competition from

heterozygous types, with a much greater range of vari-

ability and consequently a higher potentiality of evolu-

tion. It is obviously impossible for the homozygous

Conifers to make headway against the characteristically

heterozygous Angiosperms. The small variations of

homozygous stocks clearly prevailed in the earlier history

of our earth, while the more rapid changes which have

ensued in later times are correlated, so far as plants are

concerned, at any rate, with marked physiographic and

No. 591] HYBRIDISM IN ANGIOSPERMS 143

climatic differentiation, and most important of all with

the phenomenon of heterozygosis.

In conclusion the situation may be summarized. The

phenomenon of variation in the older types of plants is

still unexplained and must apparently be accepted as an

ultimate characteristic of living matter. In the case of

those groups of plants, which have achieved predomi-

nance under the present climatic conditions of our earth,

hybridism has clearly played a large rôle in the accelera-

tion of the processes of evolution. The peculiar condi-

tions presented by the species of Ginothera, which have

been put forward by De Vries in favor of his mutation

hypothesis, are obviously only a particular case of the

manifestation of the natural hybridism, which is so wide-

spread a feature of the Angiosperms. The mutation

hypothesis has suffered a process of rapid disintegration

of late and it is increasingly clear on the botanical side

that where the term mutation is used it. ordinarily indi-

cates changes which are the result of previous hybridiza-

tion. Concerning the Animal Kingdom the trend of

opinion is apparently setting equally strongly against

mutation. My zoological colleague, Professor Castle, has

recently declared himself in no uncertain terms against

the hypothesis of mutation, an expression of opinion not

the less convincing because he originally held the view

that mutation was a necessary pendant to Mendelism.

He is now able to explain to himself the appearance of

new characters as a result of the summation of small vari-

ations, which is essentially the Darwinian position.

A FURTHER ANALYSIS OF THE HEREDITARY

TRANSMISSION OF DEGENERACY AND

DEFORMITIES BY THE DESCENDANTS

OF ALCOHOLIZED MAMMAIS. II

CHARLES R. STOCKARD ann GEORGE PAPANICOLAOU

DEPARTMENT OF ANATOMY, CORNELL UNIVFRSITY MEDICAL SCHOOL,

ITY

THE [NFLUENCE OF [NTERNAL AND EXTERNAL FACTORS on

THE QUALITY OF THE OFFSPRING

Table II gives the relationship between the size of the

litters and the mortality of the descendants from differ-

ent combinations. It brings out in a way the variable

internal and external factors to be considered in inter-

preting the conditions of the members of the numerous

litters of animals. The external factor considered in the

table is one of nutrition or environment, depending upon

the number of young developed in the uterus at any one

time. The table indicates the influence of an internal

factor, the germ plasms concerned in mating related or

non-related animals. Four combinations are considered :

pairs of normal non-relatives, pairs of alcoholic non-

relatives, pairs of normal relatives, and pairs of alcoholic

relatives.

The first vertical column shows that in mating together

normal non-related guinea pigs of the stocks used in these

experiments the average litter contains 1.96 individuals.

Fifty-one and eleven hundredths per cent. of the young

were found in litters of two, and 20 per cent. of the ani-

mals occurred in litters of three. Fifteen and fifty-five

hundredths per cent. of the animals were born in litters

of only one young, and 13.33 per cent. in litters of four

individuals.

The next space below in the table shows the number

and percentage of individuals living over three months

144

TABLE II

THE SIZE or LITTERS AND MORTALITY OF DESCENDANTS FROM DIFFERENT COMBINATIONS

Alcoholic Inbred 135

Normal Lines 90 Alcoholic Lines 401 Normal Inbred 30

ii 1 2 3 4 1 2 3 4 1 2 3

in litt. in litt. in litt. inlitt. | in litt. inlitt. in litt. in litt. | in i in a in a

2 214 87 1

stl pa 11%) (20%) (13.33%)|(20. wine (53.86%) (21.69%) (3.99%) Wed (60 bs ) (28

umber of young per litter|(Average number of young per litter eying: number of young per litter

coor 9) a

Lived over 3

= months.. 1 2 3 4 1 2 3 4 in itt in tie: in He

cee in litt. in litt. in litt. in litt. | in litt. in litt. in litt. in litt. 2

40 11 109 19 3 (100%) (66.66%)

? (85.71%) get bods (61.11%) (33.83%) | (64. gate (50.93%) (21.83%) (18.75%) |(All together 80%)

a (All together 74.44%) (All together 46.13%)

Died within

: months .. 2 3 4 1 2 3 4 1 2 3

in litt. inlitt. inlitt. inlitt. | inlitt. inlitt. inlitt. inlitt. | inlitt. in litt. in litt.

(14.28%) a (38.88%) (66. 66%) on 71%) (49.06%) (78.16%) (81.24% (33.33%)

(All together 25.56%) All together 53.86%) (All together 20%)

1 2 3 4 1 2 3 4 1 2 3

in litt. in litt. inlitt. inlitt. | in litt. in litt. inlitt. in litt. | in litt. in litt in litt.

ie, 0 0 0 10 0 0 0

(2.25%) (4.20%) (11.49%)

(All together 5.23%)

eS 2 oe 4 1 2 3 4 1 2 3

in litt. in litt. in litt. in litt. | in litt. in litt. in litt. in litt. | in litt. in litt. in litt.

0 0 0- 0 1 4 2

(1.19%) (1.86%) (3.44%) (12.50%) (5.55%)

(All together 2.49%) (All together 3.33%)

in litt.

in Ai

in litt.

1 2 3 4

in litt. in litt. in litt. in litt.

14%) (56.29%) (15.55%)

ee age number of young per litter

3 4

in lite in Het. in litt. in litt.

29 2 0 0

(76.31%) (26.31%)

(All together 36.29%)

1 2 3 4

in litt. in litt. in litt. in litt.

9 2 0

(23.68%) (73.68%) (100%)

(All together 63.70%)

1 2 3 4

in litt. in litt. in litt. in litt.

B 26%) (18.42%) (19.04%)

All together 14.81%)

1 2 3 4

in litt. in litt. in litt. in litt.

1 0

(2.89%) (5.26%)

(All together 3.70%)

146 THE AMERICAN NATURALIST (Vou. L

in the different-size litters. Almost 86 per cent. of the

individuals born one in a litter lived, and about 87 per

cent., 40 out of 46, of those born two in a litter lived.

When there were three in a litter, however, only 61 per

cent. lived and of those born four in a litter, it happened

that only one third of them survived, though there were

only a few in all. Of the total number of young from

normal non-related parents 74.44 per cent. lived. Judg-

ing from these statistics litters of one or two young are

the most vigorous and individuals born in litters of three

or four are not so likely to be strong and long-lived.

The next space below gives the mortality records,

which, of course, is merely another way of bringing out

the above statements. The space following contains the

number of deformed animals, but from the normal mat-

ings not one such individual has been produced. The

last space gives the number of small-size or dwarf speci-

mens also, none of which occur among these litters from

normal non-related parents.

The second vertical column contains a similar analysis

_ of the influence of the size of the litter on the mortality

and condition of the young born from non-related alco-

holic parentage. This not only includes the offspring

from directly treated animals, but also other matings of

non-relatives belonging to the alcoholic lines. Here

again the majority of all the young, 53.36 per cent., are

born in litters of two. Litters containing three are next

in frequency, followed by litters of only a single individ-

ual. Of the total number of offspring produced by al-

coholic parents 21.69 per cent. occurred in litters of three,

and only about 4 per cent. of the offspring were members

of litters of four individuals. The average number of

young in the litters from these animals is 1.79, somewhat

smaller than from normal matings.

The space below shows that in all only 46.13 per cent.

of these young survived, whereas more than half as many

more, or 74.44 per cent., from normal parentage lived

over three months. The most vigorous animals are those

No. 591] TRANSMISSION OF DEGENERACY 147

born only onein a litter. Sixty-four and twenty-eight hun-

dredths per cent. of them lived. While about 51 per cent.:

of the two-in-a-litter individuals survived, only about 22:

per cent. of the young born three in a litter were capable

of surviving, and only 18.75 per cent. of the individuals,

from the litters of four lived more than three months..

These figures indicate that the offspring from similarly.

injured parents are more capable of survival when born,

in a small litter of one or two than when contained in,

larger litters of three or four. ;

This is not on account of the fact that the treated or

degenerate mother is more incapable of nourishing the,

larger litters, since the same is true of the larger litters:

from normal mothers, as shown by the previous column. ,

The fact is that all young of large litters tend to be small

and weak at birth, whereas a single young is far better

accommodated. For these reasons it is always of im-

portance to know the size of the litter in which an animal |

was born in estimating the degenerate qualities it may

possess as compared with the qualities of another indi-

vidual. For example, one animal may appear larger and

stronger than another, and yet when bred will give rise:

to more degenerate offspring than the weaker individual.

Although having a vigorous body, its germ-cell complex

was not so good as that of the weaker animal, from a

larger litter which produces better offspring. Therefore, '

the small weak males bred to normal females do not al-

ways give the poorer results when compared with the.

matings of stronger males and normal females.

- The next space is the reverse of the one above and

shows the percentages of mortality among the offspring

derived from alcoholic non-relatives. More than half of

the young, 53.86 per cent., from these combinations die

soon after birth, a mortality record just twice as high as

that of the control animals. :

The next space shows the frequency of deformities

among such young. Here it is again clearly indicated

that the animals born one in a litter are better than those

FIG 521 albino Fə Q (two alcoholie grandmothers, both grandfathers

normal). ina only one day after birth;

- > meninges of the brain were filled

ood. Gross tremor and complete p raais of right side. ‘ataracts, both

crystalline lenses being cas opaque. The photograph shows the iho yes d

paralyzed right extremities while e left legs are held in a normal position in

heir effo o support the body. (Birth weight, hale grams. )

Fic. 2. 506 mouse and

and the aternal

oh Grea tr emor and

e : $ of right eye opaque.

Photograph shows the powerless «

right legs attempt

gs with the

figures are at di

condition of the outs aia p left le

pting to support the grams.) wo

e body. (Birth weight, 57

magnifications,

> g

No. 591] TRANSMISSION OF DEGENERACY 149

from litters of two, which are in turn better than the mem-

bers of the litters of three individuals. Only 2.25 per

cent. of the 84 individuals born in litters of one were de-

formed. While 9 of the 214 individuals born two in a

litter, or 4.2 per cent., almost twice as many, were de-

formed. And 10 of the 87 animals born three in a litter

were deformed, or about 11.5 per cent., which is almost

five times higher than the number of deformities found

among the animals born in litters of single individuals. |

Among the descendants of alcoholic non-relatives there

was in all 5.23 per cent. of deformed specimens, whereas

not one deformed animal arose from similar normal

matings.

The last space of this column indicates the number of

dwarf or undersize animals produced in the different lit-

ters from non-related alcoholic lines. Among the 84 ani-

mals born one in a litter only a single individual was of

unusually small size. The 214 animals born in litters of

two were all of average size except four, or stated exactly,

1.86 per cent. of them were undersize. In the litters of

three 3.44 per cent. of the animals were small, while 12.5

per cent. of the members of litters of four were small

specimens. Here again it is shown that the members of

large litters are not so uniformly up to the standard of

size and vigor as animals born in litters of only one or

two individuals.

The third vertical column gives a similar analysis of

_the few normal inbred individuals which have been pro-

duced during the time of the experiment. There are not

many such matings, as a general effort has been made to

avoid inbreeding the control animals since this might be

considered to vitiate the results.

The few young produced by inbred normal matings have

all been in litters of only one or two offspring, so that

the size of the litters averages only 1.43 individuals. The

size of the litters is, therefore, smaller than from either

the non-related normal or alcoholic animals. Eighty per

cent. of the young have survived, more, however, from the

agouti, yellow and white, normal Fs 4. A normal animal from

generation of the control; slightly inbred, natural size. Birth

ams.

599 black, white and red. Fs 4 degenerate animal ae

alcoholic lines; no inbreodine = the paternal grandm er had

ae the maternal grandfather had both parents ‘alecholic.

: lic reat andmothers and

rs;

e one

zri I $ rere normal

I í no ale sm i aj

(2 (AL Ọ (AA) (NN).

shows the front limbs bent under

the body and the : animal

head. t weighed only ií e

While the above normal

at birth

> time of death,

animal weighed rams, actually

ind all normal animals increase in weight rapidly from

No. 591] TRANSMISSION OF DEGENERACY 151

litters of one than from the litters of two. Not one de-

formed animal has resulted from these normal inbred

matings and only one individual of the thirty was less

than two thirds the average weight. The few normal in-

bred young here considered are then equally as good as

the young from normal non-relatives and necessarily su-

perior to the alcoholic lines. Judging from the results of

others, there is little doubt that a more extensive inbreed-

ing might produce deleterious effects.

The last vertical column indicates the effects of in-

breeding alcoholic animals and their descendants. This

combination shows the:poorest quality offspring found in

the table. Here again the members of the larger litters

are at a disadvantage when compared with those born in

the small litters. The average number of young in a

litter is 1.62, somewhat smaller than the litters produced

by mating alcoholic non-relatives. Thirty-eight litters

contained a single individual each, the same number of

litters contained two individuals, while only seven litters

consisted of three young, and these were the largest litters

produced. The inbred alcoholic animals, therefore, have

a tendency to produce a large proportion of small litters

and this tendency aids in strengthening their offspring. |

Of the 135 young resulting from these matings, only

36.29 per cent. of them survived; this is the poorest life

record shown by any combination. Of those born only

one in a litter, however, 76.31 per cent. survived, which is

a record equal to the average of the control. Therefore,

even in this very bad combination, alcoholic inbreeding,

when only one young is produced at a litter by an animal

ordinarily capable of producing two or more, this one

young is so well nourished and accommodated that it is

somatically vigorous. Yet on breeding such individuals

it almost always happens that very inferior offspring re-

sult. The germ cells, at any rate, may possibly be

stronger than those in the weaker individuals which oc-

curred in litters of two or three. Only 26.31 per cent. of

the animals born in litters.of two were capable of sur-

No. 591] TRANSMISSION OF DEGENERACY 153

viving. The mortality here happens to be about three

times higher than among the single-litter individuals.

And further, not one of the 21 specimens born in litters

of three lived. Among the offspring from the alcoholic

inbred lines, judging from the numbers now available,

the difference between the vitality of individuals born in

litters of one and those born in larger-size litters is most

striking. |

The space below is the reverse of the one just consid-

ered and gives the percentage of young dying in the dif-

ferent litters. Only 23.68 per cent. died from litters of

one individual, while 73.68 per cent. died in the litters of

two individuals, and every one, 100 per cent., of the ani-

mals born in litters of three died within three months and

usually within a few days.

The proportion of deformed animals occurring in the

different-size litters again emphasizes the same differ-

ences in quality. All together 14.81 per cent. of the 135

individuals were grossly deformed; this is the highest

percentage of deformed animals occurring in the several

combinations represented in the table.

Of those animals born one in a litter only 5. 26 per cent.

were deformed; of those born two in a litter 18.42 per

cent., or more than three times as many, were deformed,

and of those specimens born in litters of three 19.04 per

cent., or about one in five of them, were grossly deformed;

The proportion of deformities, therefore, conforms to

the mortality records, being very much higher in the

larger litters, and not unusually high among the individ-

uals born one in a litter as compared with the average

percentage of deformities from alcoholic non-relatives.

Therefore, the bodily quality of the offspring is not mate-

rially worse from alcoholic inbred animals than from

matings of alcoholic non-relatives, provided only one in-

dividual is born in the litter. But when more than one `

individual occurs in the litter, the alcoholic inbred com-

bination is the most disastrous for the vitality and form

of the offspring of all the combinations considered.

154 THE AMERICAN NATURALIST [Vou. L

The last space shows that 3.7 per cent. of these off-

spring were less than two thirds the normal size. This

again compares unfavorably with the other combinations,

and here also the individuals born one in a litter show a

superiority over those born in litters of two.

From a consideration of this table it may be concluded

that the vigor of a guinea pig varies inversely with the

size of the litter in which the animal is produced, and this

is equally true whether the animal is born from normal

or alcoholic parentage. However, the differences be-

tween the mortality of animals born in litters of one, two

or three from normal parentage are not nearly so great

as comparable differences between the members of the

small and large litters from alcoholic lineage. For ex-

ample, the difference in mortality between normal ani-

mals in litters of one or two is about 1 per cent., or

scarcely any; between these and the mortality of speci-

mens born three in a litter there is a difference in mor-

tality record of about 24 per cent., to the discredit of the

larger litters.

The comparable differences in the alcoholic lines is ever

so much greater. There is almost 14 per cent. higher

mortality among individuals from litters of two than

from litters of one, and actually about 43 per cent. higher

mortality among animals from litters of three than from

litters of one. The difference between the mortality per-

centages in the litters of one and the litters of two from

alcoholic inbred animals is 50 per cent. In other words,

the mortality is three times as high among individuals

from litters of two as from litters of one in inbred alco-

holies, while the normal individuals born in litters of two

are equally as good as those in litters of one. The par-

ents from the injured alcoholic lines are incapable of pro-

ducing large litters of strong individuals. The sub-

normal fetus fares pretty well alone in the uterus but is

put at a great disadvantage by having to divide its uter-

ine nourishment with brothers and sisters.

Another almost equally plausible explanation of this

No. 591] TRANSMISSION OF DEGENERACY 155

striking difference in quality and vitality among the

members of small and large litters might be given. It

may be supposed that the growth capacity of the eggs

maturing in the ovaries of normal and subnormal indi-

viduals depends somewhat upon the number of eggs ma-

turing at any one time, or ovulation period. A normal

animal may be capable of developing two entirely good

eggs at an ovulation, or possibly three, whereas a weak-

ened, less vigorous individual has ovaries incapable of

producing more than one well-nourished or well-devel-

oped egg at any one time. Of course, it is understood

that the small size of a mammalian egg would make it

seem as though it required very little stored food from

the ovary, yet that little must be of an extremely fine

quality, since so much of the energy of early development

is derived from the materials stored within the egg.

One point which might be interpreted to favor such an

explanation is the fact that the small, weak young con-

tained in the large litters do not recover and make their

shortage good after birth, as might be expected if their

inferior condition was simply due to a lack of nourish-

ment available in the overcrowded uterine environment in

which their late stages of development were passed.

Lack of intra-ovarian nutrition would certainly produce

a more lasting effect, since it occurs at an earlier stage

than lack of uterine nutrition, though of course we do not

pretend to deny that poor uterine nutrition would also

leave its persisting mar.

When only one young was produced in a litter the aver-

age growth rate of such individuals during the first

month after birth was 85.09 grams. Such specimens

were not only largest at birth, but they grew fastest after

birth. Animals born in litters of two increased 68.46

grams during the first month after birth, while those born

three in a litter gained only an average of 63.6 grams

during the same period. In other words, the last group

only gained 75 per cent. of the amount gained by simi

specimens which were fortunate enough to be developed

alone in the ovary and in the uterus. : a

156 THE AMERICAN NATURALIST [Vou. L

A second conclusion indicated by Table II is that in-

breeding the defective alcoholic stock produces a quality

of offspring decidedly inferior to that produced by the

alcoholic lines when not inbred. This involves the inter-

nal factors of the germ cells. When a modified germ cell

is united with a related one probably modified in a closely

similar manner, a summation of the modification pro-

duces a more decidedly modified individual than would

result from the combination of two non-related germ

cells, even though they also be modified. In other words,

as is shown in much of the data on heredity in higher ani-

mals, relatives probably respond to the treatment more

nearly in the same way than do non-relatives, and there

fore inbred defectives produce the most disastrous results

obtainable.

THE RELATIVE CONDITIONS oF THE MALE AND FEMALE

DESCENDANTS FROM PATERNAL AND FROM MATERNAL

ALCOHOLIZED ANCESTORS

We may now consider the possibility of analyzing the

relative influences of various alcoholized ancestors on

their offspring of different sex and the descendants of

such offspring. The problems may be stated thus: are

the offspring from alcoholized males more or less degen-

erate or modified than those from alcoholized females,

and is there a difference in the degree of degeneracy be-

tween the male and female offspring? Are the descend-

ants from alcoholic grandparents on the father’s side

more or less defective than the descendants from alco-

holic grandparents on the mother’s side, and do alcohol-

ized grandfathers and grandmothers show an equally

strong tendency to stamp their grandchildren? Do the

grandsons and granddaughters show relatively different

conditions, depending upon whether they are descended

from alcoholized grandfathers on the father’s or the

mother’s side or from alcoholized grandmothers on the

paternal or the maternal side?

Table III, which excludes all inbred animals, is a sum-

TABLE III

Tar i as DEGENERATIVE INFLUENCE OF MALE AND FEMALE ALCOHOLIZED ANCESTORS ON THEIR MALE AND FEMALE DESCENDANTS.

)

(ALL INBRED

ANIMALS ARE EXCLUDED

: are aun aru Po om art

Un- of Un- - f Un

Sex Sex Sex Sex

with with with with th

Alcoholized Alcohi Grandfath Alcoholized Grandfath Alcoholized Grandmoth Alcoholized Grandmoth

=a Fae A MoD on tho Father's Side > = the Mother's Side pæ the Father’s Side s a the Mother's Side

nur 44 43 70 37 23 38 36 38 26 34 32 48 37 33 36 23 18 25

mths.....| 37 33 34 22 0 26 23 23 19 0 25 23 18 14

: 7 84.1% 76.74% 91.89% | 95.65% 72.22%| 60.52% 67.64% | 59.87% 67.56%| 69.69% | 78.25%| 77.77%

44.58% 57.14% 49% 36.84% 45.28% 48.48%

ed within 3

momha... 7 | 10 70 3 1 38 10 15 26 11 13 48 12 10 36 5 25

115.9% 8.1% | 4.34% | 100% | 27.77%| 39.47%| 100% | 32.35% | 40.62%| 100% | 32.43%! 30.30% 100% | 21.74%! 22.22%| 100%

55.41% 42.85% 51.0% 63.15% 54.71% 51.51%

=o 1 3 3 0 0 0 3 4 1 4 T 2 3 7 2 0

- 12.27%] 4.65% | 4.28% 8.33% | 10.52% 2.64% 12.5% |14.58%| 5.40% 9.037% 16.66%, 8.69%

3.82% 0 7.0% 10.52% 10.37% 2.77%

158 THE AMERICAN NATURALIST [Vov. L

marized analysis of these questions. The male and fe-

male descendants from six different lines are tabulated.

The table is not perfectly pure, but merely represents a

mass result, since, for instance, in giving the young from

alcoholized fathers some of these young had also alco-

holized grandparents, etc. The same is true of the other

lines. But the large majority of the figures are from

unmixed matings, so that these mass results do have some

real significance.

In the first vertical section is given the records of off-

spring from alcoholized fathers. Forty-four males, 43

females and 70 young of unknown sex are considered.

Of the males 84 per cent. lived, and 76.7 per cent. of the

females lived. These numbers are very high, since in the

early part of the experiment only those young which sur-

vived were catalogued for sex. Therefore, all of the 70

young of unknown sex were animals which died at birth

or soon after, and as the table shows more than half of

the animals from alcoholized fathers died soon after birth.

The mortality among the male offspring from alcohol-

ized fathers was 15.9 per cent., while among the female

offspring it was considerably higher, being 23.25 per cent.

The same difference in quality between the sexes is illus-

trated by the percentage of gross deformities. Only

2.27 per cent. of the males were deformed, while twice as

great a proportion, or 4.65 per cent., of the female off-

spring from treated fathers were deformed. In all 3.82

per cent. of the offspring from alcoholized males were de-

formed and the female offspring were inferior in quality

to the male.

The next section of the table presents similar data for

the offspring from alecoholized mothers. There were 37

male, 23 female and 38 offspring of unknown sex. Again,

the offspring in which the sex was ascertained during

part of the experiment were only those that survived,

therefore, their mortality record is very good, while all

the animals of unknown sex were individuals that died

shortly after birth. Yet the records of the males and fe-

No. 591] TRANSMISSION OF DEGENERACY 159

males are based on exactly the same kind of data and are

to be fully compared. Eight and one tenth per cent. of

the males died, while only 4.34 per cent., proportionately

about half as many, females died. Not one grossly de-

formed animal was found among the offspring of alco-

holized females.

Thus from the mortality records the sons of alcoholized

mothers appear more affected than their daughters. And

taken as a whole the records of the alcoholized mothers

are superior in quality to those of the alcoholized fathers,

thus indicating that the male germ cells are more injured

by the treatment than the female germ cells.

The third section shows the records of the male and fe-

male descendants from aleoholized grandfathers on the

father’s side. Here the mortality record of the males is

much better than that of the females; 27.77 per cent. of

the males died soon after birth, and 39.47 per cent., a

very much higher proportion of the females, died.

The mortality of these animals from alcoholic grand-

parents seems much greater than that of animals from

treated parents; this is due, however, to the fact that the

sex of many more of these that died at birth was ascer-

tained as they occurred later in the experiment when this

point was being watched. The totals are the only figures

in the horizontal mortality columns that are to be com-

pared. The total mortality of descendants from alco-

holic grandfathers on the father’s side was 51 per cent.,

which is higher than the mortality of the offspring from

aleoholized mothers, 42.85 per cent., but lower than the

mortality of offspring from alcoholized fathers, which

reached 55.41 per cent. :

Among the ascertained male descendants from an alco-

holized paternal grandfather 8.33 per cent. showed gross

deformities, while 10.52 per cent. of the descendants as-

certained to be female were deformed. Considering all

the animals in this group, 7 per cent. were deformed,

which is almost twice as great a proportion as occurred

among the offspring of alcoholized fathers. The deform-

160 THE AMERICAN NATURALIST [Vou. L

ities in the F, generation are more frequent than in the F.

The fourth section shows the influence on the grand-

children of an alcoholized grandfather on the mother’s

side. This is the most injurious combination shown.

Only 36.84 per cent. of the offspring survive. Of the male

descendants of an alcoholized maternal grandfather 32.35

per cent. died soon after birth, while proportionally many

more, or 40.62 per cent., of the female descendants died.

In all a total of 63.15 per cent. of the descendants from

alcoholized maternal grandfathers died, which is the high-

est mortality record obtained.

Among the grandchildren of alcoholized maternal

grandfathers 10.52 per cent. were deformed, a very high

proportion. But of the grandsons only 2.64 per cent.

were deformed, while almost five times as many, or 12.5

per cent., of the granddaughters were grossly deformed.

Thus the females of the F, generation from a treated

maternal grandfather are poorer when considered from

the standpoint of mortality record and bodily structure

than the male F,’s from the same source.

The fifth line to be considered is that of an alcoholized

grandmother on the father’s side. The result of this

treatment as shown by the grandchildren is very bad, but

not quite so bad as from the alcoholized maternal grand-

father just discussed.

From alcoholized paternal grandmothers the conditions

of 37 grandsons and 33 granddaughters are to be com-

pared. About the same survival record is shown by both

sexes: 67.56 per cent. of the male grandchildren lived and

69.69 per cent. of the females lived. Of all the descend-

ants from this combination, including those in which the

sex was not determined, only 45.28 per cent. survived,

giving a mortality record of 54.71 per cent., considerably

better, by almost 10 per cent., than that of animals from

an alcoholized maternal grandfather in the preceding

section.

A large proportion of the animals from alcoholized pa-

ternal grandmothers were deformed, 10.37 per cent.

However, only 5.4 per cent. of the grandsons were de-

No. 591] TRANSMISSION OF DEGENERACY 161

formed, while many more, 9.03 per cent., of the grand-

daughters were deformed and among the young of unde-

termined sex 16.66 per cent. were deformed.

In the last section the records of descendants from alco-

holized grandmothers on the mother’s side are given.

There were 23 males, 18 females, and 25 young which died

with their sex unascertained. Forty-eight and forty-

eight hundredths per cent. of the animals lived. The

mortality among the males was 21.74 per cent., about the

same as that of the females which was 22.22 per cent.

The total mortality being 51.51 per cent. From this com-

bination there occurred a low percentage of deformities,

confined entirely to the grandsons. So that 8.69 per

cent. of the grandsons from alcoholized maternal grand-

mothers were deformed, while none of the granddaughters

showed any gross structural abnormalities.

TABLE IV

MORTALITY DURING THE First THREE MONTHS OF THE DESCENDANTS OF

KNOWN SEX From ALCOHOLIC ANIMALS (Nor INBRED

Males Females All Together

Treated with a re ee EE DOG

Alcohol Total Died Tow Died Total Died

ag, Early Mortality — Early Mortality eho Early Mortality

Tü o 44 | 7 |15.90%| 43 | 10 |23.25%| 87 | 17 |19.54%

Mother -sa 37 3 8.10%! 23 1 4.34%, 60 6.66%

Grandfat

her

on father’s side.| 36 | 10 | 27.77%) 38 | 15 |39.47%| 74 25 | 33.78%

mother

on father’s side.| 37 | 12 |32.43%| 33 | 10 | 30.30% 70 | 22 | 81.42%

Grandfather on

her’s side 34 | 11 |32.35%| 32 | 13 | 40.62%) 66 | 24 86.86%

21.95%

Grandmother on

mother’s side...| 23 5 | 21.74%) 18 4 | 22.22%) 41

Table IV presents in a more concise manner certain of

the figures considered in the foregoing Table III. Only

the mortality records of the male and female descend-

ants from different sources and the total mortality of the

several groups is shown by the table and thus a ready

comparison of the conditions may be made. Among the ©

= offspring from alcoholic fathers 15.9 per cent. of the

-males died and 23.25 per cent. of the females. The fe-

162 THE AMERICAN NATURALIST [Vou. L

male are more injured than the male offspring of treated

fathers. The next horizontal line shows that the off-

spring from treated mothers are far better than from

treated fathers, having a much lower mortality. The

male germ cell is more affected by the alcohol than the

ovum, therefore treated fathers produce poorer offspring

than treated mothers.

The heterogeneous female descendants from an alco-

holized paternal grandfather are more affected than the

male, 39.47 per cent. mortality to 27.77 per cent.

The male and female descendants from an alcoholized

paternal grandmother show about equal conditions, 32.43

per cent. male mortality to 30.3 per cent. female mortality.

The heterogeneous female descendants are inferior to

similar male descendants from an alcoholized maternal

grandfather, 40.72 per cent. female mortality to 32.25 per

cent. male mortality.

The male and female descendants are about equally

strong from an alcoholized maternal grandmother, 21.74

per cent. male mortality to 22.22 per cent. female mor-

tality.

Although explanations of the above differences between

the ways in which the male and female guinea pigs are

affected by the treatment, as well as explanations of the

different records of the grandsons and granddaughters

from alcoholization of different ones of the four grand-

parents, are difficult to give at the present stage of the

experiment, a tentative explanation based on the compo-

sition of the chromosomal complex is certainly suggested.

GENERAL CoNSIDERATIONS

In the case of the male guinea pig, according to the

studies of Miss Stevens (’11), two kinds of spermatozoa

are produced. The one has a large X chromosome, the

female-producing spermatozoon, and the other contains

a corresponding small Y chromosome, which from homol-

ogy with other forms we may consider to be the male-

producing. The two classes of spermatids are different

Alcoholic

oy ae

yf Xin) 15.9% Mortality 23.2

2.21% 3

5% (a)X f || x

Deformed 4.65% :

Descendants Descendants

1% Mortality 63.15%

1% Deformed 10.52%

Alcoholiv

of N”

“i” 3.1% Mera 4.34% Xini Fox

O O

orm&

Descendants

54.11% Mortality 5l. 51%

10.31% D porone 2 TES

arg in quality between

les, and indicating a grows

jur chromosome gives to the males an

smaller va a f deformities and a lower mortality recor

wer at of the shows equal amounts of ured chroma

the black X

passing TRA the treated pote to her male and female

chromosomes ounts of pries chromat’n are e ed with

unequal amounts of normal praa from the pra

Y chromosomes. The pring, 7 ese have in

odified and less

unequal white X and

their chromoso tu proportionally normal

than the female offspring which have equal am norm d

alcoholic. And the male offspring and their actually found to

rtion of deformities than the fe-

show a higher mortali ter pro

male spring and thet ty and a great propo

164 THE AMERICAN NATURALIST "Vou. L

morphologically and we may expect them to differ in their

susceptibilities to the alcoholic treatment. One class may

be more affected than the other. This might be due sim-

ply to the reason that one class of spermatozoa actually

according to mass has more chromatin to be acted upon

than the other. And this difference in mass of affected

chromatin might be sufficient to give a difference in qual-

ity between the individuals arising from the two classes

of spermatozoa.

At any rate, as the accompanying diagram indicates,

there is a decided difference between the records of the

male and female offspring from treated guinea pigs. The

upper half of the accompanying diagram shows that the

mortality is higher and the gross deformities more fre-

quent among the female offspring sired by alcoholized

male guinea pigs than among the male offspring. This

difference we may venture to suppose is due to the fact

that the female offspring actually receive more modified

or injured chromatin from the alcoholic father than do

the sons. The diagram is an attempt to represent this

larger mass of injured chromatin, the large black X chro-

mosome passing to the daughters, while the smaller

black Y chromosome is received by the sons.

Another possible explanation might be that the two

heteromorphic sex chromosomes, the X and Y, respond

differently to the influence of the alcoholic treatment, the

X being the more affected. Such an opinion has some

basis, since these chromosomes in the later development

of the two sexes seem to carry such a number of contrast-

ing qualities according to the splendid evidence presented

by Morgan and his associates. One may be permitted to

assume on probability, at any rate, that the X and Y

chromosomes are qualitatively different in their finer

chemical constitutions, and this qualitative difference

would necessitate a different response to the chemical

treatment on the part of each of the two chromosomes.

There is also important evidence from the partheno-

genetic groups, as, for example, the Phylloxerans and

No. 591] TRANSMISSION OF DEGENERACY 165

Aphids (Morgan, ’09), which might lead one to believe

that the two classes of spermatids or finally spermatozoa

are never quite equally active or vigorous. This differ-

ence may vary from apparent equality in most higher

animals to cases such as the parthenogenetic Phyllox-

erans and Aphids, in which one class of spermatids are

actually degenerate and non-functional.

In this connection an experiment performed with a

quite different problem in view by Cole and Davis (714)

with alcoholized rabbits is suggestive. They. found that

when two male rabbits were mated with a female super-

fetation occurred in most cases so that part of the result-

ing litter of young were sired by one male and part by

the other. The two males differed in their ability, so

that one more often sired the majority of young of a given

litter and in the total number of competition matings sired

the greater number of young. When this male with the

fertilizing advantage was treated for a short period of

time, a month or more, with fumes of alcohol he was then

affected in such a way that when mated in competition

with the same male he normally had beaten he now failed

to sire any young. Yet if mated singly or alone with a

female he still had the power to beget offspring. The

alcohol treatment had in some way lowered the power of

his spermatozoa to fertilize an egg. Thus these sperma-

tozoa could no longer fertilize an egg in the presence of

the spermatozoa from a male which was originally less

potent than they.

All of these data indicate differences in the behavior

and reactions of the individual germ cells, and such dif-

ferences probably account for the discrepancy existing

between the conditions of the male and female offspring

from an alcoholized father. Since this point has only

recently been discovered in the experiments, we now have

very few definite matings to test its meaning by back

crosses with the normal. But a large number of hetero-

geneous matings have been made during the last few

years and their gross results serve to verify the fact that

166 THE AMERICAN NATURALIST [Vou. L

the difference in quality between male and female off-

spring is actual, although such matings furnish no definite

analysis of the conditions.

In the first place, the upper half of the diagram shows

that the mortality is higher and the defects more frequent

among the female offspring of treated males than among

their sons. The products of the heterogeneous matings

in which these male and female offspring have taken part

go to indicate that the first apparent difference in their

records was a real difference. The mortality record of

the mass descendants from the sons of alcoholized fathers

is about 20 per cent. better than the mortality record from

the descendants of the. daughters of alcoholic fathers.

And the proportion of deformities is 50 per cent. higher

among the descendants of the daughters than among the

descendants of the sons. These conditions of the de-

scendants prove that the female offspring from the alco-

holized males are actually worse than the male offspring

in the following respects: their mortality record, the fre-

quency of deformities, and the quality of young to which

they give rise. The only plausible way to account for the

origin of this difference is to assume that the female-

producing spermatozoa were more modified by the treat-

ment than the male-producing spermatozoa. Whether

such an increased modification is due to the presence of

a greater mass of chromatin to be injured in the one case

than in the other or to a difference in response on the part

of the two heteromorphic sex chromosomes it is impossi-

ble to state. The difference, however, is a fact!

The lower half of the diagram illustrates the different

qualities of the male and female offspring from alcohol-

ized mothers. Here each sex of the offspring in accord-

ance with prevalent cytological views receives an equal

amount of chromatin from the treated mother. And,

moreover, as far as the treated mother is concerned simi-

lar chromosomal complexes are conveyed to both sexes

of the offspring. The two classes of young should, there-

fore, show similar conditions, but such is not the case.

No. 591] TRANSMISSION OF DEGENERACY 167

The mortality of the male offspring is higher than that of

the female. This condition may probably be explained

on the same principles we have employed above. The

two sexes receive equal amounts of injured chromatin

from their aleoholized mother, but this injured chromatin

in the case of the female individuals is mixed with a

larger amount of normal chromatin from the father than

is the case with the male. The female combination of

equal amounts of good and bad chromatin gives rise to a

better product than the male combination of a larger

amount of modified chromatin with a smaller amount of

good. Therefore the records of the male offspring are

inferior to those of the female offspring.

The female combination, ovum and spermatozoon with

equal amounts of chromatin, good and bad, is proportion-

ately less injured than the male combination, ovum with

a larger amount of bad and spermatozoon with a smaller

amount of normal chromatin. The diagram represents

the black X chromosomes equal in size passing to the

daughters and the sons to be combined with the large

white normal X in the case of the daughters and with the

small white normal Y in the case of the sons.

Again the descendants from heterogeneous matings of

these males and females prove that there is an actual

difference in quality. The descendants from the sons of

alcoholic mothers show a slightly higher mortality and a

much greater proportion of deformities than are found

among the descendants of their daughters.

We believe that these results actually show a difference

in response to the treatment on the part of the male- and

female-producing spermatozoa. Such a difference log-

ically follows the cytological differences in structure

which Wilson and others have so clearly demonstrated

during the past ten years. If this structural difference

is of any significance, as it surely must be, then such

physiological differences in behavior as are indicated in

our results should sooner or later be found.

On such a basis as this the sex ratio in different classes

168 THE AMERICAN NATURALIST (Vou. L

of animals may possibly be explained. A species such as

man, which constantly seems to produce more males than

females, may be said to form more active or vigorous

male-producing spermatozoa. In the competition to fer-

tilize the egg such spermatozoa win an advantage and in

the sum total more males than females arise, the ratio

depending upon the extent of the advantage the one class

of spermatozoa has over the other.

We now have under way a number of matings which

are designed to test these propositions in an analytical

fashion. One of us (Papanicolaou, 715) is in possession

of data giving reason to believe that a second explanation

may be offered to account for the different conditions pre-

sented by the male and female offspring produced by alco-

holized females. Such an explanation is based on the

supposition that the female guinea pig as well as the male

has a share in the determination of the sex ratio and may

produce two kinds of ova. Such an explanation in its

final analysis is extremely complex and unnecessary in

the present discussion, though it will be presented in a

future consideration of the regulation of the sex ratio in

these animals.

Admitting, as is suggested above, that the two groups

of spermatozoa differ in their response and resistance to

the treatment, we may also admit that there are other

normal differences in their vitality and behavior. These

normal differences must also vary within certain limits.

In one group of animals the female-producing sperma-

tozoa may be more active and possess a higher degree of

fertilizing power than the male-producing spermatozoa.

Such a group would show a sex ratio below one hundred,

there being more females than males produced. In other

species of animals with a sex ratio of more than 100 the

reverse condition obtains; the male-producing sperma-

tozoa possess on an average a higher fertilizing power

than the female-producing. But the advantage of the

male-producing sperm may be slight and no doubt many

individual males tend to form female-producing sperm

No. 591] TRANSMISSION OF DEGENERACY 169

with a higher fertilizing power than the male-producing.

Such individuals will more frequently beget female off-

spring. Slight differences in the physiological behavior

of the two classes of spermatozoa would account for the

sex ratios in all animals, and finally, as Morgan has

shown, the extreme difference between the qualities of the

two classes of spermatozoa leads to the degeneration of

one entire class and the necessary production of only one

sex from the fertilized eggs of these species. Such spe-

cies must also be parthenogenetic in order to produce in-

dividuals of the other sex.

This discovery by Morgan suggests, as Wilson (711)

brings out in his review of the sex chromosome question,

a plausible explanation of the sex ratios in different

classes of animals. And we believe the evidence pre-

sented above lends further support to such an interpre-

tation.

A rather old popular idea in attempting to explain the

sex riddle may have some ground of fact from the stand-

point of the variations in the differences of fertilizing

power of the two classes of spermatozoa. It has often

been claimed that one testis is male-producing and the

other female-producing. Every one knows that this 1s

untrue. Yet one testis may have a tendency to produce

spermatozoa of the female class with a higher fertilizing

power than the male sperm of this testis, and the other

testis might have an opposite tendency, since the condi-

tions of behavior often differ in two organs of a bilateral

pair. An animal which has produced a large proportion

of male offspring may after semi-castration produce al-

most all female offspring. A possible explanation for

such an occurrence would be that the removed testis had

produced more vigorous male sperm than female and the

spermatozoa of this testis possessed the higher fertilizing

power, while the remaining testis tended to produce more

potent female sperm. On removing the one testis the

other came into supremacy. In the same imaginary case,

if the opposite testis had been removed, there would have

170 THE AMERICAN NATURALIST [Vou. L

been no change in the tendency to produce offspring of

a certain sex, since the remaining testis originally pos-

sessed an advantage.

Finally, then, from the above experiments there is no

question that the material basis of the hereditary quali-

ties has been injured, since alcoholized males have trans-

mitted the injury to four generations during a period of

almost five years. In other words, as stated above, chro-

matin injured five years ago is now living in the great-

grandchildren of the individuals in which it was injured.

Bardeen with the X-ray and Oscar Hertwig with ra-

dium have induced similar injuries by directly treating

the spermatozoa, but these cells were so greatly injured

that only the immediate effect upon the developing em-

bryo was shown. The present experiments, however,

demonstrate the passage or transmission of the injured

chromatin from generation to generation during a period

of years. The behavior of the carriers of heredity be-

comes pathological just as any other organ with a normal

function may behave in an abnormal or pathological

manner.

Mammals are particularly adapted to the study of such

features of heredity as this, on account of their typical

structure and large, easily observed organs. The com-

plexity of their structure and behavior further permit the

possibility of slight modifications becoming visible

through abnormal conditions of their nervous system, ete.

Thus with such material as guinea pigs a few experiments

of this kind may furnish certain clues to the processes of

behavior of the chromosomes that less plastic and simpler

forms might never present in such a manner as would be `

recognizable.

On the other hand, the small litters and comparatively

slow breeding render these higher animals unsuitable for

an exhaustive analysis of many of the intricate problems

of normal heredity.

No. 591] TRANSMISSION OF DEGENERACY ae!

SUMMARY AND CONCLUSIONS

In the foregoing pages we have considered the results

of an experiment now in progress for more than five years

which analyzes to some extent the influence on the off-

spring of alcoholizing either one or both parents and the

manner of hereditary transmission of the induced effects

to subsequent generations,

The experiments have demonstrated on two different

stocks of normal guinea pigs that the parental germ cells

may be so modified by chemical treatments that they are

‘rendered incapable of giving rise to a perfectly normal

offspring. This incapacity is probably due to modifica-

tions of the chromatin or carriers of the hereditary quali-

ties within the germ cells, since the great-grandchildren,

the F, generation, from the treated animals are usually

more decidedly affected and injured than the immediate

offspring (F,) of the aleoholized animals.

This then becomes a study of the behavior of diseased

or pathological chromatin in heredity. Chromatin ren-

dered pathological more than four years ago is still living

and has now been passed on to the F, generation from the

alcoholized great-grandparents. The F, animals are

almost without exception incapable of reproduction and

are in many ways subnormal and degenerate.

Studies of abnormal heredity may possibly furnish a

means of analyzing the normal methods of action by

which the minute carriers of hereditary qualities con-

tained within the fertilized egg are capable of causing

complex developmental and structural changes to reoccur

from generation to generation in so wonderfully consist-

entamanner. Justas the knowledge furnished by studies

of experimentally modified embryonic development has

supplied valuable data towards a clearer understanding

of the normal processes and changes which occur in the

developing embryo. :

The treatment of adult guinea pigs by an inhalation

method with daily doses of alcohol through several years

produces little if any noticeable effect upon the organs

172 THE AMERICAN NATURALIST [Vou. L

and tissues of the animal’s body. The direct action of

alcohol fumes tends to injure the respiratory mucosa and

to render the cornea of the eye dull or opaque. These

changes, however, do not inconvenience the animals in

any perceptible way, and they remain strong and hardy

and live as long and actively as the untreated guinea pigs.

In spite of their healthy appearance the injurious influ-

ence of the alcohol inhalation is very decidedly shown by

the quality of offspring to which the treated animals give

rise. And the descendants of these offspring are even

worse than the F, generation when compared with the

different generations of control animals produced under

identical cage and food conditions.

The males seem to be more injured by the treatment

than the females, taking as an index of injury the quality

of their offspring and descendants. Stating it differ-

ently, the spermatocytes or spermatozoa are more sensi-

tive to the changed chemical condition of the tissues than

are the female germ cells.

There is a larger proportion of degenerate, paralytic

and grossly deformed individuals descended from the

alecoholized males than from the alcoholized females.

The records of 682 offspring produced by 571 matings

of animals of various types have been tabulated to show

the kinds of litters of young produced and their ability to

survive. One hundred and sixty-four matings of alcohol-

ized animals, in which either the father, mother, or both

were alcoholic, gave 64, or almost 40 per cent., negative

results or early abortions, while only 25 per cent. of the

control matings failed to give full-term litters. Of the

100 full-term litters from alcoholic parents 18 per cent.

contained stillborn young, and only 50 per cent. of all the

matings resulted in living litters. Forty-six per cent. of

the individuals in the litters of living young died very

soon after birth. In contrast to this record 73 per cent.

of the 90 control matings gave living litters and 84 per

cent. of the young in these litters survived as normal,

healthy animals.

No. 591] TRANSMISSION OF DEGENERACY 173

The mating records of the descendants of the alcohol-

ized guinea pigs, although they themselves were not

treated with alcohol, compare in some respects even

more unfavorably with the control records than does the

above data from the directly alecoholized animals.

Of 194 matings of F, animals in various combinations

55 have resulted in negative results or early abortions, 18

stillborn litters of 41 young occurred, and 17 per cent. of

these stillborn young were deformed. One hundred and

twenty-one living litters contained 199 young, but 94 of

these died within a few days and almost 15 per cent. of

them were deformed, while 105 survived and 7 of these

showed eye deformities. Among 126 full-term control

young of the same stock not one has been deformed.

The records of the matings of F, animals are still

worse, higher mortality and more pronounced deformi-

ties, while the few F, individuals which have survived

are generally weak and in many instances appear to be

quite sterile even though paired with vigorous, prolific,

normal mates.

The structural defects shown by the descendants of

aleoholized animals seem to be confined chiefly to the cen-

tral nervous system and special sense organs. Many o

the young animals show gross tremors, paralysis agitans;

the hind legs, fore legs or both legs of one side may be

paralyzed (Plates I and II). Eye defects are very com-

mon, such as opaque cornea, opaque lens, various degrees

of monophthalmicum asymmetricum, and finally several

cases of complete anophthalmia have occurred, the entire

eyeballs, optic nerves and optic chiasma being absent

(Figs. 1 to 3 and Plate III).

The quality of individuals from the same parentage

varies inversely with the size of the litters in which they

are produced. Animals born one in a litter are rather

strong, even though derived from very bad alcoholic lines.

This difference between the members of small and large

litters is also shown by the normal animals, but the differ-

ence in quality between members of large and small litters

174 THE AMERICAN NATURALIST [Vou. L

is ever so much greater in the alcoholic lines. There is

also some tendency on the part of the alcoholic animals to

produce a greater proportion of small litters and this aids

somewhat towards the perpetuation of their lines.

Inbreeding tends to emphasize the alcoholic effects.

This is probably due to related animals responding to the

treatment in closely similar ways on account of the simi-

larity of their constitutions. Inbreeding, as such, may

be harmful. But inbreeding added to the alcohol effects

produces a much worse condition in the offspring than

either inbreeding or alcoholism alone could do.

The data from alcoholized male lines indicate that the

female offspring from alcoholic males are less viable and

more frequently deformed than the male offspring. And

heterogeneous matings of such male and female offspring

further emphasize the same inferiority on the part of the

female offspring from treated males. This is a very sig-

nificant fact.

The fact that the offspring of one sex differ in quality

from those of the opposite sex, and that the female off-

spring of an alcoholic male are inferior to his male off-

spring suggests at once a difference between the germ

cells concerned in the production of the male and female

young. Miss Stevens showed that the spermatocytes of

the male guinea pig contained a heteromorphic pair of

chromosomes and half of the spermatozoa would be ex-

pected to receive one member, the X chromosome, of the

heteromorphic pair and one half of the spermatozoa

the other member, the Y chromosome, of the heteromor-

phic pair. We now have two possibilities in explanation

of the above facts. In the first place, it may be assumed

that the alcohol acts similarly on all of the chromatin to

injure it. Thus a mass action would cause the sperma-

tozoa carrying the larger member of the heteromorphic

pair to deliver more injured chromatin and the other

spermatozoa with a less total amount of injured chro-

matin would deliver less when they fertilize eggs contain-

ing equal amounts of normal chromatin. The fertilized

No. 591] TRANSMISSION OF DEGENERACY 175

egg giving rise to the female, therefore, contains a greater

proportional amount of alcoholic chromatin to normal

chromatin than does the egg giving rise to the male.

And so the female product is actually more injured than

the male.

A second possible explanation of these conditions may

be that the X and Y chromosomes themselves respond

differently to the treatment, the X being the more sensi-

tive of the two. But in either case the two classes of

spermatozoa certainly seem to respond differently to the

treatment and this shows a physiological difference in

behavior to correspond with the well-known morpholog-

ical differences so often found between the two groups of

spermatids of many animal species.

The data from alcoholic female lines indicates that the

male offspring from alcoholic females are inferior in

quality to the female offspring. And heterogeneous

matings of such male and female offspring further prove

the inferiority on the part of the male offspring from

treated mothers. This is also significant. How can it be

put in accord with the above chromosomal explanations

for the difference in quality between the female and male

young of alcoholized fathers?

If we admit that all of the eggs arising from an alco-

holized female guinea pig are homomorphic and contain

groups of chromosomes equal in mass, it follows that her

male and female offspring receive the same amount of

injured chromatin and should be affected by such chro-

matin to equal degrees. But this is only part of the case,

the injured female chromatin is combined with normal

chromatin from the normal father when the eggs are fer-

tilized and here the difference arises. The female off-

spring receives from the normal father a larger amount

of normal chromatin than do the male offspring. So that

the female arises from an egg in which equal amounts of

good and injured chromatin are present, while the male

offspring arises from an egg in which a larger amount of

injured chromatin is united with a smaller amount of

176 THE AMERICAN NATURALIST [Vou. L

normal. Therefore, proportionally, the male offspring

from treated mothers have more injured chromatin in

their entire bodily make up than do the female offspring,

and are comparatively in a more abnormal condition.

Another explanation of these differences between the

male and female offspring of alcoholized females could be

based on the possibility of the female being heterozygous

for sex. This involves a very complex discussion, but one

for which there is some ground on the basis of the regu-

lation of the sex ratio in these animals.

Finally, then, the experiments show the hereditary

transmission through several generations of conditions

resulting from an artificially induced change in the germ

cells of one generation. And they furnish data of im-

portance bearing upon the pathological behavior of the

carriers of heredity as well as the differences in behavior

between the two types of germ cells produced by an ani-

mal carrying heteromorphic chromosomes.

LITERATURE CITED

Bardeen, C. R. 1907. Abnormal Development of Toad Ova Fertilized by

Spermatozoa Exposed to the Roentgen Rays. J. Exp. Zool.,

iv; pai

a ee

Cole, L. J., and Davis, ©. L. 1914. The Effect of Alcohol on the Male

Germ Cells, Studied by Means of Double Matings. Science,

. 476.

Hertwig, O. 1913. Verka an Tritoneiern über die Einwirkung bestrahl-

ter Samenfäden auf die tierische Entwicklung. Arch. f. Mikr.

Anat., 82, Abt. II.

Morgan, T. H. 1909. A Biological ait Cytological Study of Sex Deter-

mination in Phylloxerans and Aphids. J. Exp. Zool., VII,

39.

1912. ~ Elimination of the Sex Chromosomes from the Male-pro-

cing Egg of Phylloxerans. J. Exp. Zool., XII.

1910 to 1938. Numerous Studies on serapi in Drossphiis. J. Exp.

Zool., Biol. Bulletin, AM. Nar,

Papanicolaou, Garge. 1915. Sex D Sd and Sex Control in

uinea Pigs. Science, N. S.,

Stevens, N. M. 1911. Heterochromossinor Í ii the Guinée Pig. Biol. Bull.,

XXI, p

Stockard, C. R. 1910. The Influence of Aleohol and Other So as on

Embryonic Development. Am. J. Anat.,

1912. An Experimental Study of Racial onset te in k Mammals

Treated with Alcohol. Arch, Internal Med., X

No. 591] TRANSMISSION OF DEGENERACY 177

1913. ee Effect on the Offspring of Intoxicating the Male Parent and

e Transmission of the Defects to Subsequent Generations.

e Nart., XLVII, p. 641.

1914. A Study of Further Generatións of Mammals from Ancestors

Treated with Alcohol. Proc. Soc. Exp. Biol. and Med., XI,

. 136.

1914, The Artificial Production of Eye Abnormalities in the Chick

mbryo. Anat. Record, 8, p.

Wilson, E. B., 1905-09. Studies on Gioii odios I to V. Jour. Exp, Zool,

Vols. 2, 3, and 6.

1911. The Sex Chromosomes Arch. Mikr. Anat., 77.

1914. Croonian Lecture: The Bearing of Cytological Research on

Heredity. Proc. Royal Soc., 88.

SHORTER ARTICLES AND DISCUSSION

VARIABILITY UNDER INBREEDING AND CROSS-

BREEDING

AN unusually thoughtful and suggestive discussion of evo-

lutionary problems is contained in Professor Walton’s paper on

‘‘variability and amphimixis’’ published in the November, 1915,

number of this journal. But the paper is in some danger of

neglect because the conclusions reached are apparently so revo-

lutionary that to a casual reader they may seem freakish. Yet

it will be seen by one who reads the paper more carefully that

the radical character of its conclusions is due in part to the fact

that certain familiar ideas are here viewed at a new angle.

Nevertheless the new point of view has, it seems to me, to some

extent, caused the author loss of perspective in relation to some

of the phenomena which he discusses, for which reason further

consideration of them may be profitable.

e occasion of Walton’s discussion was a biometric study

which he made of two sorts of zygospores produced by Spirogyra

inflata, one sort produced by union of cells in the same filament

(called by him ‘‘close fertilization’’), the other by the union of

cells in different filaments (called ‘‘cross fertilization’’). Zy-

gospores of the former sort (‘‘close fertilized’’) were found to

be on the average larger and more variable than those of the

latter sort, contrary to the prevailing idea that cross fertiliza-

tion leads to increased variability. It may however be ques-

tioned whether Walton’s material is such as to throw new light

on this question, for it is by no means certain that cells of

Spirogyra which unite in lateral conjugation are the exact

equivalents morphologically and physiologically of those which

unite in scalariform conjugation. It is conceivable that zygo-

spores formed in lateral conjugation may be larger and more

variable because the cells which gave rise to them were as a

group larger and more variable. It is conceivable that cells

which resort to scalariform conjugation are not such as can

satisfy their physiological demands for conjugation by uniting

with a sister-cell in the same filament. For it is known that in many

plants sexual union occurs only as a last resort, when conditions

178

No. 591] SHORTER ARTICLES AND DISCUSSION 179

are unfavorable for further continuance of the organism either

by vegetative reproduction or even by parthenogenesis. But

whether or not Walton’s own observations are considered per-

tinent, the question which it leads him to consider is one of pro-

found evolutionary significance—does cross fertilization (as com-

pared with close fertilization) tend to produce greater or less

variability.

In parts of his discussion Walton fails to keep clearly in view

distinctions, which he nevertheless recognizes, between the re-

Spective variabilities of F,, F, and mixed populations. The

fact has been known since the days of the early plant hybridizers,

and is expressed clearly in one of Focke’s laws of hybridization

that the first generation (F,) offspring of a hybrid cross are

not, as a rule, more variable than the more variable parent race.

In other words the generalization which Walton attacks, that

crossing produces variability, is not commonly, if at all, held by

biologists to apply to F, populations but only to the conditions

obtaining in subsequent generations. But Walton’s own obser-

vations are made exclusively upon F, zygotes. Supposing his

two classes of zygotes to be morphological and physiological

equivalents of each other (which, however, may reasonably be

questioned) there was no ground for expecting one sort to be

more variable than the other, so far as existing knowledge of the

effects of inbreeding and cross-breeding is concerned.

Walton cites two experimental investigations, in support of

his own observations on Spirogyra, to show that close fertiliza-

tion produces greater variability than cross fertilization, viz.,

that of Jennings on Paramecium and that of Barris and my-

self on Drosophila. But neither of these invesi ations deals

with the same sort of cases as Walton’s. i he gs is compar-

ing the variability of conjugants with: that of De ek sa conia.

This is a case where sexual is contrasted with asexual reproduc-

tion and is in no way comparable with a case in which the ef-

fects of cross and close fertilization are compared with each

other. I quite agree with Walton’s conclusion that the results

are statistically considered far from conclusive, and would add

that they are quite aside from the question which Walton is con-

sidering. Barrow’s comparisons were made between single lines

of Drosophila inbred (brother with sister) for from 30 to 61

generations and a culture derived from two original pairs of

1See Coulter, 1914, ‘‘The Evolution of Sex in Plants.’’

180 THE AMERICAN NATURALIST [Vou L

Drosophila the descendants of each pair being allowed to inter-

breed freely. As to the results we said (p. 776)

These experiments show no appreciable effect of inbreeding (on vari-

ability). In every case the brood reared under the best and the most

uniform conditions has the highest average number of teeth (in the sex-

comb), irrespective of whether or not inbred. The same may be said

of variation in size. Inbreeding has diminished neither the average

size nor the variability in size.

Walton considers these conclusions justified by our statistical

constants in the case of the sex-comb, but believes that a signifi-

cant difference is observed in length of tibia, which we found to

be both greater and less variable in the culture not inbred. He

criticizes our failure to calculate coefficients of variation for

tibia length (as we had done for sex-comb) and upon caleulating

such coefficients finds the greater variability of the inbred lots

significant. But the same difference in variability was indicated

by the standard deviations (which we gave) and the calculation

of the coefficient of variation adds nothing to the force of the

demonstration. We considered then and still consider the dif-

ferences observed sufficiently accounted for by external condi-

tions, 7. e., we considered them purely phenotypic. We showed

that lant of tibia is greatest and its variability least when

food and temperature conditions are best. The difference be-

tween two inbred races (M and N) inbred practically the same

number of generations (viz., 31 and 30, respectively) but treated

very differently as regards food, was found to be several times

greater than the difference between the inbred culture M and

the not-inbred culture X. Hence it is not probable that the in-

breeding had anything to do with the differences found in

variability.

It is difficult to understand how on any theory of heredity in-

breeding could be expected to increase variability within a single

inbred line, such as one of our inbred cultures of Drosophila.

On a Mendelian theory it would be expected that inbreeding,

brother with sister, for a large number of generations (61 in our

A series) would result in the production of a number of homo-

zygous lines, each of which by itself would be entirely devoid of

variability, except that due to environmental agencies. If all

the possible derived lines descended from a pair under inbreed-

ing were combined into one mass of material, it would seem

probable, on a Mendelian theory, that if any genotypic varia-

No. 591] SHORTER ARTICLES AND DISCUSSION 181

tions had occurred, this material would show greater variability

than the ancestral race before inbreeding began. This, I take

it, is the point which Walton has in mind when he asserts that

inbreeding has a tendency to increase variability. But this is

very different from the condition to be expected in any single

line considered separately, as in one of our inbred lines of Droso-

phila. Such a line should be less variable than the population

from which it arose, provided that population contained any

genotypic variations whatever!

The question is decidedly worthy of consideration, which Wal-

ton’s paper suggests, is evolution more rapid in a self-fertiliz-

ing or habitually close fertilizing population on one hand or in

a habitually cross fertilizing population on the other hand.

The importance of the question is not lessened by the fact that

Walton has brought into the discussion material wholly irrele-

vant, including his own observations on the zygospores of Spiro-

gyra and the observations of Jennings on Paramecium and Bar-

rows on Drosophila. But the work of Hayes on the variability of

pure races of tobacco and of their hybrids, which Walton cites,

does bear directly on this question. By combining the observa-

tions on the parent races into one mass of data and treating

this statistically, Walton has shown that self-fertilizing lines

mixed together would form a population more variable as re-

gards number of leaves and height of plant than the popula-

tion produced by cross-breeding of these same lines. Hayes’s

observations verify Focke’s law already cited, that the variabil-

ity of F, does not exceed that of the more variable parent race,

but that F shows increased variability. Theoretically F, should

show the maximum variability. Walton’s figures indicate

clearly that this maximum variability under cross-breeding is

less than the variability of a mixture of the two inbred races

and consequently that continuous self fertilization within a

mixed population will produce a more variable population than

will result from continuous cross fertilization. This is an im-

portant generalization which reagent: will hold good in

all cases in which ‘‘intermediate’”’ or ‘‘blending’’ inheritance

occurs. It would not hold good bes eases in which completely

dominant and quantitatively invariable Mendelian factors are

concerned, but it is doubtful whether such cases occur, as I have

elsewhere tried to show. It is the great variability of self fer-

tilizing populations and the stability of variations arising under

self fertilization (since no variations will be ‘‘swamped by cross-

182 THE AMERICAN NATURALIST [Vou. L

ing’’ in a self fertilizing population ) that allow of the forma-

tion and perpetuation of ‘‘little species,’’ side by side and yet

quite distinct, within highly variable taxonomic species such as

the dandelion. These same characteristics of self fertilizing

populations furnish much of the material which plant breeders

use. Following Vilmorin, they find it necessary only to isolate

and propagate by themselves the variations which spontaneously

arise. The task of the breeder who is dealing with a continu-

ally cross-breeding organism is much more complex. He often

finds it necessary first to inbreed his stock, in order to learn

what potential variations it contains, or, if one prefers so to

express it, in order to induce variability, though this form of

statement is not strictly accurate. Such inbreeding of a natu-

rally cross-breeding organism often causes temporary loss of

vigor, as notably in the case of maize, and frequently in domes-

tic animals. But when the desired variations have been isolated,

vigor can usually be recovered by increasing the stock to such

an extent that matings become possible within the race and yet

not involving union of closely related individuals.

Notwithstanding the utility of inbreeding in securing varia-

tions, there are important sources of variability which are found

in cross-breeding alone. Supposing that under inbreeding

variation has already occurred in different directions and the

original condition has been wholly lost, it is often possible to

recover it again by crossing. This is the familiar phenomenon

of reversion upon crossing. It is also possible by crossing to

combine in one race variations which have occurred separately

in different races, a thing which would be impossible under con-

tinuous inbreeding. But a certain amount of inbreeding must

usually in such cases follow up the cross-breeding in order to

isolate and make secure the combinations desired.

It is not wise, therefore, unduly to exalt either inbreeding or

cross-breeding as evolutionary processes or tools of the breeder.

Each has its utility at the proper time and place. They are like

pick and shovel, each supplementing the work of the other.

The question is worth considering in this connection—what

effect will inbreeding and cross-breeding respectively have on

the variability of single characters. This is a question to which

I have given considerable attention for several years and the

answer to it is, I think, becoming clear. A single character

which Mendelizes has its variability increased by crossing. Some

explain this as due to actual modification of the unit character

No. 591] SHORTER ARTICLES AND DISCUSSION 183

through crossing, others as due to the introduction of modify-

ing factors by means of the cross. Whichever view is adopted,

the fact is perfectly clear that modification of single Mendeliz-

ing characters occurs in cross-breeding. Under continuous in-

breeding we should expect that single Mendelizing characters

(within single lines but not within an entire inbred population)

would attain a condition as devoid of variability as it is possible

for them to attain and observation confirms this expectation.

As regards characters which ‘‘blend’’ in heredity, these are

not inherited as single characters; they do not Mendelize in the

ordinary acceptation of the term. The characters of the respec-

tive parent races disappear in the cross, being replaced by a

common intermediate condition or blend. This blend persists

into the F, and later generations but with a certain amount of

variability which is at a maximum in F, and beyond that point

tends to disappear in the absence of any special selection. It

points to imperfection of the blending process or, in the view

of those who prefer a Mendelian terminology for such cases, it

points to plurality of factors determining the character. All

the cases with which Walton has dealt in the paper under re-

view are cases of blending inheritance and as regards them it is

true, as already indicated, that continuous inbreeding tends to

the production of a more varied population (but not of more

variable separate lines) whereas cross-breeding tends to produce

a less variable population (devoid of differences between fami-

lies) but nevertheless a population more variable than the single

lines of a self fertilizing or constantly inbred population. |

W. E. CASTLE

BUSSEY INSTITUTION

NOTES AND LITERATURE

MIMICRY IN BUTTERFLIES

American biologists have been somewhat in a quandary of

late as to what to believe and to teach about ‘‘mimicry’’ in in-

sects. The consideration of chance resemblances in animate and

inanimate things in which mimicry in the strict sense could not

possibly exist, and the widespread skepticism of natural selec-

tion as an effective, creative agency in evolution have made many

of us inclined to bury mimicry in the same grave with telegony,

prenatal influences, the inheritance of acquired somatic charac-

ters, and sexual selection. Meanwhile, the Oxford school of

zoologists, under Professor Poulton’s leadership and the inspira-

tion of an orthodox faith in the potency of natural selection,

have continued to accumulate a rich array of newly discovered

models and mimics among African butterflies.

Many of these and other cases of mimicry are described in the

opening chapters of Professor R. C. Punnett’s interesting book

and admirably portrayed in the sixteen plates, twelve of which

are in colors. With a remarkably clear and convincing style

that has become familiar to us through his popular little book

on Mendelism Punnett here recounts the history of the theories

of Bates and Miiller, mentions some of the morphological fea-

tures upon which real affinities among butterflies depend, and

describes in some detail examples of mimicry from various parts

of the world.

Of particular interest to us in the United States is his brief

discussion of the supposed mimicry of Papilio philenor by P.

troilus, by the black southern variety, usually called glaucus,

of the female of our common turnus, and by a third species,

P. asterius (usually known by us as P. polyxenes, or P. asterias).

The northward extension of the range of troilus into Northwest

Canada, far beyond that of the supposed model philenor, is

thought to weaken this as a case of mimicry, and the author con-

cludes that

1‘*Mimiery in Butterflies,’’? by R. C. Punnett, Cambridge Univ. Press,

1915, 8vo., pp. 159, 16 plates.

Punnett transposes these names, following Poulton (vide Annals Entom.

Soc. America, Vol. 2, 1909, p. 225), who adopts Rothschild and Jordan’s

revision.

184

No. 591] NOTES AND LITERATURE 185

On the whole it seems at present doubtful whether any relation of a

mimetic nature exists between P. philenor and these three species of

Papilio.

The blue female of the southern fritillary, Argynnis diana, and

our ‘‘red-spotted purple,” Limenitis (Basilarchia) astyanaa,

which Professor Poulton has conceived also to be mimics of P.

philenor, are likewise regarded as ‘‘ very problematically mimetic. H

The striking resemblance of our ‘‘viceroy,’’ L. (B.) archippus,

for the ‘‘monarch,’’ Danais (Anosia) plexippus, is mentioned,

though no allusion is made to Abbott’s biometrical study of 87

specimens of the supposed ancestral type, L. (B.) arthemis, from

which the mimic, archippus, is thought to have arisen. Abbott,”

by the way, found that the color markings involved in the Poul-

ton hypothesis of gradual change by natural selection (e. g.,

reddish spots) are much less variable than the blues and other

colors not considered in that theory, the color pattern of arthemis

showing no tendency to break up or to shift in the direction of

the Anosia type.

Punnett next examines critically Wallace’s well-known laws

or conditions of mimicry, discusses the evolution of a Ceylonese

‘“‘mimicry ring” (a group of five superficially similar butter-

flies), describes the case of Papilio polytes, the trimorphic

‘‘mimetic’? and ‘‘non-mimetic’’ females of which are geneti-

cally separated from one another by two Mendelian factors, con-

siders the enemies of butterflies, and, finally, the relation of

seasonal and local variation to mimicry. He arrives at the gen-

eral conclusion that there are two prominent difficulties in ‘‘ac-

cepting the mimicry theory as an explanation of the remarkable

resemblances which are often found between butterflies belong-

ing to distinct groups,” viz., ‘‘the difficulty of finding the agent

that shall exercise the appropriate powers of discrimination,

and the difficulty of fitting in the theoretical process involving

the incessant accumulations of minute variations with what is at

present known of the facts of heredity.’’* In view of these diff-

culties, taking his cue from genetics, he suggests that

Each group of Lepidoptera contains, spread out among its various

members, a number of hereditary factors for the determination of color

pattern. . . . Some factors may be common to two or more groups, in

which case some of the permutations of the factors would be similar in

the groups and would result in identical or nearly identical pattern.*

* Washington Univ. Studies, Pt. 1, No. 2, 1914.

P. 139,

*Pp. 145, 146,

186 THE AMERICAN NATURALIST (Von. L

Thus, referring by way of illustration to the somewhat analogous

ease of the coat colors of rodents, he says:

In certain features the rabbit might be said to “mimic” the mouse,

in other features the guinea-pig.

It is a significant fact in this connection that the various models

‘‘mimicked’’ by the different species of a polymorphic species

are almost always closely related, and hence may be expected

to exhibit color patterns based on different combinations of iden-

tical factors.

In criticism of Wallace’s laws of mimicry, Punnett points out

the fact that although the mimic and model usually occur in the

same locality this is not always the case, the cooperation of mi-

gratory birds being invoked to explain the exceptions.

Regarding the defenselessness of mimics as compared with

models, it is noted that the ‘‘mimic’’ is often a swifter flyer, and

hence better prepared for defense than the model.

Exceptions are given to the rule that the models are more

numerous than the mimics, and that the mimics differ from the

most of their nearest allies. The Pierid genus Dismorphia, for

example, includes prominent South American mimics which

differ strikingly from the ‘‘whites’’ of the Temperate Zone but,

unfortunately for the theory of mimicry, only about a dozen

of the seventy-five described species are white, the rest present-

ing a ‘‘wonderful diversity of color and pattern.” Among them

are species clearly non-mimetic as regards color, which by simple

substitution of one color for another in the spots would be trans-

formed into a ‘‘mimetic’’ species.®

The author concludes that

It is on the whole unusual to find cases where a single species departs

widely from the pattern scheme of the other members of the genus and

at the same time resembles an unrelated species.

Two of the best examples are our American ‘‘viceroy’’ and the

pierid Pareronia. ‘‘Mimicry tends,’’ he adds, ‘‘to run in certain

groups” and ‘‘in many cases at any rate little meaning can be

attached to the statement that the imitators differ from the bulk

of their allies.’’

5 The reviewer recently observed in Porto Rico a case bearing upon this

point, in Leptalis (Dismorphia) spio, which closely resembles in color ana

general shape the very common Heliconius charitonius, A color variety ot

the former, however, is found in certain localities on the island, in which

orange replaces yellow § in the color pattern, rendering the resemblance to the

Heliconian less apparent. A simple mutation of orange into yellow woula

make this an excellent example of ‘‘mimicry.’’

No. 591] NOTES AND LITERATURE 187

In the chapter entitled ‘‘Mimicry Rings’’ the author considers

the difficulty of explaining the protective value of the minute

initial variations in the direction of a model. As an illustra-

tion, a group of five superficially similar butterflies in Ceylon is

described. This ‘‘mimicry ring’’ includes two hypothetically

distasteful Danaines (D. chrysippus and D. plexippus) and the

females of three very unlike males (Hypolimnas misippus, Elym-

nias undularis, and Argynnis hyperbius). The coloration of one

of these males (E. wndularis) is a.deep purple brown, like that

of ‘‘satyrs’’ generally. If this represents the original type from

which the gay orange and black pattern of the female has been

derived, how has the change come about? Slight initial varia-

tions of the Satyr in the direction of the orange Danaine could

not possibly be mistaken by birds for the model. The absurdity

is pointed out of assuming, on the other hand, that the Danaine

was originally like the male Satyr, and acquired its warning

coloration pari passu with the mimic, for the Danaine model

can hardly have been originally like all of the three very di-

versely colored males of the mimicking females. Mutation in

each of the three types, however, may have produced females

so similar to the Danaine as to be mistakable for it, and if nat-

ural selection indeed operates in this case, it may act in ‘‘putting

on the finishing touches,’’ or in preventing regression.

In the two following chapters the author discusses the resem-

blance of two of the three varieties of female Papilio polytes to

the two ‘‘poison-eating’’ Papilios of India and Ceylon, P. aristo-

lochiew and P. hector. As is well-known, Punnett? has himself

studied in Ceylon the behavior of these species, and Freyer” has

continued the work, making extensive breeding experiments on

the polymorphic ‘‘mimie.’’

A study of the geographical distribution in this case shows a

general correspondence between the range of each mimic and its

model, but notable differences are discovered. Regarding the

value of the resemblance between mimic and model, Punnett

had no difficulty in distinguishing between model and mimic on

the wing, even at a distance of forty to fifty yards, while near

at hand the brilliant scarlet of both models, which covers the

body and is conspicuous in spots upon the wings, is seen to be

very different from the softer red found upon the wings (not

e “<Spolia Zeylanica,’’ Vol. 7, Part 25, ge

T Phil. Trans. Roy. Soc., London, Vol. 204, 1

8 Vide: Lutz, AMERICAN NATURALIST, Vol. a .. 190, 1911.

188 THE AMERICAN NATURALIST [Vou. L

upon the bodies) of the mimics. Dried specimens of models

and mimics are likely to be confused, but not the living butter-

flies.

Freyer’s breeding experiments bring out the fact that a

simple Mendelian relation exists between the three varieties of

female in P. polytes, the males of which, though phenotypically

alike, correspond genotypically to the three kinds of female. Of

these three the one resembling the male (‘‘non-mimetic’’) is reces-

sive to the mimetic forms, lacking a factor, X, possessed either

in simplex or duplex condition by the ‘‘mimetic’’ females. The

male likewise is latently either xx, XX, or Xx, as the case may

be, but retains a uniform appearance in all cases owing to the

presence of an inhibitor factor for which he is heterozygous

(li), the female being recessive (ii). The male, unlike other

Lepidoptera, so far as they have been investigated, is also sup-

posd to be heterozygous for a sex factor which we may for brev-

ity call M, responsible for maleness, with which the inhibitor

factor is completely coupled, so that the male-producing sperms

(MI) always contain the inhibitor factor, the female-producing

always lack it (mi).

The two mimetic varieties of female are differentiated from

each other by the presence or absence of another factor, Y,

which acts merely as a modifier of the factor X when that is

present, and transforms the aristolochie-like female (XXyy or

Xxyy) into the hector-like (KXYY , XXYy, XxYY, or XxYy).

This color modifier, responsible for intensifying and extending

the red markings, is supposed to occur in either homozygous or

heterozygous condition, or to be absent (recessive) in the male-

like form of female and also in each biotype of the male, though

when present without X, or in the presence of I, it has no visible

effect. Thus there are 9 biotypes of males and 3 of male-like

females, all phenotypically alike. Referring to Poulton’s theory

of the gradual evolution by natural selection of the male-like

type of female into the aristolochiw-like, and subsequently into

the hector-like, Punnett argues that crossing the hector-like

(double dominant) with the male-like (double recessive) germ

plasms and inbreeding should show the hypothetical intermediates

postulated by Poulton, but no such intermediates have appeared

in the breeding experiments.

Freyer’s random sampling of the population of polytes gave

49 of the two mimetic females to 40 of the male-like coloration,

or roughly 5 dominants to 4 recessives, a proportion indicating

No. 591] NOTES AND LITERATURE 189

stable equilibrium between the mimetic and non-mimetic varie-

ties. Scanty historical data tend to show that the mimics were

as common fifty years ago, and probably a century and a half, as

to-day, so the author concludes ‘‘that in respect of mimetic re-

semblances natural selection does not exist for P. polytes in

Ceylon,’’ or at least there is ‘‘no effect appreciable to the neces-

sarily rough method of estimation employed.’’

The author next considers the evidence that the enemies of

butterflies could have played the part assigned to them by the

advocates of the mimicry theory. Predaceous insects evidently

pay no attention to warning colors; certain lizards devour butter-

flies freely, but do not exercise any discrimination in the species

which they attack. Hence neither insects nor lizards can be sup-

posed to play any part in establishing a mimetic resemblance.

Birds destroy butterflies in considerable numbers, but

Some of the most destructive appear to exercise no choice in the

species of butterfly attacked. They simply take what comes first and is

easiest to catch.

Monkeys and baboons often eat butterflies. They show strong

likes or dislikes for certain species. The monkey may be re-

garded as ‘‘the ideal enemy for which advocates of the mimicry

theory have been searching—if only it could fly.” The con-

clusion is reached that

even a slight persecution directed with adequate discrimination will in

time bring about a marked result where the mimetic likeness is already

in existence. It is not impossible therefore that the establishing of such

a likeness may often be due more to the discrimination of the monkey

than to the mobility of the bird.

In the final chapter on ‘‘Mimicry and Variation”’ the author

describes Carpenter’s observations on the polymorphic mimic

Pseudacrea eurytus, the four forms of which show an extraordi-

nary resemblance to acræine ‘‘models’’ of the genus Planema.

These butterflies inhabit the shores of Victoria Nyanza in Cen-

tral Africa where the models are very abundant, the polymorphic

mimics less common but still numerous, and intermediates be-

tween the different types of mimic rare, but not unknown. On

Bugalla Island in the lake, on the contrary, the mimetic Pseuda-

creas are very abundant, and their respective Planema models

relatively rare. Here intermediates between the varieties of

the polymorphic mimic occur in proportionately larger numbers

than on the mainland, owing as Dr. Carpenter believes to the

190 THE AMERICAN NATURALIST [Vou L

cessation of natural selection in the absence of sufficient models

to familiarize the hypothetical enemy with the several warning

color patterns of the models. On the mainland, however, any

of the aberrant intermediates that might be produced by inter-

breeding of the different varieties of the polymorphic species

would meet an enemy having constant experience with the warn-

ing colors of the different models, and tend to be eliminated.

The enemy, in other words, would avoid the perfect mimics,

while aberrant individuals suggesting two models at once pre-

sumably would be attacked and eaten. This interesting case de-

serves thorough investigation.

The author makes a faux pas when he says regarding seasonal

variations in butterflies, due to ‘‘changes in the conditions of

later larval and earlier pupal life’’

Dr no case are they known to be inherited, and in no ease consequently,

could variation of this nature play any part in evolutionary change.

In the example cited (Araschnia levana-prorsa), which pre-

sents two color patterns alternately through the year, it is

obvious that both patterns are inherited. The environment in-

deed decides which shall appear, but the hereditary basis common

to both seasonal types is no less real than that of any butterfly

of seasonally uniform pattern. Although the seasonal color

patterns of A. levana and prorsa apparently can not behave as

Mendelian allelomorphs to one another as do the color patterns

of other non-seasonal polymorphic insects, they are by no means

outside the pale of Mendelian heredity. It is not too much to

expect that future studies will disclose colors or color patterns

allelomorphic to A. levana-prorsa’s shifting coloration. The re-

viewer would not, with Professor Punnett, rule seasonal varia-

tion entirely out of court as possible stages in ‘‘the develop-

ment of a mimetic likeness’’ or, rather, in the evolution of the

remarkable likenesses, alleged to be mimetic, which this book

brings so well to our attention.

The author is so strongly influenced by the idea that minute

variations are fluctuations always controlled by the environ-

ment rather than by the internal conditions that result in he-

redity that he treads upon uncertain ground in discussing an ex-

ample of local variation cited by Poulton.2 A small white spot

on the wing of Danais chrysippus varies in size locally from a

conspicuous marking, in China, to a faint dot tending to dis-

® Bedrock, 1913, p. 300. Cited by Punnett.

No. 591] NOTES AND LITERATURE , 191

appear, in Africa. Punnett suggests that the details in pattern

may be in slight measure affected by the plant species on which

the caterpillars have fed, thereby producing local races. Trans-

portation of a local race to a region inhabited by another dis-

tinct local race ‘‘would help us in deciding whether any varia-

tion by which it is characterized had a definite hereditary basis

or was merely a fluctuation dependent upon something in the

conditions under which it had grown up.’ We may well ask:

are these two propositions mutually exclusive? May not a de-

tail of color pattern to a certain degree at certain times be sub-

ject to environmental influences and at the same time may not

its variations have a ‘‘definite hereditary basis?’’ The re-

viewer has had so much experience in observing the transmis-

sion in Colias philodice and C. eurytheme of spots comparable

to that mentioned in D. chrysippus that he is convinced that a

definite hereditary basis (consisting presumably of multiple `

factors) underlies every fluctuating detail of color pattern. By

artificial selection from inbred stock, using uniform food plants,

and exposing the caterpillars and pup# to similar conditions,

the breeder of butterflies may decrease even to elimination or

increase within certain limits a detail of color pattern like that

mentioned. The champions of the theory of mimicry are en-

titled to this crumb of comfort. ‘‘For if it can be supposed,’’

remarks the author, ‘‘that small differences of this nature are

always transmitted, it becomes less difficult to imagine that a

mimetic resemblance has been brought about by a long series of

very small steps.’’

Yet the facts which the mimicry theory seeks to explain clamor

for explanation. Punnett sets forth at the end some that are

most insistent.

Certain color schemes are characteristic of distinct geographical re-

gions in South America, where oes may occur in species belonging to

very different genera and families

In Central America a ET occurs that is common to sev-

eral Heliconines, Ithomiines, Nymphalines of two or more gen-

era, and Pierids; in eastern Brazil another pattern in which

“‘all the various genera which figure in the last group are again

represented.’’ On the upper Amazon a still different pattern is

common to the same group of genera from that just mentioned,

only two notable genera being absent. Finally in Ecuador, Peru

and Bolivia a widely different pattern occurs in a group lacking —

192 THE AMERICAN NATURALIST [Vou. L

a Pierid or Danaid but containing in exchange ‘‘a Papilio, an

Acrea, and two species of the Satyrid genus Pedalodes.’’

Assuming that one of these patterns must have been the most

primitive, he asks why a distasteful genus should change from

one efficient warning pattern to another quite distinct one.

Though the premise is not necessarily true nor even probable,

yet if the ancestral pattern were a generalized type of any

warning’’ character whatever, the question would still be per-

The author suggests that a newly acquired color scheme, like

one of these, may be ‘‘associated with a certain physiological

constitution which places butterflies possessing it at an advan-

tage as compared with the rest,’’ just as the melanie variety of

peppered moth that is ousting the typical form in Britain and

on the Continent may have associated with its deep pigmenta-

tion a greater hardiness. This, however, goes but a short way

toward the explanation of the extraordinary local associations of

unlike South American butterflies showing similar coloration.

This is a live question that challenges the attention of any stu-

dent of evolution who has opportunity to undertake experimen-

tal work in the tropics of South America.

So if we must sooner or later consign Mimicry to its last rest-

ing place, with its less infirm but already moribund parent,

Warning Coloration, let us do so filled with gratitude for the

pioneer work accomplished by its champions in opening up

promising fields of investigation, where we or our descendants

may hope to discover new factors in evolution or to gain a deeper

insight into those now only dimly understood. Thanks are

meanwhile due to the author of this attractive volume for his

keen diagnosis of the present condition of the mimicry theory

and for his admirable description of the phenomena which it has

attempted to explain.

JOHN H. GEROULD

VOL. L, NO. 592" APRIL, 1916

The American Naturalist

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

sent to the Editor of THE AMERICAN NATURALIST, Garrison-on-Huwson, New York.

Short articles containing summaries of research work bearin g on the

problems oe organic evolution are especially welcome, and will be given preference

in publica

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

E ikapa reprints will > supplied at cost

Subscriptions advertisements should be sent to the aeee The

bscription price is Ro dollars a year. Foreign postage is fifty cents and

Canetion postage twenty-five cents additional. The peta for single copies is

forty cents. The advertising rates are Four Dollars for a p

THE SCIENCE PRESS

Lancaster, Pa. Garrison, N. Y.

NEW YORK: Sub-Station 84

Entered d tter, April 2, 1908, at the Post Office at Lancaster, Pa., under the Act of

Congress of March 3, 1879.

B O Ol KS FOR S ALE | JAPAN NATURAL HISTORY SPECIMEN

Perfect Condition and Lowest Prices.

Specialty: Bird Skins, Oology, Entomology, Marine

NATURAL HISTORY Animals aad others. Catalogue free. Correspond-

Catalogue on ence soliéited. »

T. FUKAI, EASA ralist.

«oa JOHN D. SHERMAN, J misi ge

PUBLISHED BY THE CAMBRIDGE UNIVERSITY PRESS

| MIMICRY IN BUTTERFLIES |

Bound i in Buckram. ae 8vo. a vili+188. sia 16 plates (12 of which

are colored). Price $4.50

CF By TR a Pen, F. e Fellow of Gonville and Caius geek

q e DA e of Genetics in the University of Cambri

~ EXTRACT FROM THE PREFACE —

the hope that it may appeal to several

t im ai ad different callings in life

en pagead Bepsie Ai erei be reasonably short, w well ilas-

soa E cory and not too hard to Sepeni I have always

n obliged- ar ı that I know of nothing in our language answering t0 | —

descripti and if is lar gely as an attempt to remedy this deficiency that the z

I opo sn wl be ot finer Jo rer whe live i e or visit tro tropes a

l are attracted by the b z y or ory BiS poani haie f

THE

AMERICAN NATURALIST

Vou. L. April, 1916 No. 592

THE MECHANISM OF CROSSING-OVER

HERMANN J. MULLER

Rice INSTITUTE

Ir is the object of this paper to give an account of the

most important evidence thus far gained in regard to the

manner in which separation of linked factors—often inter-

preted as ‘‘crossing-over’’—takes place, and to describe

an experiment in which a new method for studying the

occurrence of such separation is employed. This experi-

ment is still under way, but as it may be a considerable

time before the results are obtained in full, it would not be

advisable to withhold longer an account of this work and

of other work that bears on the nature of the ‘‘crossing-

over.’’

I. Tus Discovery or INTERCHANGE BETWEEN HOMOLOGOUS

HROMOSOMES

The question as to whether separation of linked factors

is due to pieces of homologous chromosomes changing

places with each other, carries us back to the question

whether the factors lie in the chromosomes at all. As is

well known, there is a large body of evidence from cy-

tology and from experimental embryology, showing that

the chomosomes are persistent, self-perpetuating struc-

tures which have a profound influence upon development.

But the first definite evidence that the Mendelian factors

are contained in the chromosomes lay in the striking cor-

respondence which was found between their respective

193

194 THE AMERICAN NATURALIST [ Vou. L

methods of distribution—the segregation of Mendelian

_ allelomorphs exactly paralleling the pairing and separa-

tion of homologous chromosomes during the maturation

of the germ cells, and the random assortment of Men-

delian factors belonging to different, independently segre-

gating pairs, paralleling the random assortment of chro-

mosomes belonging to different pairs (Sutton). Still,

there was no indication of a connection between any par-

ticular chromosome and a particular character until the

work of McClung, Stevens and Wilson and others proved

that in many animals the ‘‘X-chromosome’’ contains, or

at least invariably accompanies, a factor for sex, inas-

much as all fertilized eggs which receive two X-chromo-

somes develop into females, while those with one X be-

come males.

The next time that particular factors and chromosomes

were found to be correlated was in 1910, when Morgan

pointed out that the factor in Drosophila determining

whether an individual shall have red or white eyes, as well

as Several other factors, must also be located in the X-

chromosome (or atleast must accompany it in its segre-

gation). For, to put the whole argument briefly, the fact

that a red-eyed male bred to a white-eyed female produces

red-eyed daughters and white-eyed sons shows that the

female-producing spermatozoa—those that receive the X-

chromosome—also receive the factor for red, but the

male-producing sperm—which do not receive the X— also

fail to receive red. In other words, the factor for red was

judged to be in the X-chromosome, because in the male it

is always distributed to precisely the same spermatozoa as

those to which the X’s happen to be distributed. Bridges

has recently obtained evidence that in the female, too, such

“*sex-linked’’? factors accompany the X-chromosome in

Segregation. Ordinarily there is no opportunity for at-

tacking this question in the female, since the female con-

tains two X’s, which are of course indistinguishable to the

eye, so that it would be impossible to tell whether or not a

particular one of the X’s was always distributed to the

No. 592 | THE MECHANISM OF CROSSING-OVER 195

same eggs as a particular sex-linked factor. But in

Bridges’s cases of non-disjunction, the maturation divi-

sions are often abnormal, so that some eggs are found to

have retained both X’s; and in accordance with this it is

found that some of the offspring of such females have like-

wise received two sets of maternal sex-linked factors.

These cases, therefore, show that in the female also the

sex-linked factors ‘‘follow’’ the X-chromosome (1, 2).

Morgan next studied the relation of different sex-linked

factors to each other in inheritance, and then another re-

markable fact came to light. Theoretically, the dihybrid

females resulting from a cross of a red-eyed fly having

rudimentary wings by a white-eyed fly having long wings

(both of these pairs of characters are sex-linked) should

have contained in one of their two X-chromosomes the

factors ‘‘red’’ and ‘‘rudimentary,’’ and in the homologous

X-chromosome the factors ‘‘white’’ and ‘‘long’’; the ma-

ture eggs should retain either one X or the other and

should therefore have contained either red and rudimen-

tary or white and long. In other words, red and rudi-

mentary should be completely linked in their inheritance,

and similarly white and long. But the results showed that

these factors sometimes separate in heredity, for not only

the above types of offspring are produced, but also some

red longs and white rudimentaries (7); in fact, about 42

per cent. of the offspring belong to one or the other of the

two latter classes. If we admit that white and long were

originally present in the same chromosome, the only way

to account for this separation of the factors is to suppose

that in some of the cells of the hybrid female the X-chro-

mosomes interchanged parts before being distributed to

theeggs. For if the factor for ‘‘long’’ of the chromosome

containing ‘‘white’’ and ‘‘long’’ should somehow change

places with the ‘‘rudimentary’’ of the homologous chro-

mosome, then when homologous chromosomes are sep-

arated at the maturation division, the egg may come to

contain either an X-chromosome with white and rudimen-

tary or an X with red and long. | |

196 THE AMERICAN NATURALIST [ Vo. L

Recent work on Drosophila has borne out in a striking

way the conclusion that the separation of factors just dis-

cussed is due to chromosomal interchange. It will be re-

membered that the pairs of factors in the example under-

went recombination in only about 42 per cent. of the eggs,

i. e., they held together more often than they separated,

and so might be said to be partially linked. Their mode

of inheritance therefore forms a contrast not only to com-

plete linkage, but also to the familiar cases of random

assortment, where two pairs of factors are found recom-

bined in about 50 per cent. of the offspring, and thus show

no linkage at all, presumably because they lie in different

pairs of chromosomes which segregate independently.

Further investigation showed that not only ‘‘white’’ and

‘‘ rudimentary,” but all the known sex-linked factors, in-

stead of segregating independently, are ‘‘partially linked”?

to one another in greater or less degree. This then was

additional evidence that these factors did not lie in differ-

ent pairs of chromosomes, as in familiar cases, but in the

same pair of chromosomes, and that their separation or _

recombination was therefore dependent upon chromo-

somal interchange. But furthermore, if these linked fac-

tors all lie in the X-chromosome (being sex-linked), then

it might be expected that other groups of interlinked fac-

tors also would be found, that lie in other chromosomes.

A factor in any one of these other groups would not be

sex-linked, but would be linked in greater or less degree

to every other factor of the same group, since it lay in the

same chromosome with it, but it would undergo 50 per

cent. of recombination with factors in other groups. This

expectation has been fulfilled. In 1911 Morgan and Lynch

found two pairs of factors in Drosophila(black versus gray

body color; vestigial vs. long wings) that were linked to

each other, but that were not sex-linked (10). These were

designated as lying in group II or Chromosome II. Later,

Sturtevant found that two other pairs of factors (pink vs.

red eyes and ebony vs. gray body color) were also linked

to each other, but were neither sex-linked nor linked to the

No. 592} THE MECHANISM OF CROSSING-OVER 197

other non-sex-linked group; these were assigned to Group

III or Chromosome III (15). Incidentally, it was evi-

dent that these cases are of exactly the same nature as

those previously discovered by Bateson and Punnett in

the sweet pea, and termed by them ‘‘coupling”’ or ‘‘re-

pulsion.’’ Moreover, the chromosome interpretation

made it clear why the factors should be ‘‘coupled”’ or ‘‘re-

pelled’’ according to whether the hybrid received them

from the same or from opposite parents. There was only

one difference in detail between the facts in the two spe-

cies: it was discovered by Morgan that in Drosophila the

linkage is always complete in the male, the separation of

factors that are linked to each other occurring only in the

female (9); in the plants, on the other hand, recombi-

nation occurs in the genesis both of eggs and of sperm.

Since that time the inheritance of over one hundred

pairs of factors of Drosophila has been studied. This

investigation should give an extensive experimental test

of the theory of chromosomal interchange, for if linked

factors are those carried by the same chromosome, there

should be the same number of groups of interlinked fac-

tors as there are pairs of chromosomes. There are four

pairs of chromosomes in Drosophila—two pairs of long

ones, the pair of moderately long sex-chromosomes, and

a pair of very small chromosomes. By 1913, work had

been done upon a large number of factors, and the results

showed that all these factors were linked in one of the

three groups already discovered. But in 1914 the author

found a pair of factors independent of these (bent vs.

straight wing), i. e., constituting Group IV (12), and not

long afterwards Miss Hoge found another pair of factors

in this fourth group (eyeless vs. normal eye), (3). Ac-

cordingly, the number of groups and of chromosomes now

correspond, and not only that, but the relative sizes of the

groups correspond in a general way with the relative

lengths of the chromosomes. Can it be mere chance that

one hundred factors fall into this particular grouping?

But if it is admitted that these groups are carried in the

198 THE AMERICAN NATURALIST [ Vou. L

chromosomes, then, as above pointed out, the separation

of factors in a group means chromosomal interchange.

Il. A MECHANISM oF INTERCHANGE ALREADY PROVIDED BY

THE THEORY OF Crosstnc-OvER (CHIASMATYPE)

Janssen’s ‘‘chiasmatype theory,’’ based on cytological

observations of spermatogenesis in Batracoseps, described

just such a process of interchange between the homologous

chromosomes as Morgan’s evidence from genetics required

(4). Agreat bulk of evidence has accumulated to show that

during the period of synapsis, homologous chromosomes

come into contact, and in many cases chromosomes can be

seen to be twisted around each other during one stage or

another of synapsis. The essential point postulated by

the chiasmatype theory is that, as the paired chromosomes

draw apart again, they do not always untwist completely,

but may break at some points where they are crossed—

thus, in Fig. 1, the upper piece of the light-colored chro-

Fie. 1.

mosome (L), which was on one side, loses its connection

with the lower part of L, that has crossed to the other side,

but becomes united instead with the lower piece of the

dark chromosome (M) which, on account of the crossing

of the strands, now follows it on the same side; similarly,

of course, the upper part of M becomes united with the

No. 592] THE MECHANISM OF CROSSING-OVER 199

lower part of L; in this way a recombination of parts is

accomplished. Morgan and other workers on Drosophila

base their acceptance of this essential point in Janssens’s

chiasmatype theory upon the evidence (from cytology)

that homologous chromosomes do twist about each other

during synapsis, taken together with the evidence (from

genetics) that these chromosomes emerge as new combina-

tions. Janssens, on the other hand, maintains that certain

details in the appearance of the chromosomes during that

stage in synapsis called ‘‘strepsinema’’ give ocular evi-

dence that crossing-over occurs at this particular period

and in a particular manner. As it would seem possible,

however, to put another interpretation upon his figures,

this question may be deferred until later.

Janssens had intended the chiasmatype theory to explain

the supposed fact that there might be more pairs of factors

T era a=

4 Qs

quam:

capable of recombination than there were pairs of chro-

mosomes. (It might be mentioned in passing, however,

that at that time this fact had not yet been demonstrated ;

there are even now probably no published facts except

those recently discovered in Drosophila which prove this

point). As shown above, Morgan went further than this

with the chiasmatype theory by applying it to explain, spe-

cifically, the recombination of linked factors (8). More-

over, he pointed out at the same time an important corol-

Fig. 2.

200 THE AMERICAN NATURALIST [Vor. L

lary to this theory. It has already been stated that he had

found different degrees of linkage to-exist between the

various factors of a group: for example, the proportion

of cases in which separation occurs between white and

rudimentary was said to be 42 per cent., whereas the fre-

quency of separation between white (eye color) and the

factor for yellow body color is only about 1 per cent. In

explanation of these different degrees of linkage, Morgan

pointed out that, on the chiasmatype theory, the closer the

proximity of two factors to each other in the chromosome,

the smaller would be their frequency of separation, for

the less would be the chance for a crossing-over of the

chromosomes to oceur between them. Thus, in Fig. 2, the

factors A and C become separated both in case (a) and

in case (b), because A and C lie so far apart that in both

cases the point of crossing-over falls between them, but in

only one of the cases do A and B separate, and in one case

B and C, since these are so near together that the point of

crossing-over may often be beyond instead of between

them. In other words, on the chiasmatype theory, fre-

quency of recombination must be, to a certain extent at

least, an index of the distance apart of factors along the

chromosome. Since the time when these ideas were pro-

posed (1911), two important series of facts have come to

light in the studies on Drosophila, in support of the chias-

matype theory of interchange and of these extensions of it.

III. A VERIFICATION oF THE THEORY oF CrRossrnc-OVER.

Tue Law or LINEAR LINKAGE

Ir occurred to Sturtevant in 1911 that, if the factors

are carried in the chromosomes, then, owing to their linear

arrangement, the distance along the chromosome between

any two factors (A and C) must be either the sum or the

A Ces A B ©

eee ey D aaa O

20 20 lo

Fre. 3.

No. 592] THE MECHANISM OF CROSSING-OVER 201

difference of their distances from any third factor (B) of

the same group, îi. e., length AC = length AB + length BC,

the + or — depending on whether the third factor is be-

tween or beyond the other two (see Fig. 3). Accordingly,

if, as Morgan suggested, the frequencies of separation

(linkage values) between factors depend on their distances

apart, then the frequency of separation (degree of link-

age) between the two factors, A and C, should be predict-

able, given the frequency of separation of each from the

third factor, B. To put the matter diagrammatically, A,

B and C have been represented in figure 3 as points along

a line; A and B, we will suppose, separate from each other

in heredity in 20 per cent. of cases, to correspond with

which they have been placed the same number of units (20)

apart in the diagram; similarly, B and C, which we will

suppose to separate 10 per cent. of the time, have been

placed 10 units apart. (As above pointed out, there are

obviously two possible diagrams to choose between, de-

pending on whether C is beyond A and B or between

them.) Then, if it be true that the frequency of separa-

tion between any factors is always precisely proportional

to their distance apart, it will follow that the per cent. of

separations between A and C will be equal to the number

of units of distance on our diagrammatic chromosome be-

tween A and C; this in turn equals AB + BC = 30 or 10.

If separation frequency bears a less simple relation to dis-

tance, but is nevertheless determined by it (see below),

frequency AC will not equal distance AC (i. e., AB + BC)

but can be calculated from the latter. On the other hand,

if our premises are false, and there is no linear relation at

all between the factors that determine their frequency of

separation, then frequency AC will not be equal to dia-

gram distance AC (i. e., to AB+ BC), nor even, in the

case of different sets of factors, will it bear any constant

relation to diagram distance AC; that is, it would not be

possible to discover any constant rules for calculating the

third frequency from the two others which will hold, even

approximately, for various sets of factors (BCD, LMN,

ete.). eee

202 THE AMERICAN NATURALIST [Vou L

Sturtevant found that there is indeed a linear relation in

the frequencies of separation (14, 16). In the case of

smaller per cents. of separation, per cent. AC always is

precisely equal to the sum or difference of per cents. AB

and BC (within the limits of probable error), so that the

per cents. of separation for all combinations of these fac-

tors is accurately represented by a linear diagram. In the

case of higher per cents. of separation(long distances), the

highest of the three frequencies (let us call it AC) falls

short of the sum of the other two (AB + BC), and so it is

a smaller number than the distance representing it on the

diagram, but it nevertheless (within the normal limits of

error) can be calculated from this diagram distance AC,

for a constant relation was discoverable between this hy-

pothetical distance and the actual frequency. Thus the

different frequencies do not bear any random relation

to each other that is mathematically possible, but bear

relations that disclose a linear connection between the

factors.

It remains to consider the meaning of the fact that in

cases where there is a high per cent. of separations, the

highest per cent.—that between A and ©, let us say—is

not as great as the value of the distance AC representing

it on the diagram, i. e., it is less than the sum of the per

Fic. 4.

cents shown between A and B, and B and C, respectively.

If, whenever A and B or B and C separate, A and C sep-

arate also, as shown in Fig. 2, (a) and (b), then per cent.

AC would be equal to per cent. AB plus per cent. BC, but

since per cent. AC is lower than this, there must be cases

in which, although A and B or B and C separate, A and C

fail to separate. It is obvious that in these cases, where B.

separates from A but A does not separate from C, that B

No.592] THE MECHANISM OF CROSSING-OVER 203

must have separated from C also, i. e., a separation has

occurred between A and B, and between B and O, coinci-

dentally. On the view above presented, a separation

means a crossing-over of chromosomes, and so in these

cases the chromosomes must be thought of as crossing-

over at two points coincidentally, as shown in Fig. 4. This

process has been named by Sturtevant ‘‘double crossing-

over.” As shown in the figure, where crossing-over oc-

curs coincidentally, both in AB and in BC, the chromosome

crosses and crosses back again between A and C, hence the

latter factors do not become separated.

When the frequencies of separation (diagram distances)

between A and B, and between B and C are both small, it

is to be expected as a matter of pure chance, if the factors

are joined in line in the manner described, that such coin-

cidences will occur very rarely, even in proportion to the

small frequencies involved, and so the per cent. of separa-

tions between A and C will be practically as great as that

between A and B plus that between B and C. Hence per

cent. AC will be accurately represented by diagram dis-

tance AC. On the other hand, if separation is frequent

between A and B or between B and C, there should be

more chance of coincidence of these separations, and the

number of separations between A and C will fall corre-

spondingly short of AB + BC, which is the value of AC

shown on the diagram. Consequently, in predicting fre-

quency AC on the basis of AB and BOC, allowance must be

made for these coincidences. But the author ventures to

point out that, as the number of these coincident separa-

tions is found to be largely determined by the frequency

of separation, greater frequencies being accompanied by

a larger proportion of coincidences, as has been shown,

then the amount of allowance to be made can be approxi-

mately calculated for any given distance on the diagram;

accordingly, the frequency of separation between A and C

can be calculated from AB + BC, i. e., from the distance

AC on the diagram. The precise manner in which coin-

cidence of separations increases with their frequency is a

204 THE AMERICAN NATURALIST [ Von. L

question which will be reserved until later. But it is

clear that, since coincidence does not vary independently

of ‘‘distance,’’ a linear relation holds between the linkage

values, in that these values can be calculated from a

linear diagram much more exactly than would be ex-

pected on chance relationships; in mathematical terms,

the frequencies of separation between all combinations of

the different factors in a group are largely a function of

the distance apart of these factors in a linear figure.

It will now be desirable to consider these same facts

from another angle. As it is possible to represent the

linkages between any three factors of a group in terms of

their distances on a linear diagram, it follows that all the

factors of a group together can be represented in one

linear diagram. Suppose that such a diagram has been

made, and that the order of the factors in it is ABCDEFG.

Now, as has just been explained, since per cent. AC nearly

equals per cents. AB plus BC, it must follow that a separa-

tion between A and B rarely coincides with one between B

and C; the same fact may also be expressed by saying that

when A and B separate, C stays with B rather than with

A. Similar relations, of course, hold for the other fac-

tors, too; thus D also stays with B and C when A and B-

separate, but it stays with C when B and C separate. The

linkage of D with B, then, is only due to its linkage with

C, for, although it usually stays with B, it very rarely

stays with it except when C does. Thus D is linked to B

only through C, and to A only through B. Similarly, all the

other factors also are linked together in a chain, each to

the one on either side: just as D is linked on the one hand

to E, and on the other hand to C, but not to any other fac-

tors except through one of these, so C is linked on the one

hand to D, on the other hand to B, but is linked to E only

through D, and to A only through B, ete.; moreover, all

the factors are linked to the others in the same order.

Separation of factors in such a group accordingly means

the breaking of the chain at just one or two points, for it

has been pointed out that when B and C separate, A and

No. 592] THE MECHANISM OF CROSSING-OVER 205

B rarely separate coincidentally, but usually remain to-

gether, and C, D, E, etc., all remain together also, separa-

ting in a body from A and B. In other words, the factors

are not interchanged singly, but stay together in sections,

according to their positions on the diagram, and whole

sections are exchanged at once.

It may be objected that these conclusions are in many

cases based on linkage values obtained in different experi-

ments; that it is unwarranted to conclude, for instance,

that when C and D separate, E remains with D, simply

from the fact that in one experiment the frequency of

separation between C and D had the value m per cent., in

a second experiment DE was n, and in a third experiment

CE was m + n, for this conclusion would only be true on

the supposition that in all three experiments each factor

had the same frequency of separation with each of the

others as in the particular experiment where that fre-

quency was determined. The answer is that numerous

experiments have been performed in which three pairs of

factors (or more) could be followed at the same time, and

these experiments have given results precisely the same

in kind, although more accurate than the preceding. But

in experiments of the latter type, the coincidence of the

various separations and non-separations does not have to

be calculated out as in the case above, but is given directly

_ by the results. Thus in a hybrid which has received ABC

from one parent and the allelomorphs abe from the other,

gametes in which coincident separation between B and A

and between B and C has occurred will be distinguishable

by having either the composition aBe or the composition

AbC (see Fig. 3), and the number of such offspring can

thus be directly counted instead of it being necessary to

caleulate them from the relations between separation

values for A and B, B and C, and A and C. And in the

experiment of the author’s given in Section V, where the

inheritance of a large number of factors is followed simul-

taneously, the results show directly and graphically that

206 THE AMERICAN NATURALIST [Von. L

the factors, as arranged in line in order of their linkage,

are exchanged in whole sections at a time.

In the first section, evidence was presented, showing

that groups of factors are connected with particular chro-

mosomes, and segregate with them at the maturation divi-

sions; this was in fact proved to be true in the case of

sex-linked factors, which are found always to segregate

with the X-chromosome during spermatocyte divisions.

Yet it was conceivable that the factors were not actually

in the chromosomes, but rather tied to them by some

obscure connection (chemical, physical or metaphysical),

although the fact that the relative sizes of the groups cor-

respond to the lengths of the chromosomes might be taken

as evidence against such a view. On that view, a separa-

tion of linked factors would be considered not a physical

interchange between the chromosomes themselves, but a

transference, by a factor, of its invisible bond, from one

chromosome to the homologous one. But Sturtevant’s

evidence just presented shows that however one may have

conceived, a priori, the chemical attraction or physical-

connection that makes linked factors tend to segregate to

the same pole in the maturation divisions—this connec-

tion binds them in a linear manner, one after another, in a

chain. This unique result, then, constitutes specific evi-

dence that the factors are actually in the chromosomes, in

an order which can be determined by their linkage rela-

tions, and that the separation of linked factors is conse-

quently a real interchange between parts of the chromo-

somes themselves.?

1 The fact of linear linkage does not connote that frequency of crossing-

over is necessarily entirely dependent upon distance, for it is still possible

to escape the conclusion that crossing-over occurs equally “BS in all parts

of the chromosome, by assuming that coincidence of separations A-B an

B-C usually oceurs with not very different frequency from oa aa of

separations G-H and H-I, even if different actual lengths are involved,

in the chromosomes in the same order as on the dia iagram. In fact, no

matter how great the differences in frequency of crossing-over in -different

parts of the chromosome might be, the linkage order of the factors would

still be the same as their real order so long as coincident crossing-over in

any two regions did not occur as often as single crossing-over in either region.

No. 592] THE MECHANISM OF CROSSING-OVER 207

Furthermore, admitting this conclusion that the linkage

diagrams really represent the chromosomes, the fact that

the factors are exchanged in sections proves that whole

pieces of the chromosomes change places at once, as oc-

curs in the process of ‘‘crossing-over’’ postulated by

Jannsens, instead of small parts or factors in the chromo-

somes being separately exchanged. The idea that the

interchange during synapsis may be a kind of exchange

of separate particles from one container to another seems

to have been held by a number of geneticists. On this

view, the chromosomes might be considered as a sort of

pod, containing the factors within them like so many

beans; when the chromosomes synapse, the pods open

towards each other, so that a factor in one might change

places with a factor in the other. Conceivably—if we

adopt this view—certain factors might be harder to dis-

lodge than others, and so different frequencies of separa-

tion would exist between different factors. But such a

mechanism of interchange would not result in a mode of

linkage that may, in the sense explained above, be called

linear, for separation of factor B from A would, on this

mechanism, have no influence at all on whether or not C

separated from A. This difficulty could be partially met

by supposing that interchange of one factor in some way

facilitates interchange of the neighboring factors, but the

type of linkage which is actually found goes much further

than this, and shows that the whole group of factors re-

mains intact except at one or two points, interchange being

in two or three entire sections. This can only mean, then,

that interchange is a process of crossing-over, if it occurs

by means of synapsis at all.

It might be claimed, however, that this ae of

whole sections of the chromosomes need not occur at syn-

apsis, and therefore need not be of the nature of crossing-

over at all. The only alternative to crossing-over, how-

ever, would be to suppose that, during the resting period

of cells, the chromosomes might break up into pieces, and

that then, in reuniting, a fragment of one chromosome

208 THE AMERICAN NATURALIST [ Von. L

might become joined with a piece of the homologous chro-

mosome instead of with a piece of the same chromosome.

But on the fragmentation theory it must be supposed that

the fragments reunite in exactly the original order, and,

further, that the two homologous chromosomes break at

precisely the same point before interchanging— otherwise

one reformed chromosome would lack certain factors and

the other would have too many; nevertheless, this point

can not be a fixed point, as interchange may occur any-

where. Since interchange, when it occurs, usually takes

place at one point only, it must also be assumed that the

frequency of the recombination just described is so nicely

regulated that in about half of the cases it has happened

just once (and at one point in the chromosome) during

the sum total of resting periods of all cells ancestral to

any particular egg cell that shows interchange. For in

about half the eggs a particular chromosome has ex-

changed in only two sections, and in very few have there

been more than three points of interchange. Moreover,

in the ancestry of the rest of the eggs, no interchange

whatever can have occurred. Finally, the fallacy of the

fragmentation idea becomes obvious when we consider

that if interchange took place in the resting period of an

embryonic cell, most of the eggs derived from this cell

would show that particular recombination, and hence the

individual in question would give an unusually large pro-

portion of offspring of this sort. Thus different individ-

uals of the same strain would differ greatly in their link-

age values, there being scarcely any constancy at all.

Since this is not true it would have to be assumed that

interchange takes place only a short time before the mat-

uration divisions, owing to some peculiarity in the chro-

mosome processes occurring in the cells at this period.

Thus we return again to the conclusion that interchange

occurs during synapsis.

Further evidence that interchange occurs during syn-

apsis is to be found in some results obtained with Bridges’

“‘non-disjunctional’’ flies. Non-disjunctional females of

No. 592] THE MECHANISM OF CROSSING-OVER 209

Drosophila contain, besides their two X-chromosomes, a

Y-chromosome (owing to previous mitotic abnormalities).

The presence of the extra homologous chromosome in

some way causes the X’s, in some of the oocytes, not to

separate properly at the reduction division (presumably,

this is because they did not pair with each other as usual,

but one of them paired instead with the Y, leaving the

other X free to go either to the opposite or to the same

pole as the first X). Thus some of the eggs in which the

above process has occurred come to contain two X-chro-

mosomes, whereas normal eggs contain only one. Now,

it is found that in those eggs which receive both X’s, no

interchange has taken place between them, whereas in the

eggs containing one X, interchange has taken place about

as often as usual. Hence interchange is connected with

whether or not the Y allows the two X’s to unite and sep-

arate properly, i. e., interchange seems to be a result of

the way in which chromosomes pair and separate during

synapsis, and, as we have seen, if interchange occurs at

this period, it must be by crossing-over.

B. THE CORRESPONDENCE BETWEEN SEPARATION FRE-

QUENCIES AND CHROMOSOME LENGTHS

Tn the present section still another possible test will be

given of the conclusions arrived at by Morgan, that the

factors are in line in the chromosomes, and that the order

in which they lie determines in a general way the relative

frequencies with which they separate from one another.

And it has just been explained that evidence for these

ideas is also evidence for crossing-over: that if the dia-

grams do represent the chromosomes and show the factors

in their real order, then the facts of linkage demonstrate

that, during synapsis, whole sections of the chromosomes

change places at once, i. e., cross-over.

The second test of the validity of the chromosome dia-

grams is as follows: If the order of the factors shown by

their linkage relations, and represented in the diagrams,

is their real order in the chromosomes, then it would be

10 THE AMERICAN NATURALIST [Vou L

possible, by adding together the frequencies of separation

between all adjoining factors, to obtain the total frequency

of crossing-over in the chromosome. This total frequency

would be represented in the diagrams by the total length

of the latter, since it is always the per cent. of separations

between the most closely adjoining factors which is chosen

to determine the number of units of length in any region

of the diagram. Now, the total frequency of crossing-

over in a chromosome ought, we should expect, to be de-

termined by the length of that chr . Accordingly,

we should expect to find differences between the total fre-

quencies of interchange (or the diagram lengths) of the

different groups of factors exactly paralleling the size

differences existing between the chromosomes themselves.

It will be seen, however, that such an expectation assumes

also (1) that crossing-over occurs with equal frequency

in all parts of a chromosome, and in equal parts of dif-

ferent chromosomes, and (2) that the factors available

for working out the total frequency of interchange do not

lie in any one limited region of the chromosome, but are

more or less scattered, some of them lying near each end.

A negative result from our test, then, might merely mean

that one of these two assumptions was incorrect, and this

would not disprove any essential point in the theory of

crossing-over previously outlined. On the other hand, a

positive result would seem to be too much of a coincidence

to happen by mere chance, and so would seem to prove the

correctness both of our main theory and of the two latter

points.

In regard to the size relations existing among the chro-

mosomes themselves, as determined by cytological ob-

servations, the work of Stevens (13), taken in connection

with the later work of Metz (5), and of Bridges (2), shows

that there are four pairs of chromosomes in Drosophila:

a pair of moderately long sex chromosomes, two pairs of

very long ‘‘autosomes,’’ and one pair of minute ‘‘auto-

somes.’’

We may next consider the lengths of the genetic groups,

No. 592] THE MECHANISM OF CROSSING-OVER 211

as determined by their ‘‘total frequencies of interchange.’’

The length of the first, or sex-linked, group of factors has

been found to be about 66 units; a unit, it will be re-

called, is a section of the chromosome of such length that

breakage occurs within it, on the average, one time in a

hundred cases. The evidence then shows that, in a hun-

dred cases, the first group breaks 66 times. This does

not mean that it breaks in as many as 66 cases out of 100,

for it may break two, or, very rarely, even three times,

coincidently (at different points along the chromosome)

in the same case (‘‘double or triple crossing-over’’). As

previously explained, when two breaks thus occur coinci-

dentally, the extremes of the chromosome come to lie on

the same side, and so a factor at one end of the first group

does not separate nearly 66 times in a hundred from a

factor at the other end; owing to these coincident breaks

it really separates in only about 45 per cent. of cases. The

number 66 is consequently not obtained by merely deter-

mining the frequency of separation from each other of the

two most frequently separating factors, but, as mentioned

above, it must be derived by adding together the fre-

quencies of all the smallest parts of the chain (frequencies

of AB + BC + OD, ete.). In the case of the first group,

the determination of this ‘‘total-length’’ has been accom-

plished by the combined efforts of a large number of

people, although by the work of Morgan, Sturtevant and

Bridges particularly.

Group II has a much greater length. It is probably

over a hundred units long, and is certainly over 90. This

result has been obtained principally by the work of Bridges

and Sturtevant, although, as before, others have helped

very materially. Mention must here be made of the fact

that Sturtevant has discovered in this group specific

mutant factors which, when heterozygous, lower the fre-

quency of separation in certain regions of the group very

much, although the order in which the factors are linked

is not changed (16). The variation certainly proves, how-

ever, that (if the groups represent the chromosomes)

212 , THE AMERICAN NATURALIST [ Vor. L

then, under certain special conditions of heterozygosis,

different regions of the same chr may differ in

regard to the frequency of crossing-over within them, for

different regions were not affected in the same way by

these factors; it also shows that equal lengths of differ-

ent chromosomes may have different frequencies of cross-

ing-over, for these factors affected only group II appre-

ciably.

In group III, crosses involving several combinations of

different factors have been made by Sturtevant, Bridges

and Dexter, but the order of none of the factors has until

recently been worked out by them nor has any consistent

general scheme been attempted. The information has, in

fact, not been adequate for this purpose, and much con-

fusion has also arisen on account of the great linkage

variation in this group, which seemed to occur very fre-

quently. Sturtevant, who, as stated in section I, discov-

ered the first case of linkage in the third group—namely,

that between pink and ebony—had determined the initial

positions of these factors, placing them at about 4 units

apart, and next Bridges, who had found kidney (eye),

had determined its position at about 15 units from pink,

though he did not determine the relation between kidney

and ebony. As a matter of fact, however, the kidney

determination had been made with flies in which there

was a greater frequency of crossing-over than in the ex-

periments of Sturtevant, and, as will appear later, in any

given experiment kidney is really nearer to pink than is

ebony. From time to time after this other mutants were

discovered (peach, Bridges, May, 1913; sepia, Wallace,

May, 1913; spineless, Bridges, May, 1913; deformed eye,

E. Cattell, 1913; band, Morgan, 1913; rough, Muller,

June, 1913; sooty, Sturtevant, Oct. 1913; spread, Dexter,

Nov. 1913; dichete, Bridges, July, 1915), and the fact

that these mutants were members of the third chromo-

some group was determined (peach, sepia, spineless, band

and dichete by Bridges; deformed by Cattell; rough by

Muller; and sooty by Sturtevant). The author mean-

No. 592] THE MECHANISM OF CROSSING-OVER 213

while undertook experiments with a view to determining

the order of these factors, their frequencies of separa-

tion, and the manner in which these frequencies vary, and

also sought to correlate with the results the data previ-

ously obtained.

It has developed from this work that group III is of

the same ‘‘order of magnitude’’ as the first and second.

This is the result required by the cytological facts. To

complete the parallel, it should be found that the third

group is longer than the first and, in fact, of just about

the same length as the second group. Whether this is

true can not yet be stated definitely, but the results indi-

cate that it is. It is certain that the length of the group

of factors dealt with is at least 55, but another estimate,

which, for reasons given below, would seem more prob-

able, gives the length as over 100. It should also be

borne in mind that not as many factors have as yet been

worked with in this group as in the other two, and it may

well be that other factors will be found to lie beyond any

of the twelve which have so far been approximately

‘‘placed.’? Thus, even if 55 should be the normal value

for the factors dealt with, the whole group may very well

be considerably longer. In the first and second groups,

factors lying well beyond all the others were discovered

after the positions of more than a dozen had already been.

determined.

The reason for the uncertainty in regard to the total fre-

quency of separation among those factors which have

been worked with is to be found in the linkage variation.

Sturtevant had discovered that certain races, containing

the mutant factor for ebony body color, gave extremely

low frequencies of separation; that is, ebony flies, when

crossed to those with pink eyes (pink is also in group

IIT), gave a hybrid in whose germ cells very little recom-

bination between pink and ebony occurred. Other races

(e. g., those containing sooty, an allelomorph of ebony),

when crossed to pink, gave higher values. He therefore

concluded that the ebony flies contain a factor (let us call

214 THE AMERICAN NATURALIST [ Vor. L

it C) which reduces the frequency of separation, and

which is dominant, since it produces an effect in the hybrid.

I have found that two other races of flies, one having the

factor for spread wings (also in group III), and the other

showing no ‘‘visible’’ mutant factors, also contain C, as

they behave in the same way as ebony. However, the hy-

brids produced when these races are crossed with ebony

give high frequencies of recombination again! This re-

sult shows that, as in group II, these races do not really

contain a factor which normally reduces separation fre-

quency, for, when both homologous groups of an individ-

ual contain the factor C—i. e., when it is homozygous—

the frequency of separation is high again. (This also

explains an irregularity observed by Dexter, who obtained

a high frequency of separation in a cross involving ebony

flies. )

It happens, however, that these high separation fre-

quencies obtained when C is homozygous are even higher

than those occurring in crosses not involving C at all, and

so presumably homozygous for its allelomorph, e. By

analogy with Sturtevant’s findings in the second group,

this would mean that in most crosses hitherto made not

involving C there has nevertheless been another factor

heterozygous, which has a similar, but lesser, effect on

the regions of the group studied. Some support for this

interpretation is found in the fact that occasionally higher

frequencies are obtained in these crosses not involving C,

which appear to overstep the limits of chance variation.

The evidence thus far secured on groups II and II

points to the conclusion that the highest frequency ob-

tained is that which should be regarded as the normal

value, and that very marked departures from this, which-

affect only a particular group, are generally due to heter-

ozygosis in special factors of that group. If it should be

found that marked differences affecting the total fre-

quencies of particular groups do occur, in cases where

the flies are homozygous for whatever factors influencing

linkage they may contain, we might naturally expect that

No.592] THE MECHANISM OF CROSSING-OVER 215

such variations would have gradually accumulated in the

course of evolution, until no correspondence remained

between the relative lengths of the chromosomes and the

total separation frequencies. But the parallel which does

exist between the observed chromosome lengths and the

usual (homozygous) total frequencies, would seem too

close to be meaningless, and so we should be led to be-

lieve that for some reason marked variations in the fre-

quencies of particular groups, even though, they may be

possible, do not generally persist; in other words, the

frequencies seem usually to stay at least roughly propor-

tional to the actual chromosome lengths, and to furnish

another verification of the theory of crossing-over. Fur-

ther evidence of this will be met with when we consider

group IV.

As the data whereby the positions of the factors and

the total frequency of separation have been determined

in group III, have not hitherto been published, it,may be

of interest to present some of them here. In order to

obtain data on as many combinations of factors as pos-

sible in the same cross, so that the linkage values between

the different factors would be comparable, I have en-

deavored to make up, by cross-breeding, stocks containing

six or seven mutant factors in group III at the same time.

Since on account of the baffling linkage variation, the

order of these factors could not well be determined by

combining the results of separate experiments each of

which dealt with only two factors at a time, it required a

great many trial matings before such multiple stocks

could be made up, as of course the crosses have to be.

made in a certain precise order, to secure a combination

of many linked factors. To obtain stock ABC, for ex-

ample, it would not suffice to make up AC and then mate

it to B, for it would then require two coincident recombi-

nations (which might never occur) to secure ABC. More-

over, as a first step it had been decided to get combina-

tions of ebony with various factors, and very much time

was lost in this attempt, as it was not then known that

216 THE AMERICAN NATURALIST [Vor L

when ebony is crossed to most other stocks recombination

of the factors is nearly impossible. A stock has finally

been obtained, however, combining the following charac-

ters belonging to group III: sepia eye color, dichete

(bristles, and wing), pink eye color, spineless body,

kidney eye, sooty body color, and rough eye. Data have

not yet been secured with this final stock, but the follow-

ing experiment, in which most of these factors were in-

volved, may be regarded as typical of crosses not involv-

ing factor ‘‘C,’’ and consequently giving moderately high

frequencies of separation. Hybrid females from a cross

of sepia flies with flies containing dichete, spineless,

kidney, sooty, and rough, were backcrossed to the quin-

tuple recessive stock—sepia, spineless, kidney, sooty,

rough (dichete being a dominant). The count of off-

spring is shown below. The classification as regards

kidney has not been given, as this character can not be

distinguished with certainty in eyes which are also rough.

No Separation

se dic sps

sor

131 109

Separation Occurring at a Single Point

1. Between positions| 2. Between dic 3. Between sps 4. Between so

of se and dic and sps | and so

se dic normal | se sps dic | se sps dic ser sps

sps sor sor sor dic so

9 16 20 15 | 2 19 40 28

Ire Occurring Coincidently at Two Points

1; 2. Between se rot dic; a 3. pps se er dic; | 1; 4. Between se and dic;

dic and sp and s À

se dic sps sor | se dic sor

| Sps

se dic f

sps so

1 2 | 1 2

5 o

2: 3. asien dic and SPs; IR A iti 2s and sps;| 3; 4. Porere sps and so;

and so andr

‘|

se sps dic so r | mae n beee

dic

| sps r

k addition, 1 Di spineless fly appeared, ae must pone resulted from

a separation at three points coincidently (‘‘triple ecrossing-over’’), namely,

between die and sps, sps and so, so and r.

No. 592] THE MECHANISM OF CROSSING-OVER 217

The above classification of the flies, in respect to where

separation of factors occurred, is based on the assump-

tion that the factors are linked in the order: se-dic-sps-

so-r, aS on any other arrangement the above results

would show many more coincident separations between

certain factors, than single separations. The reader may

convince himself of this by working out the numbers of

the different kinds of separations on any other scheme.

We may say, then, that the above results prove that the

factors are linked in the order just given. Turning to

the individual separation frequencies, it will be seen that

se and dic separated 25 times when there was no other

point of separation, and 14 times when there was a coin-

cident separation, i. e., 39 times in all. As there was a

total of 432 flies this means that these factors separated

in 9 per cent. of cases, i. e., are 9 units apart. In a

similar way the results show that the distance between

dic and sps is 10, between sps and so is 11, and between so

and r 19.5, giving a total distance between se and r of

9+10+11+19.5 = 49.5, which agrees well with the

value 55, obtained by combining all the records of crosses

of this general sort.

Number of Flies on

“ Distance” Be- which Result is

Factors Involved tween Them Based

Depia dichitte o 0 0552 <5 2 epee y Oe eases es on 624

Dicheote apineless. -oaren es Peaswnsun ILD cd his ews 683

Sepia spineless o LI en coc E aga a a AO A 1,014

Spineless sogty ct. 7.4 ape ee ae Dh ty fos Seine eS a gs 1,198

Booty weigh 2. cous ce Se ee a eee a oe ae 1,097

Pink, Rpielemn’ soo 5 os ee i o eT canines oe oe 825

Pink kidney o.s.: oaan oa RED os e OE ses 963

Kidney sowy oR Ae owe O48 iy. Set eh gE 885

Piok sooty a hs ae, A T EAR pees 1,566

Detovnad: pink -aa an ae BA Oe ee a ie 166

Kidney baad i:a... oae taa A E SEAS ees N O 237

[r a)

e oR 5 2 29 35 55

Fie. 5.

The following summary of ‘‘distances’’ between vari-

ous factors of the third group Suares two at a time)

218 THE AMERICAN NATURALIST [Vou L

was obtained by averaging together the results of various

experiments not involving C, in which I had followed the

inheritance of both of the factors listed at the same time.

Many of the data listed in separate lines are of course

derived from the same experiment, as in most experi-

ments more than two pairs of factors were followed. It

should also be added that in one of the cases below (se

sps), the distance was slightly lengthened to allow for

coincident separations, the approximate proportion of

which was known, but the actual number of which had not

been determined in most of the crosses.

Having determined the order of the factors, these re-

sults may now be combined, in order to obtain a series of

values based upon as many data as possible, and to con-

struct a diagram of the group. The diagram so made is

shown in Fig. 5. The numbers underneath the symbols

of the factors represent the ‘‘distance’’ of the latter from

sepia, which, as it lies at one end of the group, is used as

a common point of reference. Although the distances

shown will undoubtedly be subject to revision, the order

of ‘all the factors shown, except deformed, band, and

beaded, is certain. Deformed (eye) is surely between

sepia and pink, but it is not yet quite certain that it is to

the right of dichete; band (thorax) is near sooty, but on

which side is not known; beaded (wing) is very near

rough, but it has not been established whether it comes

before or after it (a count of 50 flies showed no crossing-

over between them). It was Sturtevant who first deter-

mined the position of beaded (found by Morgan, May 710)

with reference to this series (to the right of sooty), and

Bridges who first determined that dichete lies between

sepia and pink (about 4 to the left of pink). The data

listed merely confirm these findings, so far as these two

factors are concerned. And it may here be repeated

that numerous other crosses of factors in this group have

also been made by these investigators, although the cros-

ses have not been of a sort to show the arrangement of

the factors studied.

No. 592] THE MECHANISM OF CROSSING-OVER 219

We may next consider the disposition of the factors of

group III in a diagram based upon data from flies hetero-

zygous for C. The separation frequencies which I have

obtained are given below:

Factors Per Cent. of Separations Number of Flies

Opis Spineless ois i soa ce pee h OD 0 iss uae E ee 527

Poinelens. MIGKCY erior eee ie T 0.0... sa:9 EE 527

Pink: kidney 2503 PEN SN eee es 1S ia eee 868

Kidney sooty (or ebony) .......-..... 0.24. Rae Meee 674

Sooty (or ebony) rough .............. 1R Ji s T EAEE 843

Kidney rough sisis ears bees o a BESET T ERE 1,211

A diagram based upon these data would show sepia at

0, pink at 19.5, spineless and kidney at 20.9, sooty and

rough at 30, and the total length would thus be 30. The

tenth of a unit of distance between kidney and sooty is

based upon one fly, in which separation had taken place

between these factors. Tests of the fly (which was a

sooty rough, resulting from a backcross of a female con-

taining p sps k so r from one parent and spread C from

the other) showed that the factor C had remained with

spread, and that no recombination had taken place be-

tween the positions of spread and sooty. This one fly,

therefore, proved that both spread and C were to the

right of kidney. The factor C is thus seen to lie right in

the heart of the region where it exerts its maximum effect,

as Sturtevant has also found in the case of the similar

factors in group II.

Sturtevant has obtained slightly higher frequencies of

separation between pink and ebony in some of his crosses

heterozygous for C. The lower value here recorded may

then be due to the flies being heterozygous for another

factor besides © which disturbs separation frequency,

and which is also met with in the crosses not involving C.

The following values have been obtained in crosses in

which the factor C was homozygous:

220 THE AMERICAN NATURALIST [ Vou. L

Factors Per Cent. of Separations Number of Flies

Sepia pink ........6...0ssec sees eee ees = 5h Saran aa es 136

Pak bay 62.62. ase. + <2 ee Bi fet wav WE ee ers 290

Besides this, it is found that spread is about two thirds

of the way between pink and ebony. Ebony, it will be

remembered, is an allelomorph of sooty, and therefore

occupies the same position. As no recombination has yet

occurred between C and rough in flies heterozygous for C,

it has not been possible to obtain these factors together

and so, in crosses homozygous for ©, the linkage of

rough has not yet been discovered. The length of the

group between sepia and ebony is 75 in these flies, as will

be seen from the above data. Although these figures are

based on a relatively small number of flies, the difference

between this and the shorter value (35) found in flies not

containing C is marked enough to be significant, espe-

cially since it occurred in various crosses of this sort. Tf

the distance between sooty and rough is expanded in the

same way, the group would have a length of much over

100. If, however, this distance is of the same length as in

flies without C, the total length would be 95. The reasons

have been given which incline us to the opinion that these

values obtained in crosses homozygous for C may repre-

sent the ‘‘normal’’ figures for this group rather than

those obtained in the experiments earlier cited. Further

investigation of this point, however, is being undertaken.

Group IV corresponds with the pair of small chromo-

somes in that it contains so few factors. For this reason,

the author, in his account of the inheritance of bent

wing, in 1914, said:

It also seems probable that when other mutations are discovered in the

fourth group, the genes in which they occur will be found to be linked

strongly to the gene for bent wings, since the fourth chromosome is prob-

ably the small one, and so any genes in it must lie near together.

One other mutant factor, ‘‘eyeless,’’ has since been found,

by Miss Hoge, to lie in this group. But although Miss

Hoge has made numerous attempts (3) to combine eye-

No. 592] THE MECHANISM OF CROSSING-OVER 221

less and bent, no recombinations between them have so

far been obtainable. Group IV, therefore, forms a marked

contrast to all the other groups as regards the frequency

of separation within it, and this result is the more strik-

ing, not only because it shows that there is a group of

factors corresponding in separation frequency to the pair

of short chromosomes, but also because it happens that

this group is the same one as that which had previously

been identified with the pair of short chromosomes by

reason of the fewness of the mutant factors discovered

in it.

It is therefore evident, not only that the relative sizes

of the chromosomes are in a general way like the separa-

tion frequencies of the groups, but also that where there

is evidence from another source indicating in which

chromosome a certain group lies, this is the very one to

which the group corresponds by its total frequency of

separation. It has been shown that this is true in the

case of the fourth group. In the case of the first group,

the sex-linked inheritance of the latter connects it with

the X-chromosome, and since this is the moderately long

chromosome, it is just this one with which group I would

be identified by its frequency of separation. The other

two groups, both of which are long—one certainly very

long, and the other probably so—are thus left to corre-

spond with the remaining chromosomes, both of which

are very long and indistinguishable in appearance.

In the remainder of this article, therefore, the word

‘chromosome’? will be used instead of ‘‘group’’ and

‘‘erossing-over’’ instead of ‘‘separation of linked

factors.”

(To be continued)

INDIVIDUAL DIFFERENCES AND FAMILY RE-

SEMBLANCES IN ANIMAL BEHAVIOR

HALSEY J. BAGG

` INSTRUCTOR IN Brotocy, New YORK UNIVERSITY

In experimental work on animal behavior, but little

attention has been paid to individual differences, and

practically none to family resemblances. In studying the

inheritance of conduct in man, experimental methods can

not be used. Students of eugenics depend on observa-

tions difficult to verify. In the work here described an

attempt has been’ made to apply the methods of genetics

to the study of conduct. Such work was begun by Pro-

fessor J. McKeen Cattell some fifteen years ago, but the

results obtained by him and his students were not pub-

lished, and the problem has been given to me.?

The plan of the experiment is to measure individual

differences in behavior, to determine the extent to which

the animal which departs from the average in one direc-

tion will depart in others, to measure the resemblances in

families and in lines of descent, and to determine the

degree to which kinds of conduct can be established in

family lines by selection. It is evident that such a prob-

lem can be solved only by many years of work and with

the facilities of a research institution. In the present

paper there are described the individual differences and

family resemblances of 90 mice, as determined by the time

required to find their way through a maze. The same mice

have been tested in other ways, and further experiments

are now in progress with the F* and F° generations.

1 Basset has recently published an article on ‘‘Habit Formation in a

Strain of White Rats with Less than Normal Brain Weight.’’ Behavior

Monograph Series, No. 9, 1914. Maedowell in Science for November, 12,

1915, gives a brief abstract of work on ‘‘Parental Alcoholism and Mental

Ability. A Comparative Study of Habit Formation in the White Rat.’’

2 The greater part of the work presented in this paper was done at Co-.

lumbia University, the results being used for a master’s thesis.

222

No. 592] INDIVIDUAL DIFFERENCES 223

A maze, designed by Professor Cattell, was used, the

plan of which is shown in Fig. 1. The animal has in the

First . Second

Compartment CompaRtmMeNt

5m.

Foon

> ) t

ARtMent

A A Comp

ee B

20 cm 20 ¢m. 200CM.

20cm

Fig. 1. Diagram of Maze. A and A’=elosed wire gates; B and B’= open

wire gates; E = entrance from above.

first compartment the alternative between two gates, one

of which can be pushed open while the other is locked, and

then it has the same alternative in the second compart-

ment. When it takes the correct way in. both compart-

ments, it finds itself in the food compartment. The path

that the animal must follow can be altered by changing the

gates which are locked. ‘‘Unit construction”’ is used in

the dimensions, which are adjusted to the size of the ani-

mal, and in the fact that any desired number of standard

units can be added.

_ Preliminary tests were made with albino rats, but these

were given up for mice, which are more active and more

easily handled. The mice were given one trial each day,

and were tested at as nearly the same time as possible.

Light was found to play but a minor rôle in the tests, day-

light and artificial light serving equally well. At the out-

set the age of the mice when first tested was not always

known, but later when the various litters were obtained

the young mice were tested at or about four weeks old.

The mice were rewarded fora successful trial by a mixed

diet of milk, bread, oatmeal and sometimes meat. They

always had a little dry bread in their cages. Besides satis-

fying their hunger, the mice had the additional reward of

a place for exercise and the companionship of the mice

that had just been tested. The order of the tests was

224 THE AMERICAN NATURALIST [ Von. L

varied day by day. In ease the way through the maze was

not found in 360 see. the animal was removed and tested

again the following day. 360 sec. is thus the maximum

record for a single trial.

Seventeen trials were made with each individual. This

was a desirable number for two reasons: first, because this

number was sufficient for the average mouse to learn the

maze, and secondly, because the seventeen trials could be

divided into three somewhat homogeneous groups. The

first two trials are largely affected by chance, so, although

given here for completeness, they are not averaged in the

final ratings for each individual. The second group of

five trials represents the period of more rapid learning,

the third group of ten trials the results when the learning

is slow or completed. In this paper the averages of the

last fifteen trials are used as the index of performance.

For some purposes the last ten or the last five trials might

be preferable. The rate of learning as determined by the

relation between the first and last groups may also prove

to be of value.

In addition to the time, the number of errors, i. e., the

number of cases in which the mouse tried to go through

the locked gate, is given in the tables, as this is a measure

of the activity of the animal. With only a few exceptions,

however, the error and time curves correspond. Con-

sideration of the correlation between error and time, and

between performances at the beginning and the close of

the trials is postponed until data can be given for a larger

number of individuals and the records for other kinds of

behavior.

In Tables I to V are given the complete records of the

90 mice tested, grouped in families as described below.

The average time is 44 sec. per trial for the last fifteen

trials. The distribution of the individuals is shown in

Fig. 2. In 41 cases the time was under 20 sec., in 19 cases

between 20 and 40 sec., in the remaining 30 cases between

40 and 200 sec. None of the mice failed to learn the maze.

When the experiment started, several colored mice—

chocolate, agouti, gray, black and yellow—were tested,

No. 592] INDIVIDUAL DIFFERENCES 225

TABLE I

COMPLETE TIME AND ERROR RECORDS FOR THE YELLOW FAMILY

a mlo aSa a EA N mho 9/2 ale » u

esea La asg Seigsiseieaiee

No. pd Pas N i o

: AIEG GE ; BE|SE\SE/E|G5

Me Be ee 360/228) 58/115'1.6|38Yo" 360/277) 98 157|3.6

21W.. 2 41 1| .2139Y g . . 1860/143| 25 2:8

BAM i eo ees 234| 28) 14| 19| .4 J40Y 7 360| 47 1| .8

BRAG A AA 2 9| 5441Y 9: 210 20 | 2) 6

PMN glia Le 21| 9157Y F \360/ 171/109, 130/3.7

25AGW 9. Pepe war os: 63}1.7 [58Y F7 357/130| 51| 77|2.4

MIP ss ve 0A SOO LOOT Felaed at leon ola 360 275) 98,156/3.3

Heo 3i 51 | |

In the first column is given the catalogue number, color and sex of the

animals. In the second are the time averages (in seconds) for the first

two trials; in the third, for the next five trials; in the fourth, the last ten

trials, and in the fifth column the average of the two preceding columns.

The error average for the last 15 trials is given in the last row of figures.

This order is followed in all the subsequent tables.

One day’s record has been omitted for mice Nos, 27, 28, 29 and 31 be-

cause the poor records for that day were obviously due to a constant error,

on account of traveling, etc. These are the only cases where such a con-

dition has occurred.

TABLE II

CoMPLETE RECORDS FOR THE WHITE FAMILY

Error

NOOR SO MON | Average

1.0

226 THE AMERICAN NATURALIST © [Von. L

TABLE III

CoMPLETE RECORDS OF A FAMILY CONSISTING MOSTLY OF YELLOW

INDIVIDUALS

aa) 2\fa\3a8 PRPS E

No EEEERERE No geleSleeie8£8

EEJZE JEJER EESE ge 85 8%

et eee 225/109 71 83.1.7 |61Cho -e 3360 202) 80, 87 1.0

SUR eel 154/18 ea a 312, 90| 18 42/1.5

OAV es ce 186| 88| 39] 56/1.0 |63YWQ..........| 243130) 38 S 1.2

36Y F. 5| 28| 37| .9 |64Y9........... . |360)182/113/136|2.3

SOV Ge aa 137| 20| 16| 17|1.2 esYW @. pa ni 71| 65| 67/|2.0

SAW Oe ee 360| 36| 34| 35) .7 |J69Yo". BE TT | oori

55Y F 1360; 242/103 ct TOY Bree eee eT 234 225 751124|2.3

ee a Me te Ce | |

TABLE IV

COMPLETE RECORDS OF A SMALL FAMILY SHOWING Goop RECORDS.

| | o

Sg alfalsal.e agiP2isaita F

No. PSRS eSEE eaga a\e

P SESE aktai 7 Ee\fe|se\ seas

ee 229) o s olowo oo k 142/13| 5| 8) 5

30GRF 33i 30 LO HOGE P... 6| ;

W 7| 16| 7|10| 5 l47B1g...... arses 29 | 31 |1.0

TABLE V 4

COMPLETE RECORDS OF THE UNRELATED INDIVIDUALS |

a2 aloe tal ng alee eal 2s F

No. Ba He eS eSlEe BSlPS 2 SleS/8 =

Ia a a |

| h :

ica myer E

WAP le 213} 48 a SI hW es 74\22| 8|12| 6

rh AES JR RE SS ea ir Re ca 103 1405. t BSW... nnn 291/38 | 5 |17 |1.0

aa gs 316 me 63i 613.3 3G ...... 130| 78 | 15 | 38 |1.0

aWe TI 18-20 16) 7 BBW eS 131| 14| 9/11] 8

_ besides a large number of albinos, and among the yellow

mice several made poor records. These mice were mated,

and they and their offspring compose a group of 27 indi-

viduals, whose average time and error record is consider-

ably in excess of the normal for the entire population.”

3 This group of 27 mice was composed (see Tables I, III and V) of Nos.

20 and 26, and their seven offspring; No. 27, the sister of No. 26; a litter

of five mice, Nos. 32 , 33, 34, 36 and 37 and their ten offspring, and finally

two unrelated yel mice, Nos. 2 and 3, that were used at the beginning

cae res The 63 remaining mice of the white group bring the

No. 592] INDIVIDUAL DIFFERENCES 227

Number og InpdiviDuals

Number oj Seconps

Fic. 2. Total distribution curve for 90- individuals for the last 15 trials.

The yellow group gave an average time of 83 +7.0 sec.,

and an average of 2.0 errors for the last 15 trials. The

other mice gave an average time of 27.5 + 2.0 sec. and .9

error per trial. The yellow mice were thus found to take,

228 THE AMERICAN NATURALIST [ Von. L

on the average, at least three times as much time and to

make twice as many errors as did the white mice. In Fig.

3, the distribution curves of both groups are given, the

oF-

ZN Yellow

Number 05 INDivIDUAIS,

\GRou

X a

x j

20 40 60 80 100 120 JJO 160 180 206

Number og Seconds

Fig. 3. Distribution curves of yellow and white groups for the last 15 trials.

yellow in a broken and the white in a solid line. The curve

for the miscellaneous group is skewed, most of the indi-

viduals falling between 0 and 20 sec. The curve for the

yellow family is nearly flat, there being about the same

number of individuals in each time group.

No. 592] INDIVIDUAL DIFFERENCES 229

Fig. 4 gives two average practise curves, one for the

group of 63 white and colored mice, and the other for the

group of 27 mice that are mostly yellow. The records of

360

4 A Yellow

vi IA Group

Number o$ Seconps

1a 34 S6 T89 101 121314151611

Number og TRIANS

Fic. 4. Average record curves for yellow and white groups.

230 THE AMERICAN NATURALIST [ Von. L

all the individuals for each successive trial were averaged,

- and the probable error calculated for each point on the

curve. In accordance with a plan proposed by Professor

Cattell, the limits of the probable error are shown by

broken lines. The chances are even that with a greatly in-

creased number of cases the time would have remained

within these limits, and a nearly smooth curve can be

drawn within them. A notable exception is the tenth trial

with the yellow mice. At this point there is an unusually

! | tA

340; x

360] 14

———

—

——— eo

im ——— a a

—_—=— m — =

T e oe ee a

(e)

—- -a

No.26 Yg

‘ i À

Number OF SECONDS

‘

‘No. 209

-No.88 wo

No.SIwNg

S 9 10 H 42 PA M475 16 17

Number OF tTRIAls

Daily record curves for two white and two yellow mice.

No. 592] INDIVIDUAL DIFFERENCES 231

large number of low records, more than the law of prob-

ability would warrant.

In Fig. 5 are given sample practise curves, showing the

daily records for two white mice, Nos. 51 and 88, that

learned the maze quickly, and for two yellow mice, Nos.

20 and 26, that were slow to learn. The arrows at the

highest points indicate that the mouse did not pass

through the maze. Thus No. 26 only got through on the

fifth trial and failed in the eleventh, fifteenth and six-

teenth trials.

The mean variation for the entire group of 90 mice was

found to be 35.6. This means that any mouse picked at

random from the mixed group would be likely to vary

from the average by 35.6 seconds. In order to find

whether mice of the same litter varied less than unrelated

individuals, the mean variations for each of the 18 fami-

lies was calculated, and these when weighted for size of

family were found to be 20.2. The resemblance in be-

havior between mice belonging to the same litter was con-

sequently nearly twice as great as between unrelated indi-

viduals. This corresponds to a coefficient of correlation

in the neighborhood of 0.5 for brothers, as found by Pear-

son, Thorndike and others.

TABLE VI

AVERAGES FoR SEX DIFFERENCES

Average Number of

No. Color Sex | Average Last 15 Trials | Probable Error Erros

32 White, etc. g 27.69 Sec. + 2.9 .9 per tı i

31 White, etc. Q 27.35 Sec. + 2.9 1.0 3

15 Yellow, ete. g .0 Sec. + 9.6 2.0 5

12 Q 75.0 Sec. +10.3 2.1

In Table VI, the males and females are grouped sepa-

rately, and their average times and errors are given. In

both groups of (mainly) white individuals, with 32 males

and 31 females, and in the group of (mainly) yellow mice,

with 15 males and 12 females, the times for the females

are on the average slightly shorter, but the differences fall

within the limits of the probable error and indicate that

there are no sex differences in this kind of behavior.

232 THE AMERICAN NATURALIST [ Von. L

We may now take up in more detail the family histories.

Fig. 6 gives a graphic representation of matings, from

ae ee

OF FIFE SO

No22 wW

Family h nan 12

THO ELIE

E hi AVERAGE = PRPA "an

Fic. 6. Diaa of yellow family. Squares denote males and circles fe-

manne

experiments have given an unusually large number of yellow offspring has been

made the subject of another investigation from a mendelian standpoint

which there were selected two mice, No. 20 YQ and No.

26 Yg, which made unusually poor records, 115 and 183,

respectively, though the other mice in the same litters had

good records. The parentage of Nos. 20 and 26 was un-

known; they were mated and gave two litters, each com-

posed of three males and one female. Three mice in these

two litters gave unusually slow records and made con-

siderably more errors than normal. Two other mice gave

poor records; two gave good records, while one died be-

fore it was tested. It is unfortunate that both females

in these litters died before further offspring could be

obtained. Table I gives the complete record of both time

and error averages for these mice. It is a question

whether or not the selection of parents having poor rec-

ords tended to produce more than the normal number of

offspring slow to learn. Further investigation can alone

afford an answer.

The mice whose records are given in Table II are

graphically represented in Figs. 7 and 8. They have been

No. 592] INDIVIDUAL DIFFERENCES 233

carried down through the sixth generation, and are still

being tested. As neither of the mice of the F+ generation

mated with No. 91, the only individual of the Fë genera-

tion, No. 91 was mated successively with four unrelated

BRO

Parents

No.8 w sae girs Qw

TOOL

Noj2w Nol, y bint No./6

AGE = SITE 15.7 Sec.

4AWw No.50w Nodiw D NaPSw

Family Sa

SOE

No. 75W No.76W No.77w No.78w

Family AveRaQe = 73.72 19.5 Sec-

ore ane 113 ajal No.67w

#.3215.9Sec. `

' Er No.7iw No.72w No.73w No.1#

o. 14W

Family AVERAGE = 66 t 22.3 Sec.

No.9 w

Fic. 7. Descent of a white family.

white females, Nos. 86, 87, 88 and 89. These females had

been previously tested and found to give exceptionally

good records as indicated in Table II. Twenty-three off-

Spring resulted from these matings. Their records are

remarkably uniform and the family averages are the low-

est so far obtained. The records of these families are

graphically represented in Fig. 8, and here only the con-

tinuation of the white family is given showing. the F* and

F° generations separately for each individual family. It

234 THE AMERICAN NATURALIST [ Vow. L

is hoped that future offspring may be obtained to con-

tinue this strain. The times for the fifteen trials do not

always correspond with the times for the last ten trials.

, BKO

No.io3w No.lo4#w NoIoS5SW No.106W

Family AVERAGE = I4t.6 Sec.

ie”

O°

No.I09w No. now Noọo.Inw No-112W No. Kw No. Aw No.sw

Family Average = 13.1 £1.7 Sec.

Nowiéw NonIw NoH8w Now No-mow. No.12IWw

“Family AVERAGE = IR to Sec.

pai S

f‘ G2) © (is)

NoJ22W Nola3w Noi4w Now2Sw Nonew Noi 7Ww

Family AVERAGE = 19.143.4 Sec:

Fic. 8. Continuation of white family. No. 91 mated with four females.

Thus No. 66 has for the fifteen trials an average time of

113 sec., but in the last ten trials reduced the time to 17

sec. The capacity of the mice can only be finally deter-

mined after the same individual has been tested by dif-

ferent methods.

No. 592] INDIVIDUAL DIFFERENCES 235

Another family, mostly yellow, was derived from a

yellow female and an unknown male, probably white. The

F! from this mating gave a litter of six, Nos. 32 to 37 in-

clusive. The records of 5 of these (one died) are given in

Table III and are graphically represented in Fig. 9. The

LUunkwown]

- PSS

32Y No-33Y No 34#Y No.35Y NoS6Y No, 37Y

mily AveRAGe = 56.6 10,18 Sec

F. G9) [ise] (9)

NoS4w NoSSY No.SbY

m RAGAN, 99 t 25.8 Sec

rog OO

FAm rly AVERAGE = 70 £2)

Fic. 9. Descent of a family of mice consisting mostly of yellow individuals.

only female of the litter (No. 37) mated but once, and it is

not known with which brother. She bore in the F?, two

females and a male (Nos. 54, 55, 56). Both females of

this generation were crossed with their brother and two

litters resulted. No. 55 X 56 gave Nos. 68, 69 and 70 in

_ the F?, and No. 55 X 54 gave Nos. 61 to 64 inclusive.

From a survey of the complete records of these mice, it

is seen that although the F? and F° generations came from

the female, No. 37 (which made the exceptionally low

record of 17), still two of her young in the F? made poor

records, and Nos. 61, 64 and 70 in the following genera-

tion did the same.

The records of a family of white and colored mice are

given in Table IV and Fig. 10. The two parents and the

four offspring all have good records. It is to be regretted

_ that no further litters were obtained from this family.

236 THE AMERICAN NATURALIST [Vou. L

No.30 GR. NO.29W,

Now No.#5w No. #¥6Gr. NO.#7BI-

Family Avera Qes: 13.7 £ 43 Sec.

Fic. 10. Descent of a small family of mice showing good records,

In Table V are given the total records for the isolated

cases, which complete all the individuals tested. Here

again the relative inferiority of the yellow mice may be

noted.

SUMMARY

1. Albino and colored mice can be used to advantage for

laboratory work on animal behavior.

2. The type of maze used seems well adapted for this

kind of work.

3. There is a marked difference in individual behavior.

4. There appears to be a resemblance among individ-

uals of the same litter.

5. There appears to be a considerable difference among

different strains.

6. The sex differences, if any, are very slight.

EVOLUTION OF THE CHIN

T. T. WATERMAN

ASSISTANT PROFESSOR OF ANTHROPOLOGY

UNIVERSITY OF CALIFORNIA

In the Smithsonian Report for 1914 is an article by

Louis Robinson, M.D., on ‘‘The Story of the Chin.” Dr.

Robinson in this article goes so far as to explain the pres-

ence of a chin in human beings as the result of the habit

of articulate speech. Quite a different explanation is pos-

sible for the existence of this extraordinary feature of

our anatomy. I should like to suggest some of the evi-

dence which would seem to indicate that Dr. Robinson’s

ideas need rather careful review.

By chin is to be understood the projection or point on

the under jaw, below the mouth (Fig. 1). The jaws of

most vertebrates have no projection or prominence in

this region.

It will therefore be recognized at the outset that the

chin is a very ‘‘human”’ trait. It is a trait that distin-

guishes man from other living primates; even from his

near relatives (compare Figs. 1 and 2). It even sets off

G 1. Recent homas lower jaw, Fic. 2. Lower jaw of an orang, show-

mowing the so-called “mental” or ing the absence of chin.

“chin ” prominence,

the man of to-day from the more ancient of his progen-

itors. The earlier fossil skeletons of man are quite chin-

237

”

238 THE ‘AMERICAN NATURALIST [ Von. L

less. The absence of this bony projection in the face is

in fact one characteristic thing in our more or less ape-

like forefathers (Figs. 3 and 4). The question is, how

the ‘‘evolution’’ of this chin is to be explained.

Dr. Robinson’s explanation seems to me to boil down

to this: that man is, before all other creatures, a talker.

In talking, the genio-glossus muscle is called upon to do

the most work. This is a fan-shaped muscle which com-

poses a large part of the under portion of the tongue, and

is attached to the inner surface of the jaw just within

The lower jaw of an ancient Fic. 4. The lower jaw of Homo

ancestor of man; the “ Hoanthropus eeoa a Sie stocene ances-

dawsoni”? an early Pleistocene form tor of recent man, found at Mauer

from Piltdown, Sussex (sketched from near ta eincte:

na restoration by Dawson and Wood-

ard).

the chin. It is, according to Robinson, larger, more

specialized in structure and more fully fasciculate in man

than in the monkeys. The chin, then, says Robinson; is

the point of origin for this elaborate muscle, which in a

minute of conversation makes several hundred separate

movements. The chin has crowded forward in its pres-

ent conspicuous form as successive generations of men

developed more adequate apparatus for speech. In other

words, the chin developed because of the use and the

consequent development of this one ‘‘talking-muscle.’’

It is only fair to remark that this is an old discussion.

Walkhoff, in a series of papers, beginning with a volume

edited by Selenka in 1901, put forward the theory now

rejuvenated by Robinson. The suggestion was critically

reviewed by Fischer in a series of articles.2, Since then

the idea has appeared in a variety of journals.

1 Ausz, Biol. Centralbl., Volum

2 Especially Anat. Ans. Volina a (1903); Volume 25 (1904).

No. 592] EVOLUTION OF THE CHIN 239

It is only proper to say, further, that Robinson’s vari-

ous statements about the matter are hardly consistent.

He states that the chin is, in origin, merely a buttress for

the canine teeth; and he also believes it to be the result of

sexual selection. Having accounted for it in these two

ways, he throws in his remarks about the genio-glossus

muscle for good measure. He closes by spending more

discussion on the genial tubercles than on the chin itself.

If the first of the statements to which reference has been

made is correct, those animals which have large canines

ought to be found with the best-developed chins. Quite

the opposite is the case. Generally speaking, animals

with very large canines, such as the baboons and others,

are conspicuous for their very lack of chin. The author

also makes certain sensational statements about the lower

jaws of the ‘‘lower’’ races, that need full discussion; as-

suming in one place that uncivilized peoples have phonet-

ically simple languages, an assumption which is start-

lingly contrary to the facts. His assumption that the

genio-glossus muscle is the one prime factor in speech is

not borne out by phoneticians, as he himself notes in one

place (page 305). Aside from such minor points, all of

which demand argument, I should like to point out what

seem to my mind to be some of the more important rea-

sons for considering his theory of the origin of the chin

imperfect.

In the first place, if man’s chin develops from his talk-

ing habit, all other animals, without exception, should

lack chins altogether. None of them have a language,

properly speaking. Robinson himself, to go no further,

mentions other animals, notably the elephants, who do

possess chins. The latter have it, as the saying is, to

spare—much more than a human being has. Robinson

points out quite correctly that talking and the chin de-

velop together, as we observe man evolving through vari-

ous types. This does not necessarily mean, as Robinson

Seems to assume, that talking produces the chin. On the

contrary, the gluteus maximus muscle undergoes tremen-

240 THE AMERICAN NATURALIST {Vou. L

dous development throughout the same period which

brought in highly specialized language. No one has ever

suggested, however, because it develops along with highly

specialized language, that this muscle is concerned in

speech. I should say that the proper method is to see

whether there is any general tendency which would pro-

Fig. 5. The lower jaw of an aged person, showing the reabsorption of the

alveolar border.

duce chins in the course of evolution, a tendency which

would operate in the case of other animals, and also in

` the case of man and his forerunners. I think there is

such a general principle, and a very simple one. I should

be inclined to explain the chin, not as a by-product of

speech, but as a result of a general reduction in the size of

the jaw.

The man-like apes have very heavy chin-less jaws,

which, in point of absence of chin, compare with the jaws

of the great dogs or cats. Fossil man, too, exhibits, in-

the more ancient types, enormously large jaws. One

general fact, then, in the evolution of modern man, has

been a reduction in the size of this part of the body struc-

ture. This reduction went along with wider intelligence

in the selection of food, and has perhaps been accelerated

in man ’s case by the invention of cooking and other ar-

tificial treatment of food-substances. It is, then, a gen-

eral tendency in the evolution of the human and related

types. If we can not explain it, we may at least recognize

No. 592] EVOLUTION OF THE CHIN 241

it. The next question is: if the jaw is in this way being

reduced, should we naturally expect it to be equally re-

duced in all directions? There are reasons why we might

anticipate that it would not.

If we consider especially the horizontal ramus of the

jaw, the fact is striking that not all parts of it are equally

permanent. The teeth themselves, and the upper border

of the ramus, are temporary structures. In old age, the

teeth are lost. The upper margin of the jaw itself is, in

late life, reabsorbed (Fig. 5), which, with a corresponding

loss in the upper jaw, produces the well-known nut-

cracker appearance of the aged human face. Without at-

tempting to dogmatize, I will go so far as to say that we

might confidently expect that the region in the lower jaw

which is lightest in structure, and the first to disappear in

the individual, would be the part which would naturally

respond first to the influence of external environment.

Put in another way, the suggestion might be worded thus:

We recognize as the important thing in the jaw the teeth.

Hence, as smaller teeth became more appropriate through

change of habit and environment, changes would first ap-

pear in these teeth themselves, and in the tissues which

immediately support them. In fact, in the fossil ‘‘Hei-

delberg’’ jaw the teeth have been reduced faster than the

jaw itself.? Granted that our ape, ancestor had a jaw, it

fo os

f ie sag j

i f

i í

1 {

l i

^ / \

yi oe i

\ r

Fic, 6. Lower jaw of the gorilla (broken line), compared with the lower jaw

of recent man (adapted from Schoetensack). If the recession in successive

gorilla jaws were more rapid in the upper than in the lower border, a prominence

would be produced as shown in the human jaw.

242 THE AMERICAN NATURALIST [Vow. L

is to be expected, it seems to me, that during a general

contraction in its size, the superior margin would retract

more rapidly than the inferior. I am inclined to think

that the chin is the persistent lower margin of our large

ancestral jaw. This margin has become retracted more

slowly than the upper margin, and therefore juts out into

space (Fig. 6).

A difficulty immediately suggests itself. If a human

chin results from reduction in the size of the jaw, wher-

ever in different species of animals jaws have become re-

duced, we ought logically to find chins. One striking case

can be cited in line with this suggestion. The case of

the elephant suggests itself at once, and very clearly. |

We know that his jaw bones are the result of a remark-

able retraction. The process is one of the most pic-

turesque that we know about.t We duly find in the geo-

logically recent elephants, and especially in the living

species, a tremendous chin (Fig. 7). Whether still other

Fic. 7. Face of the Indian elephant (recent), showing the presence of a con-

spicuous chin which has resulted from reduction in the length of the jaw.

cases of chins resulting from retraction could, or could

not, be cited, I do not know. I very strongly suspect, how-

ever, that a thorough knowledge of paleontology would

put one in position to cite a considerable number, though

possibly few cases would be so clear as that of the ele-

phant. I can not resist the feeling that in some such

process we have the explanation, not only of human chins,

but the chins of other animals as well.

3 A fact mentioned by MacCurdy in the Smithsonian Report for 1909 —

(page 570).

Interestingly described by Sir Ray Lankester, ‘‘ Extinct Animals.’’

HYBRIDS OF THE GENUS EPILOBIUM

R. HOLDEN

NEWNHAM COLLEGE

Tmar hybridism and sterility are closely related has

been long recognized in a general way, but it is only

within the last few years that a systematic and com-

prehensive investigation, at least of the plant kingdom,

Fic. 1. Flower of Chamenerion angustifolium, Southern Ontario.

has been attempted. Professor Jeffrey and his students,”

working on the flora of eastern North America, have

1 Jeffrey, E. C., ‘‘The Mutation Myth,’’ Science, N. S., 39: 488-491, 1914;

“Spore Conditions in Hybrids and the Mutation Hypothesis of De Vries,’’

Bot. Gaz., Vol. LVIII, No. 4, Oct., 1

Holden, R., ‘‘Anatomy as a Means of Diagnosis of. Spontaneous Plant

Hybrids,’’ Science, N. S., 38: 932-933, 1913; ‘‘ Anatomy of a Hybrid Equise-

tum,’’ Amer. Jour. cf Bot., May, 1915.

243

244 THE AMERICAN NATURALIST [ Von. L

demonstrated that the infertility of hybrids is due to the

abnormal development of the gametic elements, partic-

ularly the pollen grains, and have shown that whenever

the purity of a species is unquestionable, the spores are

uniform, in both size and shape, while, conversely, the

spores of hybrids are usually irregular, some appearing

normal and others being shrunken and devoid of proto-

Fig, 2. Half of Anther of C. angustifolium, Southern Ontario.

plasm. During the past year the writer has extended

these investigations to include a considerable number of

English species. Many interesting cases have been en-

countered, which will be elucidated in detail on another

occasion, but the conditions in the genus Epilobium are

so diagrammatic and typical, that it seems advisable to

describe them now.

This genus is divided into two sections, Chamenerion

and Epilobium proper, the chief differences being that in

the former the flowers are irregular and the spores not in

tetrads, while in the latter the flowers are regular and

the spores are persistent as tetrads. Both in eastern

No. 592] HYBRIDS OF GENUS EPILOBIUM 245

North America and in England, the former section is rep-

resented only by E. angustifolium (L.), while the latter

includes numerous species. Moreover, although the

species of the Epilobium section are generally recog-

nized to hybridize freely with one another, they do not

hybridize with the Chamenerion section. Accordingly,

one would expect to find only good pollen in the anthers

of E. angustifolium, and a mixture of good and bad in

all the others. Investigation of the North American

forms showed that such was indeed the case, and photo-

micrographs illustrating these conditions were published.*

When the writer came to examine English specimens,

however, a different state of affairs was discovered.

Abortive spores were found not only in E. montanum, E.

parviflorum and E. hirsutum, as might have been an-

Fic. 3. Flower of Chamanerion angustifolium (Hardwick, England). Showing

shrivelled anther.

ticipated, but also in E. angustifolium. E. angustifolium

grows wild in only two localities in the vicinity of Cam-

bridge, Hardwick and Gamlingay, but in specimens from

2 Loe. cit.

246 THE AMERICAN NATURALIST [ Vou. L

both these places, as well as in others from the botanical

gardens of Cambridge University, the same mixture of

good and bad grains was found. These facts seemed to

invalidate the conclusion that abortive spores are an in-

variable sign of hybridism, but, as has so often been the

case in scientific matters, evidence which at first seems to

discredit a given hypothesis, on further investigation is

seen actually to corroborate that same hypothesis.

Reference to systematic works shows that there are two

varieties of E. angustifolium growing in England, E.

Fie, 4. Transverse section of Chamenerion ola tease showing abortive

pollen. (Hardwick, England

‘macrocarpum (Steph.) and E. brachycarpum (Leight).

There are a number of minor differences in the length

of the stolons, shape of the leaves, flowers, ete., but the

most definite is the relative length of capsule and pedicel.

E. macrocarpum? grows sparingly but generally from

Somerset and Hants. to Orkney, while E. brachycarpum,

though cultivated commonly all over England, is found

much more rarely in the wild condition, being recorded

3 Boswell, Syme, ‘‘ English Botany.’’

No. 592] - HYBRIDS OF GENUS EPILOBIUM 247

from Shropshire, N. Wales, Yorkshire, and even near

Edinburgh. Through the kindness of Dr. Wilmott the

writer was able to examine the spores of a considerable

number of both these varieties from specimens in the her-

barium of the British Museum, and in every case the

anthers contained a mixture of normal and abortive

grains.

We have here, then, a very interesting condition—

wherever the two varieties of E. angustifolium are pres-

ent, the spores are partially abortive—indicating the bar

sinister; this state of affairs is found in England, and

probably in Europe, Asia and western North America,

where both varieties are known to coexist. Wherever,

on the other hand, as in southeastern North America,

there is but one variety, the spores are all normal. Cha-

menerion, therefore, instead of discrediting the value

of abortive pollen grains as a test for hybridism, affords

another instance of its value. It also suggests another

question—how far apart genetically must individuals be

before the spores begin to degenerate? Hitherto it has

been assumed that only crosses between recognized species

bring about that result, but in the case of Epilobium, the

varietal difference appears sufficient. This, however,

opens up the whole question of what is a species, and can

not be entered upon here.

SHORTER ARTICLES AND DISCUSSION

CAN SELECTION CAUSE GENETIC CHANGE?

Ir is almost a pleasure to have occasion for controversy with a

fellow worker who shows himself so fair-minded and generous

an opponent as does Dr. Pearl in the AMERICAN NATURALIST for

February, 1916. He credits my investigations with greater

merits than I have claimed or can claim for them. If they pos-

sess any superiority, it is not because they have been either better

planned or better executed than Dr. Pearl’s, but only because the

material used was more favorable. In my experiments with rats

I have simply undertaken a less difficult task than that under-

taken by Dr. Pearl in relation to the fecundity of fowls. Pearl

is right in supposing that I have no desire to convey the impres-

sion that his work is valueless. No one has greater admiration

than I for the masterly way in which he has analyzed the funda-

mental problems of genetics and the thorough and systematic way

in which he has attempted their solution. I regret only that he

has courageously attacked so complex a problem before certain

simpler and more elementary ones had been solved. I felicitate

myself only on having been content with a less ambitious pro-

gram.

I am pleased to learn too that we are so closely in agreement

as regards the observational facts, that in reality it is only con-

cerning the interpretation of results that our views seriously

differ.

I am quite ready to grant that we are concerned with the same

fundamental question, that of the possible quantitative change

in a character under selection, that the methods which we have

employed are substantially the same and that these methods are

open to similar objections, that random sampling occurs in the

rat experiments as well as in those with fowls, though it is in-

volved in a further degree in the experiments with fowls because

of limitations of age and sex. I am quite willing that Pearl

should recall the statement ‘‘that phenotypic variation of the

character fecundity in fowls, markedly transcends, in extent and

degree, genotypic variation,” and that he should substitute in

its stead the statement that it ‘‘may’’ so transcend. I am even

248

No.592] SHORTER ARTICLES AND DISCUSSION 249

more ready to concede the existence of genotypic variation in

this character than Pearl has shown himself to be. And I have

been reluctant to accept at its face value Pearl’s statement that

at the conclusion of his fecundity selection experiments he had

more good winter layers than at the beginning, but none better.

For in our selection experiments with rats it is very clear that

when high-grade individuals grow common, a few individuals of

higher grade are sure to put in an appearance. Genotypic varia-

tion seems to me to be of such wide occurrence that it is difficult

to believe that it is ever wholly absent, that absolutely pure lines

really exist. I quite agree with Pearl’s conclusion that somatic

character is not a sure index of genetic constitution and that it

was therefore entirely logical and necessary for him to make

progeny tests in order to classify his pullets genetically. To

establish the point it is not necessary for him, as he observes, ‘‘to

be fussily nasty’’ by citing page after page from my Mendelian

writings. I had granted the point years before it was raised.

This brings us again to what Pearl considers ‘‘the most serious

phase of Castle’s attack, namely that in which he denies the va-

lidity of my conclusions respecting the inheritance of the char-

acter fecundity in fowls.’’ Let it be made very clear at the out-

set what is attacked. Not the idea that fecundity is inherited.

I think that I am even more ready than Pearl to admit that

fecundity is a quantitatively variable character and that its va-

rious quantitative conditions are inherited. This is merely to

State in another way that genotypic as well as phenotypic varia-

tions in fecundity occur. If they occur, it is possible to isolate

them and thus to produce families characterized by them. The

conclusion which I ‘‘attack’’ is this, that the observed variations

in fecundity depend upon two and only two differential factors,

both of which are Mendelian, one sex-linked and the other not sex-

linked. Several possibilities are conceivable, which this conclu-

sion does not include, as for example that more than two genetic

factors are concerned in the variation, that one or other or both.

of the supposed factors are quantitatively variable and so capa-

ble of gradual change under selection. I am not advocating or

defending any of these possibilities. I am merely attacking the

conclusion outlined substantially as I understand Pearl to hold

it. There are really several distinct points in this conclusion,

Some of which seem to be better grounded than others. If I were

asked either to accept or to reject it as a whole (and Pearl’s pub-

250 THE AMERICAN NATURALIST [Von L

lished data leaves no alternative to this) I should reject it, and

this decision would not be influenced by the consideration that

Morgan, Doncaster, Johannsen and Plate accept it, because it

accords with the conception of the pure-line which they have

adopted. Authority does not count in science. Majorities do

not decide what is true. If they did, Mendelism would have

been false in 1868 and true in 1900. If Morgan and Johannsen

should next week decide against the pure line idea, as Jennings

has already done, what could the rest of us then do except change

our minds too, if we base our scientific judgments on authority ?

Dr. Pearl, I am sure, would be the last to advocate such an idea.

I grant to Pearl the legitimacy of his method in attacking the

problem of the inheritance of fecundity and the necessity of estab-

lishing arbitrary categories of winter egg production in which

his birds are then classified. But I regret what seems to me to

be the needless restriction of his published data to the contents

of these categories. Pearl points out that I too have made use

of arbitrary categories in dealing with the rat statistics, but I

would call attention to this difference in our procedure. My cate-

gories, + 1, + 2, ete., are indeed arbitrary, but I have not limited

the reader’s information to their contents. I have published the

data in such form that the reader may, if he chooses, form new

categories with different inclusiveness, subdividing each category

and then subdividing these again down to the lowest limit of

Observation which can be made with certainty. Pearl has not

made it possible for us thus to deal with his data. We may take

it or leave it, but we can not change it. We have no means of

knowing how many pullets laid 1-10 eggs in their first winter,

how many laid 11-20, or 21-30 eggs. In what particular are

- these ‘‘original records” which Pear] withholds ‘‘valuable’’ ex-

cept as proof of the conclusions which he sees fit to base on them?

Tf he decides, as announced, that the data are not to become pub-

lic property until he has finished his own study of them, he is

Pearl seeks to offset his own sin of omission by charging a like

offence upon me, maintaining that the scientific publie withholds

acceptance from my conclusions concerning the rat selection ex- _

periments solely because I have never presented my results ‘‘in

No. 592] SHORTER ARTICLES AND DISCUSSION 251

such form that any other interpretation of the data could by any

_ chance be tested.’’ If this statement is true, it is because of my

inability to devise any other form in which to present the data.

I have presented it in such form that the limits adopted for the

categories of variation could be shifted at will and I am ready to

be shown how its presentation can be further improved and sim-

plified. Pearl suggests that my omission pertains to the individ-

ual pedigree of the rats, in which suggestion he echoes a thought

of the Hagedoorns on which I have twice commented elsewhere,

showing, I think, that the alleged defect does not exist, for the fol-

lowing reasons:

1. It is impossible for a colony of 33,000 rats to be produced

from an original stock of less than a dozen animals, with con-

stant breeding together of those which are alike in appearance

and pedigree, and with continuous selection of extremes in two

opposite directions, without the production of pedigrees which

in the course of each selection experiment interlock generation

after generation and finally become in large part identical with

each other. This has been repeatedly verified in individual cases,

but is incapable of a more generalized statement or of demonstra-

tion in generalized form. At least I am unable to devise such

demonstration.

2. In a specifie case described on pp. 20 and 21 (Castle and

Phillips) a selection experiment was started with the hooded F,

offspring of a single selected hooded and a single wild rat and

this experiment was carried through the F, generation leading to

the production of 804 young from rigidly selected, closely inbred

descendants of a single pair. We showed (l. c., p. 21) that the

progress of selection within this inbred family follows a remark-

ably close parallel, generation by generation, to the progress of

selection in our plus series as a whole. Here there can be no

question of a difference in pedigree among the selected animals.

This is eliminated as a possible factor in the result. Can Pearl

suggest any other possible factors capable of elimination? If so,

I should be pleased to give attention to them.

I humbly beg pardon for having made the all too obvious sug-

gestion that environmental conditions, and in particular size of

flock, may affect average flock fecundity. And yet I find that

Pearl himself elsewhere lays great stress on this point. My chief

offense seems to lie in my failure to realize that he had already —

taken all possible precautions in this matter, and that he consid-

252 THE AMERICAN NATURALIST [ Von. L

ered himself in a position to vouch for the uniformity of environ-

mental conditions, not only in eight years of experiments which he

had personally superintended, but also in nine previous years of

experiments of which he had neither control nor information

until they were completed and which were made sometimes in

50, sometimes in 100, and sometimes in 150 bird flocks. Are

there not here some elements of uncertainty which at least con-

done the offense even if they do not excuse the question?

I am prepared to accept without question Pearl’s statement

that date of hatching can not possibly have had anything to do

with the rise in average flock production which has occurred be-

tween 1908 and 1915, notwithstanding his own previous state-

ments on the subject and the evidence which Phillips has pro-

duced that date of hatching of ducks affects their adult size. I

am prepared to accept the view that this rise was due wholly to

genetic changes, but I do not believe that Pearl or any one else is

in a position to say to what agencies the decline previous to 1908

was due.

And now, with Pearl, I turn with pleasure to the general

problem of selection and note that our differences are here rather

verbal than real. They lie in that philosophic pitfall of causa-

tion.

Pearl can not conceive that selection may cause or occasion or

lead to genetic change, though he can readily see how popula-

tions may change under its influence. Thus selection may in-

crease the proportion of high-grade individuals but it can not,

on his view, beyond a limited and fixed point, occasion the pro-

duction of individuals of increased grade. With these views I

squarely take issue, and I shall try to show that his view is a

purely a priori view, while mine is based on both observation and

experiment.

Pearl’s reasoning throughout rests on the assumption that the

potmtimiiy of a germ cell can not change except by a causeless

method, ‘‘mutation’”’; that no extraneous influences can change

it. Experience tianka directly the contrary, indicating that

germ-cells brought together in fertilization mutually influence

each other. Let us consider for a moment Pearl’s illustration.

He supposes an organism to exist, A,,, which is producing ga-

metes of the uniform value, ass, and can not understand how su

gametes uniting with each other can ever produce individuals of

a higher value, say Ass. No more can I, if we accept his further

No. 592] SHORTER ARTICLES AND DISCUSSION 253

hypothesis that there is to be ‘‘no mixing of germ-plasms.’’ But

what justification have we for that further hypothesis? Experi-

ence furnishes none. On the contrary, I have shown in numer-

ous specific cases that when unlike gametes are brought together

in a zygote, they mutually influence each other; they partially

blend, so that after their separation they are less different from

each other than they were before. The pure-line advocates have

adopted the procedure of dismissing such explanations as mysti-

cal, an easy way to dispose of troublesome ideas. But the stub-

born fact remains to be accounted for that partial blending does

occur (1) when polydactyl guinea-pigs are crossed with normals

` (Castle, 1906), (2) when long-haired guinea-pigs are crossed

with short-haired ones (Castle and Forbes, 1906) and, (3) when

spotted guinea-pigs or rats are crossed with those not spotted

(MacCurdy and Castle, 1907). Davenport has furnished numer-

ous instances of the same thing in poultry ; indeed he has shown

that ‘‘imperfection of dominance’’ and of segregation are the

rule rather than the exception in Mendelian crosses in poultry.

To assume that ‘‘there is no mixing of germ-plasms”’ is a con-

trary-to-fact assumption, whatever it may be in formal logic or

scientific methodology.

Let us change slightly Pearl’s hypothetical case. Let us sup-

pose, as he does, that a gamete ass has united in fertilization

with another a,, gamete producing a soma, Ags. Now what sort

of gametes may we expect such an individual to produce? Pearl

says, in effect, nothing but a,, gametes, unless a genetic miracle

occurs, a mutation, incapable of casual explanation. But we

should hesitate to characterize as miraculous anything which oc-

curs with regularity, and experience shows that this is what hap-

pens quite commonly, if not regularly, in such cases. The A, in-

dividual produces gametes a majority of which have the value ays,

but a few of which have a higher value, a,5, and a few a lower

value, a. For the correlation in value between soma and

gamete is not absolute. It is in many cases close, but not in-

variable, as I think Dr. Pearl would admit. If it be granted for

the sake of argument that gametie variation occurs, it is obvious

that we have grounds for expecting somatic variation in the fol-

lowing generation. For an a, gamete uniting with another

gamete like itself may be expected to produce a zygote of value

A... Pearl maintains that such an event is without ‘‘causation,””

is ineapable of prediction and control, that all we can do is to

254 THE AMERICAN NATURALIST [Vou. L

record its occurrence, a view I by no means share. But, it may

be asked, what control can we exercise over the event? We can

prevent or permit it at will. For observation shows that if we

permit the individual to mate only with those of inferior value,

we shall get no offspring of the highest grade. Thus agg + agg

produces commonly only A,,, rarely A,, and, we might say, never

Az. But if we permit the individual to mate with individuals

of equally high grade (and this is what selection in a particular

direction does) experience shows that a majority of the offspring

will be of that same grade, but a few will be of higher grade.

These few make possible further advances. Thus a,, makes pos-

sible the subsequent attainment of a. Whether this relation-

ship involves ‘‘causation’’ or not is a question for the logicians

and methodologists, of whom I am not one. As to the fact our

rat experiments leave no doubt. In the light of such facts it

seems to me that a view earlier held among biologists, that vari-

ability is one of the fundamental properties of organisms, comes

nearer to the truth than this modern notion of the pure, unvary-

ing line. This pure-line concept Pearl rightly characterizes as

“fone of the most useful working tools in the practical breeding ~

of plants and animals that has ever appeared.’? Why useful?

Because it has caused us to pause and take careful inventory of

our facts, and to discard as rubbish many loosely held notions.

But Pear] reminds us that not all the pure linist’s facts are in one

basket with Johannsen’s beans, nor even in that other vanished

basket with Jenning’s paramecia. There are, he reminds us,

“‘all the Svalöf oats and wheats to be reckoned with.” True and

they are mute witnesses to the cumulative effects of selection.

For all agree that these pure lines of oats and wheats are the

product of continuous self-fertilization. And what more than

self-fertilization renders possible generation after generation the

union of gamete with its like, the indispensable condition for pro-

gressive variation in a particular direction, as I have tried to

show?

Intelligent selection only accelerates this natural process of

progressive variation, for it singles out the individual which is

producing gametes of unusual value and permits the union of

such high grade gametes only with gametes of their own sort, so

that step after step in a particular direction becomes possible,

where unguided self-fertilization would give only halting and

uncertain progress. Can we doubt that it is progressive varia-

No. 592] SHORTER ARTICLES AND DISCUSSION 255

tion guided by rational selection in a particular direction that

has made possible the doubling in size that most of our domesti-

cated animals have undergone since they were taken from the

wild state? And does any one seriously think that a single selec-

tion from wild stock has produced for us the enormous horses of

Flanders, or the little ponies of the Shetland Isles, the enormous

sheep of the Scotch highlands, or the huge rabbits of Europe,

each a monstrosity in comparison with its most probable wild

ancestors, and yet producing blends in crosses with them? This

blending shows that the change has been one of slow accomplish-

ment and not the result of sudden discontinuous change.

When we compare the color varieties of domestic animals with

those of their wild ancestors, as I have been able personally to

do in the case of cavies, we are struck by the fact that the domes-

tic varieties are relatively clear and distinct in color, either more

intense, more dilute, or of purer color than we can obtain from

the wild form by simple recombination of genetic factors. For

example it is possible by crosses to obtain from wild cavies the

retrogressive varieties, black and yellow. But such synthetic

blacks lack the full intensity of blackness found in our best

strains of black guinea-pigs, and the synthetic yellows are apt to

be either pale or muddy in yellowness, lacking the intensity and

brilliancy of our best domestic varieties. It is impossible to

escape the impression that our improved domestic varieties are

not mere factorial recombinations derived from wild species, but

that they have been forced up to a higher standard by repeated

selection; that the breeder, for example, has first observed vari-

ation in intensity of blackness among his blacks (doubtless ob-

tained originally from a retrogressive sport) and that by re-

peatedly selecting the blackest available individuals he has in-

creased the blackness of the race. Thus it is no accident that

the meat and milk and wool producing capacity of our domes-

tic animals far exceeds that of any wild ancestral species. The

standard in each case has been raised and it has not been raised

by a single lucky accident (the mutation view), but by a series of

slow advances each impossible until a previous advance had been

made. I am aware indeed that Pearl at one time maintained an

opposite view, holding (if I remember correctly) that our best

strains of poultry are no better layers than some strains of jungle

fowl. But I do not believe that this view can be successfully de-

fended. I am certain that such an idea is quite preposterous in

256 THE AMERICAN NATURALIST [Von L

the case of most characters for which our domestic animals are

valued and as regards which their improvement has been at-

tempted by selection. In such cases there has been a series of

slight advances, and everything indicates that the order of the

advances is significant and necessary, that the higher stages can

be attained only by passing through the lower ones. -If this is so,

we need not quibble about ‘‘causation,’’ but we may assure our-

selves that if we wish to attain a distant goal, the first thing to do

is to make for intermediate points.

I regard it as a hopeful sign that Pearl can see no reason why

genetic changes may not be small in amount in some cases, even

though large in others. This I hope is only a first step toward

the complete abandonment of that ‘‘real, genuine pure-line body

of doctrine’’ which he still holds dear.

W. E. CASTLE

BUSSEY INSTITUTION,

Forest Hinzs, Mass.,

February 16, 1916

VOL. L, NO. 593 MAY, 1916

THE

AMERICAN |

NATURALION »

| Devoted to the Advancement of the Biological Sciences with —

The American

Naturalist

MSS. intended for peepee se and books, ete., intended for review should be

sent to the Editor of TH

Short articles containin

RICAN vahirseaie ge Garrison-on- Hudson, New York.

summari

s of research work bearin on th

problems of Real peo evolution are sapectalty welcome, and will be given preference

et yur

gsi ndred reprints of renee gee are supplied to authors free of charge.

F urther reprints = be supplied at cost

Subscriptions and advertisements should be sent to the gnc The

subscription sta is four dollars a year. Foreign paee is fifty cents and

Canadian postage twenty-five sbin additional. The c =n for single poo is

forty cents. The advertising rates are Four Dollars for a pa

THE SCIENCE PRESS

Lancaster, Pa.

Garrison, N. Y.

_ NEW YORK: Sub-Station 84

, April 2, 1908, at the Post Off

tI ancaster, Pa , under the Act of

Congress of March 3, 1879.

BOOKS FOR SALE)

JAPAN NATURAL HISTORY SPECIMENS

thas bh ae bce Perfect Condition and Lowest Prices.

Specialty: Bird Skins, Oology, Purtanainu’s Marine

NATURAL HISTORY Animals and others. Catalogue free. Correspond-

Catalogue on applicat ence solicited.

JORN D. D. SHERMAN, Jr. T. FUKAI, Naturalist,

403 Seneca MOUNT VERNON, N. Konosu, Saitama, Japan

PUBLISHED. BY THE CAMBRIDGE UNIVERSITY PRESS |

Eea IN BUTTERFLIES

SE: Bound in Buckram. e SDR Pp. un With 16 plates (12 of which

= 4 By eats ET Pisci, FR D. Fellow of Gonville a samme College,

F oes Balfour Professor of Genetics i in the University of Cam -

o- ih ~ EXTRACT FROM THE PREFACE | a I

I| £ tis Tittle book has been written to sev ue

F classes 0 in the hope that it may pet ae

quently 1 ł [ have ‘been asked by friends oF am different callings in life to F

) na bly short, well illus- |

aoe not too hard to piesen G I have always |

= I know Of Smee in our anr ee to |

THE

AMERICAN NATURALIST

VoL. L. May, 1916 No. 593

GENERAL BIOLOGY OF THE PROTOZOAN LIFE

CYCLE!

GARY N. CALKINS, Px.D.

PROFESSOR OF PROTOZOOLOGY IN COLUMBIA UNIVERSITY

For five decades after the time of Ehrenberg, the pecu-

liar conception of a protozoan as a miniature replica of

a metazoan, held by this gifted observer, influenced the

study of Protozoa. This influence gradually wore off

and, so far as morphology is concerned, ended with the

careful observations of Stein, Claparéde and Lachmann,

Engelmann, Biitschli and Hertwig, who showed that

various structures of the protozoan body are not beating

hearts, brains, ovaries and stomachs, but are simple dif-

ferentiations of the single-celled organisms.

A more lasting influence of Ehrenberg’s teaching, seen

even to-day, is the habit of regarding a single protozoon

as the complete expression of a species equivalent to an

individual worm, molluse or mammal. The individual

metazoon dies, while the protozoon does not die but grows

to full size and divides into two or more—facts which led

Weismann to his conclusions regarding mortality in

Metazoa and immortality in Protozoa.

We owe to Maupas the credit for dissipating this last

reminiscence of Ehrenberg’s teaching, and for showing

that the single cell is not the final representative of a pro-

tozoon species. We are accustomed to the idea that many

1 Opening address Subsection E, Protozoology, Section VIII 2nd Pan-

Amerian Congress

257

258 THE AMERICAN NATURALIST [ Vou. L

individuals of a polymorphic coelenterate are present in

potential in the fertilized egg of the coelenterate, but we

are less accustomed to the idea that polymorphic indi-

viduals are present in potential, in the fertilized cell of a

protozoon. Research in recent years has shown that suc-

cessive generations of Protozoa may be more or less pro-

gressively differentiated, so that a cell picked out at one

phase of the life cycle is quite a different type of indi-

vidual from one picked out at another phase. Which, for

example, would be the ‘‘type’’ individual of the dimorphic

Foraminifera? Which would be the type in the repro-

ducing flagellated and ameboid stages of Nägleria punc-

tata? of different phases in the life history of Centropyzis,

Arcella, or Difflugia? or of intestinal and blood-dwelling

stages of Plasmodium? The morphological differences

here indicate that the protozoan life history involves -

differentiation analogous to that of a polymorphic meta-

zoon, and justify the comparison of the whole life cycle

with the development and differentiation of a metazoon,

especially that of a metagenetic type such as coelenterate

or trematode.

The importance of the whole life cycle, first demon-

strated by Maupas, was fully recognized by Schaudinn

and applied by him to the study of parasitic forms. The

monographs resulting from this study, especially those

on Coccidium schubergi, Plasmodium vivax and on rhizo-

pods, are classics in the literature of Protozoa, and models

which later students have followed.

Through Schaudinn’s work, and by later researches,

the sequence of events in different parasitic types has

been made out with painstaking care until to-day, we

know the general history of the majority of injurious

human protozoan parasites, the modes of transmission

from host to host, the types of intermediate hosts and

what happens in them. In short, we know enough to

furnish an adequate basis for public and private prophy-

laxis which, in the hands of sanitary commissioners and

public health officers, has put an end to epidemics of yel- _

No. 593] THE PROTOZOAN LIFE CYCLE 259

low fever, malaria and dysentery; has rehabilitated vast

tracts of land in Italy; saved millions of dollars in South

Africa and in our southern states, and has made the

Panama Canal possible.

Such are the first, and practically the most important,

results of our knowledge concerning protozoan life cycles ;

quite enough, indeed, to justify the science of Protozo-

ology. Important as these results are, we are not at all

satisfied ; we know too little about the conditions of devel-

opment; too little about the nature of the vital processes

of the organisms themselves and their variations in struc-

ture and function under differing conditions, ignorance

which must be cleared away before much further practical

advance can be made. Further advance will be less spec-

tacular and must be based upon the biological study of

the organisms as units of protoplasmic substance, and

this will rest upon working hypotheses supported by ex-

periment. It is along such theoretical lines that I wish to

direct your attention for a few minutes, to develop a con-

ception of the life cycle as a whole, and to offer a theo-

retical interpretation of the different phases of vitality |

and of structural variations.

Let us consider for a moment, a single Ameba or a

malaria germ, not as a cause of disease, but as a unit mass

of protoplasm which, like a free-living Paramecium or

Didinium, performs all of the fundamental vital activities

common to living things, namely nutrition, excretion, irri-

tability and reproduction. The chemical composition of —

these unit masses, so far as I know, has never been made

out, but there is no reason to doubt that it agrees with

that of other living substances, since the accompanying

properties of protoplasm—metabolism, growth and re-

production—are obviously performed, and probably in

the same way. In such unit masses of protoplasm we

assume that processes of hydrolysis, synthesis, oxidation

and reduction, are constantly going on as in other proto-

plasms, and not in any haphazard way, but always orderly

and under regulative control of the organism as a whole. >

260 THE AMERICAN NATURALIST [Vou. L

The appearance of Ameba shows that the protoplasm is

made up of alveoli and inter-alveolar substances of differ-

ent density, representing colloidal and crystalloidal sub-

stances in a general mixture which Ostwald describes as

an emulsoid. Between these different substances con-

stant chemical activities are in progress, and the order-

liness which distinguishes these processes in the proto-

plasm of the living organism from similar processes which

go on in the same protoplasm when crushed, are possibly

due, as Mathews states, to the physical barriers of cellular

and nuclear membranes, alveoli, and the colloidal centers

of activity. The speed with which such processes take

place in living protoplasm, which, in itself, distinguishes

living processes from chemical processes in lifeless sub-

stances, is due to specific enzymes or catalyzers which

are manufactured as aresult of chemical activities in living

protoplasm. These bring about and control each suc-

cessive step in the long chain of chemical actions involved

in destructive metabolism, the action in each event being

conditioned by the nature of the protoplasmic substratum.

In this chain of destructive processes different substances

may be formed which undergo no further oxidation or

other chemical change, but are stored up in the proto-

plasm until disposed of by excretion, these products, lead-

ing to changes in the protoplasmic substratum, i. e., to

protoplasmic chemical differentiation, may or may not be

accompanied by visible structural differentiations. Such

products of destructive metabolism, in the form, usually,

of nucleo-proteins or their derivatives, may act as poisons

to other organisms, as melanin does to the host in ma-

laria, or as the proteolytic ferments of Entameba his-

tolytica do in dysentery; or they may play some impor-

tant part in the vital activities of the organism itself, as

in phosphorescence of Noctiluca and the dinoflagellates,

or more generally, in regeneration and reproduction.

Let me illustrate this latter point by some experiments

made on Uronychia transfuga, a ciliated protozoon. This

organism has rather a complicated structure with nine

No. 593] THE PROTOZOAN LIFE CYCLE BOL

giant cirri at the posterior end (Fig. 1). Under labo-

ratory conditions it divides once a day approximately, or,

more exactly, once in twenty-six hours. The first indica-

tion of division is the precocious formation of the giant

cirri in a central region of the body which we have called

the ‘‘division zone.” The experiments were undertaken

for the purpose of studying the relative power of regen-

eration of the single cell at different ages between divi-

sions, it having first been determined that the cell regen-

erates readily after being cut. Cells were cut with a scal-

pel at different periods subsequent to division; some

during the end stages of division; some 15 minutes after

division; some one hour after; others 2, 4, 8, 12, 16 and

20 hours after, and some were cut just prior to the next

division period, i. e., 24 to 25 hours after division. In all

-Cases of record, the cells were so cut that one portion con-

tained the micronucleus and part of the macronucleus, the

other portion containing only a part of the macronucleus.

The former, or, as I shall call it, the nucleated portion,

invariably regenerated after some hours, forming a per-

fect cell, the latter, without a micronucleus which I shall

call the enucleated portion, behaved differently as regards

regeneration, according to the age of the cell when cut.

In all cases this portion lived from three to five days after

the operation. If the recently divided cell were cut at

any period up to 16 hours after division the result was the

same; no regeneration occurred, the fragment merely

rounded out, swimming about by its adoral membranelles

(Fig. 2, 3). If the cells were cut when from 18 to 24

hours old, regeneration occurred not only in the nucleated

portion, but in the enucleated fragment as well, the per-

centage of regeneration increasing with the increased age

of the cells when cut, until at the age of 24-25 hours the

enucleated fragments regenerated perfectly in 100 per

cent. of cases (Fig. 4, 5, 6, 7).

These results indicate a gradual chemical differentia-

tion of the protoplasm as a result, probably, of destruc-

tive and constructive metabolic processes. The giant

262 THE AMERICAN NATURALIST [Vou. L

Normal adult individual of Uronychia transfuga with macro-

rge chromatin spherules; micronucleus (m); endoral

membrane (E); pre-oral membrane (P); and large posterior cirri.

2 3. Individual 12 hours old c n in 2. Part A had

- 4, 5, 6 AND 7. Individual cut at age of 25 hours as shown in 4. A

regenerated y, except for absence of micronucleus, in 24 hours (5 A);

B divided through the original division plane (indicated in 4), within a few

hours forming a minute but perfect individual (6B’), and a normal full-size

individual (7 B”).

No. 593] THE PROTOZOAN LIFE CYCLE 263

cirri which are regenerated are the visible expression of

inherited structures characteristic of the species. Since

the enucleated fragment from a cell cut when young does

not regenerate while the nucleated fragment does, we

must conclude that one essential factor at least, necessary

for the production of these inherited structures, lies in the

micronucleus.

The giant cirri, furthermore, are visible differentiations

which are precociously formed at division. This must

mean that the inherited factors find their expression at

this period, and it follows from the successful formation

of giant cirri in enucleate fragments from old cells, that

whatever may be the direct causative agent or agents in

the process they must be generally distributed throughout

the protoplasm at this time. We have no direct evidence

as to what these agents may be; possibly there is only one

and that of the nature of a specific enzyme, or perhaps

some chemical body analogous to hormones formed as a

result of mutual interaction of nucleus and cytoplasm

when the latter has reached a certain stage of chemical

differentiation through normal activities. Or it is possible

that such chemical bodies are present at all times and are

activated only when the protoplasmic substratum reaches

some particular stage of development. Thus it is possible

that, with continued metabolism, the acidity of the proto-

plasm gradually increases until a concentration is reached

in which specific enzymes, not able to act before, are now

activated.

However theoretical the interpretatio of the phenom-

enon may be, the periodic and temporary power of re-

generation is an observed fact indicating a difference in

the protoplasmic make-up at different age periods, a dif-

ference which may be satisfactorily expressed by the

phrase cumulative chemical differentiation.

Another observed fact is that the regenerative power

is exhausted with cell division, for young enucleated frag-

ments do not regenerate. This indicates a reduction of

the differentiated adult protoplasm to the condition of

young cells; or, at least, the protoplasm is restored to a

264 THE AMERICAN NATURALIST [Vou. L

state where the causes underlying regeneration are inac-

tive. This may be due to the exhaustion of specific sub-

stances which take part in the reaction of regeneration, or

it may be due to the chemical and physical changes accom-

panying cell division.

We are led through these experiments, to further specu-

lations concerning the nature of cell division. Chemical

differentiation of the protoplasm continues even after the

stage is reached when regeneration is possible. This is

shown by the fact that formation of the cirri in Uronychia

precedes the process of division in normal cells, and by

the additional fact that regeneration of cirri occurs while

cell division does not occur in enucleate fragments cut

from old cells. I would interpret cell division as due to

cytolytic action set up by enzymes or other chemical

bodies produced as a result of interaction of nucleus and

cell body differentiated chemically by age. Cytolysis may

then occur more or less extensively throughout the entire

protoplasmic mass, but it is most active in the division

zone of the organism which is more highly differentiated

than other regions (see Calkins, 1911, and Peebles, 1912).

The membrane of the cell turns in at this eytolyzed divi-

sion zone and the constriction results in cell division.

As a consequence of the activities accompanying cell

division the protoplasmic substratum is reduced from the

differentiated adult condition to the condition characteris-

tic of young cells, and the processes of growth and chem-

ical differentiation, division and de-differentiation, recur

in more or less rhythmical succession.

Viewing the life cycle as a whole, there are two phases

which must be taken into account. These are, first, the ©

encystment phase, and second, the sexual or conjugation

phase, both widespread and almost universal in protozoan

life histories. Let us first consider the encystment phase.

Encystment occurs ordinarily when the conditions in

the surrounding medium are adverse, such as desiccation,

lack of food, ete., such encysted forms emerging from the

cyst when suitable conditions are restored. In some cases

also, encystment occurs during the digestion of food. In

No. 593] THE PROTOZOAN LIFE CYCLE 265

addition to these casual encystments there is another form’

of encystment which involves more deeply-lying activities

of the protoplasm. In Didinium nasutum I have found

that encystment occurs at periodic intervals which cannot

in any way be connected with adverse conditions of the

environment or with feeding, but must be interpreted as a

normal phenomenon due to internal conditions of the

organisms. Encystment at such times persists for from

5 to 8 days and during this period no amount of coaxing

will bring the organisms out. During such encystment the

macronucleus fragments into hundreds of small chromatin

particles which are ultimately absorbed in the cytoplasm ;

the micronuclei divide, and products of their division give

rise to a new macronucleus and new micronuclei. When

the process is completed and the organisms emerge from

their cysts they possess from five to seven times the vital-

ity, as measured by the division rate, of the same race

prior to eneystment. Fermor was the first in 1913 to de-

scribe similar happenings during the encystment of Stylo-

nychia; in this case, dissolution of the old macronucleus

and absorption of the fragments, fusion of the two micro-

nuclei and formation of new macronuclei and micronuclei

from the fusion nucleus, were described.

It is well known that Paramecium does not encyst.

Nevertheless Woodruff and Erdmann (1914) have shown

that phenomena similar to those occurring during encyst-

ment in Stylonychia and Didinium, and which they refer

to under the general term ‘‘endomixis,’’ recur at periodic

intervals (about once a month) in the case of Paramecium

aurelia. Here also the old macronucleus fragments and

the fragments are absorbed in the cytoplasm, while a new

macronucleus and micronuclei are formed from the divi-

sion products of the old micronuclei.

The interpretation of this set of phenomena in the life

history of protozoa is a perplexing problem. There is not

a doubt that vitality, as measured by the division rate, is

restored. Likewise there is little reasonable doubt that a

complete chemical and physical reorganization of the pro-

toplasm takes place. The renewal of vitality was shown

266 THE AMERICAN NATURALIST [ Vou. L

‘both in Woodruff’s culture and in my Didinium culture,

and one general problem is stated in the query: how long

can such periods of reorganization continue? Woodruff

believes that they may keep on indefinitely, but in my ex-

periments with Didiniwm the race apparently lost its

power to encyst and ultimately died out after six months’

culture without encystment. So too, in my culture of

Paramecium caudatum (1902) where similar reorganiza-

tion occurred at least twice, the race ultimately lost the

power to reorganize and died out. I may have had un-

favorable forms to start with and so lost both races at

early dates. It is interesting in this connection, however,

to note that Whitney, working with the rotifer Hydatina,

a metazoon, carried a race through nearly 200 generations

by parthenogenesis when the individuals lost their power

to reproduce in this way, and many of his lines died, while

others produced sexual individuals.

The general biological effect of this process of reorgan-

ization is a new chemical combination with a new potential

of metabolic activity, and a new lease of life. Not only are

the nuclei restored to activity, but the cytoplasm is like-

wise completely reorganized by the distribution through

it of relatively large quantities of nucleo-proteins, giving

rise to successive derivatives (through hydrolysis, oxida-

tion, reduction, ete.),.all increasing the metabolic processes

and releasing more chemical energy expressed by activity

of movement and feeding, and leading to more rapid as-

similation and growth, all indicated by an increased divi-

sion rate. In short, the protoplasm is rejuvenated.

The second phase in the life history to be considered,

viz., the sexual phase, involves still more deeply-reaching

protoplasmic activities. The protoplasm of the individual

cells at this period has a different physical, and presum-

ably chemical, make-up than during ordinary vegetative

periods. In free-living forms, such as the ciliates, the -

outer protoplasm becomes sticky or glutinous so that two

cells on touching, fuse together. In this condition which I

have called the ‘‘miscible state” conjugation is possible,

and the physical condition may be so extreme that groups

No. 593] THE PROTOZOAN LIFE CYCLE 267

of cells get stuck together. I have witnessed the fusion

of nine Paramecium caudatum cells in a single amorphous

mass.

In other forms, notably the parasitic protozoa, proto-

plasmic changes at this stage follow two lines of differ-

entiation. Some cells store up metabolic products in the

form of reserves of nutriment and develop into female

gametocytes or macrogametes. Others develop into more

active male gametocytes and microgametes. In both of

these differentiated types if union or fertilization is pre-

vented, the cells die a natural death.

The effects of conjugation or fertilization are almost

the same as those following asexual reorganization through

encystment. In ciliates cytolysis of the old macronucleus

takes place and its substances are absorbed, that is,

undergo chemical changes in the cytoplasm. The ma-

jority of the maturation nuclei, both in free-living and in

parasitic forms, meet the same fate, while a new nuclear

apparatus results from the products of the fertilization

nucleus or synkaryon. The cytoplasm is renewed in a

chemical .sense and metabolic activities recommence with

renewed vigor; a new race is started. The sole difference

from encystment is that reorganization occurs after or

during amphimixis and a new hereditary complex 1s

formed in the nucleus, while even this, in endogamous

conjugation at least, can not be very different from the

condition after asexual reorganization. It is obvious that,

if conjugation is the equivalent of fertilization in metazoa,

asexual reorganization or endomixis is the equivalent of

parthenogenesis.

What is the significance of these two important phases

in the life cycle and how can they be interpreted in terms

of metabolic activities? As we have seen, there 1s reason

to believe that the cell protoplasm becomes progressively

differentiated ina chemical sense between division periods,

until just prior to division processes take place which do

not occur at earlier periods. With division this differ-

entiated condition is reduced, possibly through eytoly S15,

until a more labile protoplasm results. Now it 1s not at |

268 THE AMERICAN NATURALIST [Vou. L

all improbable that such reducing processes are more or

less incomplete, so that the protoplasmic substratum in

the second generation is different from that of the first.

We have evidence of this in the foraminifera where differ-

ences in the protoplasmic structure and in shell structure

characterize the second generation. Further evidence is

seen in the rhizopods, where increasing quantities of chro-

midia, and in some cases differences in shell structure, are

morphological indications of differentiation.

Furthermore, it is not improbable that such differences

are cumulative from generation to generation, just as

chemical differentiation is cumulative with inter-divisional

age, until a protoplasmic substratum is evolved in which

processes not possible before can now take place. Wehave

shown that Paramecium at the conjugation phase has a

different physical make-up than at other times, the cortical

plasm becomes mucilaginous and fusion results on contact,

while physiological differences are manifested by the in-

variably decreasing division rate during and after this

period when conjugation is possible. Here the proto-

plasmic substratum is differentiated, and processes occur

which are not possible at other times. So, too, in Di-

dinium, Stylonychia, ete., with successive generations a

protoplasmic substratum is gradually evolved (possibly

hastened by adverse conditions) in which the peripheral

zone of protoplasm undergoes cytolysis and forms an im-

pervious membrane—the cyst membrane—analogous to

the fertilization membranes of metazoan eggs. Further

cytolytic changes, involving hydrolysis, reduction and

other chemical activities, are set up in the cell body, espe-

cially in the cell nuclei which divide or fragment. Asa

result of these activities, which are more profound than

those accompanying cell division, the protoplasm is again

restored to a labile condition, vitality is renewed and a

de-differentiated protoplasm begins a new cycle of meta-

bolic and reproductive phases.

The phenomena of conjugation may be interpreted in a

similar way as due to processes possible only in a sub-

stratum produced by cumulative protoplasmic differentia-

No. 593] THE PROTOZOAN LIFE CYCLE 269

tion. A visible expression of such differentiation is seen

again in the chromidia formation of Sarcodina and in the

dimorphic gametocytes of foraminifera and Sporozoa.

The reorganization phenomena are quite as complicated

and as far reaching as after encystment, and the end result

is the same, a de-differentiated protoplasm and a new indi-

vidual with a high potential of vitality. If fertilization is

prevented the differentiated macro- and microgametes die

as do metazoan eggs and spermatozoa, and a similar result

follows the continued culture of free-living ciliates in

which conjugation, or its equivalent, asexual endomixis, is

prevented.

In all life histories we find more or less regular cycles

of vegetative and sexual phases, complicated by more or

less active asexual and sexual reproduction. In parasitic

forms it is possible, I may say probable, that reorganiza-

tion and renewal of vitality take place during encysted

stages as Schaudinn, Wenyon and others have held for the

genus Entameba; or, as in Paramecium, they may take

place without encystment in types like Plasmodium as de-

scribed by Schaudinn. The processes of autogamy, 50-

called, described for different types of Entameba, may be

interpreted as asexual endomixis, and the conflicting

views as to the significance of nuclear structures in Enta-

meba coli, E. histolytica, E. tetragena and E. minuta, may

all be reconciled when this possibility of asexual reorgan-

ization is applied to the various parasitic rhizopods.

With Plasmodium, the principle of asexual reorganiza-

tion and renewal of vitality, or parthenogenesis, has long

been called upon to explain malaria relapse. The process,

as described by Schaudinn, is too familiar to need repeti-

tion here. Despite the objections which have been raised

in recent years against this interpretation, it must be ad-

mitted that no à priori difficulty stands in its way. It

is evident from experiments that the protoplasm of an old

race is more stabile than that of a young race, possibly due

to accumulation of products of metabolism in the former,

either for a useful purpose, as in the storage of yolk nigel

terial in a female cell, or for some harmful purpose, as ın

270 THE AMERICAN NATURALIST [Vou. L

Paramecium caudatum during depression. In either case

if a labile protoplasm can be restored resulting in chemical

activities which ultimately bring about dissolution of

these formed products, then renewed vitality is the out-

come. Asexual reorganization effects this result, but the

same result was produced artificially by the use of salts

in my experiments with Paramecium caudatum during

conditions of depression, and in cases where the cell body

was visibly loaded with products which it could not auto-

matically dispose of. The splendid results which Bass has

obtained in cultivating Plasmodium in vitro and in the

presence of sugar, indicate the possibility of malaria

organisms while in a stabile condition being similarly

changed into a labile condition by changes in the blood

content of the host. Changes thus set up might well be

the equivalent of asexual reorganization or partheno-

genesis, or the equivalent of fertilization in restoring

vitality.

In this sketch of the protozoan life cycle I have endeav-

ored to give a comprehensive though somewhat specu-

lative account of the different phases of vitality which may

apply equally well to any type of Protozoa. Cell division,

reorganizing encystment or its equivalent, and conjuga-

tion, are all regarded as phenomena of the same general

character but differing in degree, the effect in each step

being the restoration of the protoplasm to a condition

more or less free from cumulative metabolic differentia-

tions.

REFERENCES

Calkins, G. N. 1911. Regeneration and Cell Division in Uronychia trans-

fuga. Jour. Exper. Zool., Vol. 10, No.

1911. Effects produced ge Cutting Potameciumn. Cells. Biol. Bull.,

fol. XXIL, No. 1.

1915, Didinium nasutum. 1. The Life History. Jour. Exper. Zool.,

Tol. 19, No. 2

Peebles, F. 1912. Regeneration and Regulation in Paramecium. Biol.

Bull., Vol. XXIII., No. 3.

Woodruff, L. L., and Erdmann, R. 1914. A Normal Periodic Reorganiza-

ion Process without Cell Fusion in Paramecium. Jour. Exper.

Zool., Vol. 17, No. 4.

THE EVOLUTION OF THE CELL. Il

BY THE LATE PROFESSOR E. A. MINCHIN, F.R.S.

In the phase of evolution that I have termed the pseu-

domoneral or cytodal phase, in which the organism was

a droplet of periplasm containing scattered biococci or

chromidiosomes, metabolism would result in an increase

in the size of the cytode-body as a whole, accompanied by

multiplication of the chromidiosomes. Individualization

of the cytodes would tend to the acquisition of a specific

size, that is to say, to a limitation of the growth, with the

result that when certain maximum dimensions were at-

tained the whole cytode would divide into two or more

smaller masses amongst which the chromidiosomes would

be partitioned. .

In the next stage of evolution, the protocyte with a defi-

nite nucleus, it is highly probable that at each division of

the cell-body, whether into two or more parts, the primi:

tive method of division of the nucleus was that which I

have termed elsewhere ‘‘chromidial fragmentation’’ ;?°

that is to say, the nucleus broke up and became resolved

into a clump of chromidiosomes, which separated into

daughter-clumps from which the daughter-nuclei were

reconstituted. Instances of nuclear divisions by chro-

midial fragmentation are of common occurrence among

the Protozoa and represent probably the most primitive

and direct mode of nuclear division.

It is clear, however, that if the chromatin-grains are to

be credited with specific individuality and qualitative dif-

ferences amongst themselves, this method of nuclear divi-

sion presents grave imperfections and disadvantages,

since even the quantitative partition of the chromatin 1s.

inexact, while the qualitative partition is entirely fortu-

itious. Chromidiosomes having certain specific proper-

ties might all become accumulated in one daughter-cell,

26 Op. cit., p. 101. :

. 271

272 THE AMERICAN NATURALIST [Vou. L

and those having opposite properties in the other, so that

the two daughter-cells would then differ entirely in their

properties.

I can but refer briefly here in passing to the interesting

theory put forward by Biitschli, to the effect that sexual

phenomena owe their first origin to differences between

cellular organisms resulting from the imperfections of

the primitive methods of cell-division. If we assume, for

instance, as so many have done, that one of the earliest

qualitative differences between different chromatin-gran-

ules was that while some influenced more especially the

trophic activities of the cell, others were concerned spe-

cially with kinetic functions; then it might easily happen,

after nuclear division by chromidial fragmentation, that

all, or the majority of, the kinetic elements pass into one

of the two daughter-cells, while its twin-sister obtains an

undue preponderance of trophic chromatin. As a conse-

quence, some cells would show strong kinetic but feeble

trophic energies and others the opposite condition, and in

either case the viability of the cells would be considerably

impaired, perhaps inhibited. If it be further assumed

that cells of opposite tendencies, kinetic and trophic, at-

tract one another, it is easy to see that the union and

fusion of two such cells, the one unduly kinetic (male) in

character, the other with a corresponding trophic (fe-

male) bias, would restore equilibrium and produce a

normal cell with kinetic and trophic functions equally

balanced. On this view, sexual union, at its first appear-

ance, was a natural remedy for the disadvantages arising

from imperfect methods of nuclear division.

It is not surprising, therefore, to find that the process.

of nuclear division undergoes a progressive elaboration

of mechanism which has the result of ensuring that the —

twin sister-granules of chromatin produced by division

of a single granule shall be distributed between the two

daughter-cells, so that for every chromatin-grain obtained

by one daughter-cell an exact counterpart is obtained by

the other; in other words, of ensuring an exact qualita- ;

No. 593] THE EVOLUTION OF THE CELL 273

tive, as well as quantitative, partition of the chromatin-

particles. In its perfect form this type of nuclear divi-

sion is known as karyokinesis or mitosis, and all stages

in its progressive development are to be found in the

Protozoa.

In the evolution of nuclear division by karyokinesis

two distinct processes are being developed and perfected

in a parallel manner, but more or less independently ; first,

the method of the partition and distribution of the chro-

matin-grains between the two daughter-nuclei ; secondly,

the mechanism whereby the actual division of the nucleus

and the separation of the two daughter-nuclei are effected

in the cell-division. I have dealt elsewhere?" with the

evolution of the mechanism of karyokinesis as exempli-

fied by the numerous and varied types of the process

found amongst the Protozoa, and I need not discuss the

matter further here, but the behavior of the chromatin-

grains may be dealt with briefly. The main feature in the

. process of the exact quantitative and qualitative distri-

bution of the daughter-chromatin between the daughter-

nuclei is the aggregation of the chromatin-grains or chro-

mioles into definite, highly individualized structures

known as chromosomes. In the most perfected forms of

the process of chromosome-formation the chromioles be-

come united into a linear series termed by Vejdovsky a

chr , which is supported upon a non-chromatinic

basis or axis. According to Vejdovsky, the supporting sub-

stance consists of linin; R. Hertwig, however, in his well-

known studies on Actinospherium*® considers that the

supporting and cementing substance of the chromosome

is plastin derived from the substance of the nucleoli.

However that may be, the essential feature of the chro-

mosome is the cementing together of the chromioles to

form the chromoneme, a thread of chromatin which may

be disposed in various ways on the supporting axis, some-

times being wound spirally round it (Vejdovsky).

27 Op. cit., pp. 105-120. :

28 Abhandl. bayer. Akad. (II. Cl.), XIX, 1898.

274 THE AMERICAN NATURALIST [Vou. L

The actual division of the chromatin takes place by the

longitudinal splitting of the chromoneme, in other words,

by simultaneous division into two of each of the chro-

mioles of which the thread is composed. In this way

every chromiole which was contained in the original

chromoneme is represented by a daughter-chromiole in

each of the two daughter-chromonemes. It follows that

the familiar process of the splitting of the chromosomes

in karyokinesis is a mechanism which brings about in the

most simple, sure and direct manner an exact quantita-

tive and qualitative partition of the chromatin-grains be-

tween the two daughter-nuclei. In the sequel each daugh-

ter-nucleus is built up, according to Vejdovsky, entirely

and solely from one of the two daughter-clumps of chro-

mosomes, and each chromosome is resolved again into its

constituent chromioles, giving rise in some cases to a defi-

nite portion of the nucleus, a karyomere, from which

again, at the next nuclear division, the chromosome is

reconstituted by the chromioles falling into line in an

orderly manner.

The chromatin-cycle of a cell in which the process of

division by karyokinesis takes place in its most perfectly

developed form, may, therefore, be conceived as follows:

The nucleus in its resting state contains a definite number of

companies or brigades of chromatinic units (chromioles),

each brigade spread over a certain extent of the nuclear

framework forming a karyomere. As a preparation to

division each separate brigade of chromioles falls into

line as the chromoneme, forming with its supporting sub-

stance the chromosome; there are formed, therefore, just

so many chromosomes as there were karyomeres in the

nucleus. In this disciplined and orderly array each chro-

miole undergoes its division into two daughter-chromioles,

so that each file or chromoneme of chromioles splits into

two files. At the reconstitution of the daughter-nuclei

each daughter-chromosome gives rise to a karyomere

again, the chromioles falling out of the ranks and dispos-

ing themselves in an apparently irregular manner on the

No. 593] THE EVOLUTION OF THE CELL 275

newly-built framework of the daughter-nucleus to consti-

tute their own particular karyomere. Thus karyokinesis

differs only from the most primitive method of division

by chromidial fragmentation in that what was originally

a haphazard method of distribution has become a disci-

plined and orderly manceuvre, performed with the preci-

sion of the parade-ground, but in a space far less than

that of a nutshell.

In the nuclear division of Protozoa, without going into

excessive detail, it may be stated broadly that all stages

are to be found of the gradual evolution of the tactical

problem which constitutes karyokinesis. The chromo-

somes in the more primitive types of nuclear division are

usually very numerous, small, irregular in number and

variable in size; the splitting of the chromosomes is often

irregular and not always definitely longitudinal; and dis-

tinct karyomeres have not so far been recognized in the

nuclei of Protozoa. In many cases only a part, if any, of

the chromatin falls in to form the chromosomes, and a

greater or less amount of it remains in the karyosome,

which divides directly into two. The various types of

nuclear division in Protozoa have been classified as pro-

mitosis, mesomitosis and metamitosis, for detailed ac-

counts of which those interested must refer to the text-

books and original descriptions.

I have dealt briefly with the problem of the evolution

of karyokinesis because the process of nuclear division

is, in my opinion, of enormous importance in the general

evolution of living organisms. I have expressed else-

where?® the opinion that the very existence of multicellu-

lar organisms composed of definite tissues is impossible

until the process of karyokinesis has been established and

perfected. For tissue-formation it is essential that all

the cells which build up any given tissue should be simi-

lar, practically to the point of identity, in their qualities ;

and if it is the chromatin-elements of the cell which deter-

mine its qualities and behavior, then the exact qualitative

29 Op. cit., p. 120.

276 THE AMERICAN NATURALIST [Vou. L

division of the chromatin, as effected in karyokinesis is

indispensable as a preliminary to the production of iden-

tically-similar daughter-cells by division of a parent-cell.

Hence it becomes intelligible why, amongst Metazoa, we

find the occurrence of nuclear division by karyokinesis in

its most perfect form to be the rule, and ‘‘direct’’ division

of the nucleus to be the rare exception, while, on the other

hand, in the Protista, and especially in the Protozoa, we

find every possible stage in the gradual evolution of the

exact partition of the chromatin in the process of nuclear

division, from chromidial fragmentation or the most

typical amitosis up to processes of karyokinesis as per-

fect as those of the Metazoa.

There now remains only one point of general interest

in the evolution of the cell to which brief reference must

be made, namely, the divergence of animal and vegetable

cells. Not being a botanist, I desire to approach this

question with all caution; but as a protozoologist it seems

to me clearly indicated that the typical green plant-cell

took origin amongst the Flagellata, in that some members

of this group of Protozoa acquired the peculiar chromat-

ophores which enabled them to abandon the holozoie or

animal mode of life in exchange for a vegetative mode of

nutrition by means of chlorophyll-corpuscles. It is well

known that many of these creatures combine the posses-

sion of chlorophyll with an open, functional mouth and

digestive vacuoles, and can live either in the manner of

plants or of animals indifferently or as determined by |

circumstances. It would be interesting to know exactly

what these chromatophores, at their first appearance,

represent; whether they are true cell-organs, or whether,

as some authorities have suggested, they originated as

symbiotic intruding organisms, primitively independent.

I do not feel competent to discuss this problem. I would

only remark here, that if the green plant-cell first arose

amongst the Flagellata, then the distinction between plant

and animal (that is, green plant and animal) is not so

fundamental a divergence in the series of living beings a :

No. 593] THE EVOLUTION OF THE CELL Zit

as is popularly supposed, but is one which did not come

into being until the evolution of organisms had reached

a relatively advanced stage, that, namely, of the true

nucleated cell.

I have confined myself in this address to the evolution

of the cell as this organism is seen in its typical form in

the bodies of the multicellular organisms, starting from

‘the simplest conceivable type of living being, so far as

present knowledge enables us to conceive it. But there

is not the slightest reason to suppose that the evolution

of the Protista took place only in the direction of the

typical cell of the cytologist. Besides the main current

leading up to the typical cell there were certainly other

currents tending in other directions and leading to types

of structure very unlike the cells composing the bodies of

multicellular organisms. It is impossible that I should

do more here than indicate some of the divergent lines of

evolution, and I will confine myself to those seen in the

Protozoa.

Taking as the starting-point and simplest condition in

the Protozoa a simple cell or protocyte, in which the body

consists of a small mass of cytoplasm containing a nu-

cleus, with or without chromidia in addition, an early

specialization of this must have been what I may term

the plasmodial condition, typical of Rhizopods in which

the cytoplasm increased enormously to form relatively

large masses. The nucleus meanwhile either remains

single and grows very large or, more usually, a great

number of nuclei of moderate size are formed. From

this large plasmodial type is to be derived the forami-

niferal type, characterized by the creeping habit of life,

and probably also the radiolarian type, specialized for

the floating pelagic habit. Both foraminiferal and radio-

larian types are characterized by an excessive develop-

ment and elaboration of skeletal structures, and the Seer

logical record proves that these two types of organisms

attained to a high degree of specialization and diversity

278 THE AMERICAN NATURALIST [Vou. L

of form and structure at a very early period.* The

Mycetozoa exemplify another development of the creep-

ing plasmodial type adapted to a semi-terrestrial mode of

life.

In the Mastigophora the body generally remains small,

while developing organs of locomotion and food-capture

in the form of the characteristic flagella. In this class

there is a strong tendency to colony-formation brought

about by incomplete separation of sister-individuals pro-

duced in the ordinary process of reproduction by binary

fission. The so-called colonies (they would better be

termed families) show a most significant tendency to

individualization, often accompanied by physiological and

morphological specialization of the component flagellate

individuals.

As an offshoot, probably, from ancestors of the Masti-

gophoran type arose the Infusoria, the Ciliata and their

allies, representing by far the most highly organized uni-

cellular type of living being. No cell in the bodies of the

Metazoa attains to such a complication of structure as

that exhibited by many Ciliates. In the Metazoa the in-

dividual cells may be highly specialized for some particu-

lar function of life; but a Ciliate is a complete and inde-

pendent organism and is specialized for each and all of

the vital functions performed by the Metazoan body as a

whole. From the physiological standpoint a Ciliate (or

any other Protist) is equivalent and analogous to a com-

plete Metazoon, say a man, but I can not for a moment

agree with Dobell** that the body of a Ciliate is homol-

ogous with that of a Metazoon—not at least if the word

homologous be used in its usual biological sense of homo-

genetic as opposed to homoplastic. Dobell appears to me

to negative his own conclusion when he maintains that

the body of a Ciliate is ‘‘non-cellular’’? while admitting

that the Metazoon is multicellular; how then can they be

said to be homologous? Only if the term homologous be

80 For Foraminifera see especially Heron-Allen, Phil. Trans. (B), Vol.

206 (1915), p. 229. i

31 Journal of Genetics, IV (1914), p. 136.

No. 593] THE EVOLUTION OF THE CELL 279

used in a sense quite different from its ordinary sig-

nificance.*?

In addition to the highly developed structural differen-

tiation of the body the Infusoria exhibit the extreme of

specialization of the nuclear apparatus in-that they pos-

sess, as a rule, two distinct kinds of nuclei, micronuclei

and macronuclei, composed respectively of generative

and trophic chromatin, as already pointed out. This fea-

ture is, however, but the culminating point in a process

of functional specialization of the chromatin which can be

observed in many Protozoa of other classes, and which,

moreover, is not found invariably in its complete form in

all Ciliata.

In this address I have set forth my conceptions of the

nature of the simplest forms of life and of the course taken

by the earliest stages of evolution, striving all through to

treat the problem from a strictly objective standpoint, and

avoiding as far as possible the purely speculative and

metaphysical questions which beset like pitfalls the path

of those who attack the problem of life and vitalism. I

have, therefore, refrained as far as possible from discuss-

ing such indefinable abstractions as ‘‘living substance’’ or

‘‘life,’’ phrases to which no clear meaning can be attached.

How far my personal ideas may correspond to objective

truth I could not, of course, pretend to judge. It may be

that the mental pictures which I have attempted to draw

are to be assigned, on the most charitable interpretation,

to the realm of poetry, as defined by the greatest of poets,

rather than of science.

The lunatic, the lover and the poet

Are of imagination all eompact;

Ana as imagination bodies forth

The forms of things unknown, the poet’s pen

Turns them to shapes and gives to airy nothings

A local habitation and a name.

If I might be permitted to attempt an impartial criti-

cism of my own scheme, I think it might be claimed that

32 See Appendix A.

280 THE AMERICAN NATURALIST [Von E

the various forms and types of organisms in my evolu-

tionary series, namely, the simple cell or protocyte, the

cytode or pseudomoneral stage, the micrococcus, even the

biococcus, are founded on concrete evidence and can be re-

garded as types actually existent in the present or past.

On the other hand the rôle assigned by me to each type in

the pageant of evolution is naturally open to dispute. For

example, I agree with those who derive the Bacteria as

primitive, truly non-cellular organisms, directly from the

biococcus through an ancestral form, and not at all with

those who would regard the Bacteria as degenerate or

highly-specialized cells. But the crux of my scheme is the

homology postulated between the biococeus and the chro-

matinic particle—chromidiosome or chromiole—of true

cells. In support of this view, of which I am not the orig-

inator, I have set forth the reasons which have convinced

me that the extraordinary powers and activities exhibited

by the chromatin in ordinary cells are such as can only be

explained on the hypothesis that the ultimate chromatinic

units are to be regarded as independent living beings, as

much so as the cells composing the bodies of multicellular

organisms ; and, so far as I am concerned, I must leave the

matter to the judgment of my fellow-biologists.

I may point out in conclusion that general discussions

of this kind may be useful in other ways than as attempts

to discover truth or as a striving towards a verity which

is indefinable and perhaps unattainable. Even if my

scheme of evolution be but a midsummer-night’s fantasy,

I claim for it that it coordinates a number of isolated and

scattered phenomena into an orderly, and, I think, intel-

ligible sequence, and exhibits them in a relationship which

at least enables the mind to obtain a perspective and com-

prehensive view of them. Rival theories will be more, or

less, useful than mine, according as they succeed in corre-

lating more, or fewer, of the accumulated data of experi-

ence. If in this address I succeed in arousing interest

and reflection, and in stimulating inquiry and controversy,

it will have fulfilled its purpose.

No. 593] THE EVOLUTION OF THE CELL -> 281

APPENDIX A.—THeE CELL-THEORY

The most recent attack on the Cell-theory, as it is under-

stood by the majority of modern biologists, has been made

by Mr. Dobell, who, if I understand him rightly, refuses to

admit any homology between the individual Protistan or-

ganism and a single cell of the many that build up the

body of a Metazoon. On the contrary, he insists that the

Protist is to be regarded as homologous with the Metazoan

individual as a whole. On these grounds he objects to

Protista being termed ‘‘unicellular’’ and insists that the

term ‘‘non-cellular’’ should be applied to them.

As regards the cellular nature of the Protista, it is one

of my aims in this address to show that amongst the

Protista all stages of the evolution of the cell are to be

found, from primitive forms in which the body can not be

termed a cell without depriving the term ‘‘cell’’ of all de-

finable meaning, up to forms of complex structure in

which all the characteristic features of a true cell are fully

developed. Thus in the Protozoa we find the protoplas-

mic body differentiated into nucleus and cytoplasm; the

nucleus in many cases with a structure comparable in

every detail to that of the nucleus of an ordinary body-

cell in the Metazoa; reproduction taking place by division

of the body after a karyokinetic nuclear division often

quite as complicated as that seen in the cells of the Metazoa

and entirely similar both in method and in detail; and

in the sexual process of differentiation of the gametes on

lines precisely similar to those universal in Metazoa, often

just as pronounced, and preceded also in a great many

cases by phenomena of chromatin-reduction comparable

in principle, and even sometimes in detail, with the reduc-

tion-process occurring in Metazoa. I really feel at a loss

to conceive what further criteria of homology between a

Protozoon and a Metazoan cell could be demanded by even

the most captious critice. On the ground of these and

many other similarities in structure and behavior between

the entire organism in the Protozoa and the individual cell,

whether tissue-cell or germ-cell, in the Metazoa, the case

282 THE AMERICAN NATURALIST [Vou. L

seems to me overwhelmingly convincing for regarding

them as truly—that is to say, genetically—homologous.

Looking at the matter from another point of view,

namely, from the standpoint of the Metazoa, it is true that

in the groups of most complicated and highly organized

structure the cells often develop secondary connections or

fusions due to incomplete division, to such an extent that

in parts of the body the individuality of the primitively

distinct cells may be indicated only by the nuclei (as may

occur also in Protozoa, for example, in associated grega-

rines); but in all Metazoa certain of the cells retain per-

manently their complete independence and freedom of

movement and action. Inthe Metazoa possessing the sim-

plest and most primitive types or organization, such as

sponges and celenterates, the cells composing the body

show far greater independence of action, and in the course

of ontogeny entire groups of cells may alter their relative

positions in the body as the result of migrations performed

by individual cells; while it is now well known that if the

adult sponge or hydroid be broken up completely into its

constituent cells, those cells can come together again and

build up, by their own individual activity, the regenerated

body of the organism. For these reasons it seems to me

impossible to regard the body-cells of the Metazoa other- -

wise than as individual organisms complete in themselves,

primitively as independent as the individual Protozoon,

and in every way comparable to it.

From the considerations summarized very briefly in the

two foregoing paragraphs and capable of much greater

amplification and elaboration, the view generally held that

the entire organism of a Protozoon is truly homologous

with a single body-cell of a Metazoon seems to me quite

unassailable, and to have gained in force greatly from re-

cent investigations upon both Protozoa and Metazoa. On

the other hand, any Protist, as an organism physiologic-

ally complete in itself, is clearly analogous to the entire

individual in the Metazoa—a comparison, however, which

leaves the question of genetic homology quite untouched.

No. 593] THE EVOLUTION OF THE CELL 283

As regards the application of the term unicellular or

non-cellular to the Protozoa, it is evident that if the evo-

lution of living beings had never proceeded beyond the

stage of the Protista, and if no multicellular organisms

had ever been evolved, the term cell could then never

have been invented by an intelligent being studying other

living beings, supposing for an instant the possibility of

such intelligence existing apart from a mammalian brain.

So long as the Protozoa are studied entirely by themselves,

without reference to any other forms of life, they may be

termed non-cellular in the sense that they are not com-

posed of cells. It is only when they are compared with

multicellular organisms that the term unicellular becomes

applicable on the ground of the homology already dis-

cussed between the Protozoon and the body-cell of the

Metazoon.

THE MECHANISM OF CROSSING-OVER

HERMANN J. MULLER

Rice INSTITUTE

IV. Tue Manner OF OCCURRENCE OF CROSSING-OVER

A. Interference

As soon as it seemed probable that the factors were

linked in line, and that the crossing-over was the actual

method of interchange, it became of interest to discover

and to analyze the precise mode of incidence of the inter-

change. The questions suggested themselves, for ex-

ample, what was the total frequency of crossing-over, did

any factors separate more often than they remained to-

gether, how often did crossing-over occur at two points

simultaneously, and was there any tendency, in such

cases, for the two points of crossing-over to be a definite

distance apart, or in definite positions, ete. For answers

to these questions might throw light on the mechanism of

crossing-over, what cytological phenomena it was con-

nected with, and what stage in synapsis it occurred at.

= With these points in view the author calculated the

linkage relations that would result on several possible

schemes of interchange. The simplest possibility was

that the chromosomes always twisted in loops of fixed

length, though not of fixed position, and always under-

went breakage, with recombination of homologous

strands (i. e., ‘‘ecrossed-over’’ in the technical sense), at

each place that the strands crossed one another. In such

a case there would always be a definite distance between

one point of crossing-over and another; moreover, all

factors which were separated by a distance great enough

for double crossing-over to occur between them, i. e., by

the length of at least one loop, must always have either

double or single (or multiple) crossing-over between

them. Sturtevant’s data, however, showed that this was

284

No.593] THE MECHANISM OF CROSSING-OVER 285

not true, and accordingly it had to be concluded that the

length of the loop was variable, or that ‘‘crossing-over”’

did not always occur where the strands crossed.

Another possibility was that crossings-over were quite

independent of one another, having an entirely random

or chance distribution in the chromosome, with reference

to each other. This would mean that when crossing-over

occurred at one point, another crossing-over would be

just as likely to oceur coincidently at any other given

point—whether this be very near or far away—as when

no crossing-over took place at the first point. But this

latter scheme would not be that expected on the method of

crossing-over proposed by Jannsens and followed by

Morgan, for in the stages when Jannsens supposed cross-

ing-over to occur the chromosomes are rather loosely

twisted, so that loops of very small length do not occur

as often as longer ones (thus, very near one point of

crossing-over the strands seldom cross back again). I

therefore determined the mathematical relations which

would exist between crossing-over frequencies, if cross-

ings-over had a chance distribution with reference to one

another, in order to compare these figures with those

obtained by experiment. On the assumption that sepa-

ration between A and B has no influence on separation

between B and C, if crossing-over occurs between A and

B in 10 per cent. of cases and between B and C in 20 per

cent. of cases, then, among those ten cases in a hundred

where crossing-over between A and B oceurs, 20 per cent.

(i. e., 2 eases) would be cross-overs between B and C as

well; in other words, the per cent. of double cross-overs

would be equal to the product of per cents. AB and BC

(formula 1). The easiest way to determine the correct-

ness of the assumption in any given case, therefore, is to

compare the observed per cent. of double cross-overs with

per cent. AB X per cent. BC.

Another relation besides was found to hold between

the theoretical linkage values, dependent upon the rela-

tion in formula 1. For it is easily seen that the number

of separations between A and C must always be equal

286 THE AMERICAN NATURALIST [Vou. L

to the sum of the number of crossings-over between A.

and B, and between B and C, minus all those crossings-

over contained in the cases where coincidence occurred,

and in which A and C, therefore, failed to separate,—1. e.,

minus twice the number of cases of double crossing-over.

Hence, if formula 1 is correct, then it must also be true

that per cent. AC — per cent. AB + per cent. BC—2

(per cent. AB X per cent. BC) (formula 2). This for-

mula was originally expressed not only in the above

terms, where the ‘‘per cent. of separations’’ (i. e., ratio

of separations to the total number of cases) is used as the

index of separation frequency, but also in terms of the

so-called ‘‘gametic ratio’’—the ratio of cases of non-

separation to those of separation—for this was the way

of indicating degree of linkage then used by all investi-

gators of the subject. The latter index gives much more

complicated formulas, however, and so it was pointed out

at the same time that per cent. of separations would

afford a much more useful measure of linkage.

Later, Trow also worked out and published the same

formula (no. 2)—in terms of the ‘‘ gametic ratio’’—and it

is generally known as ‘‘Trow’s special hypothesis’’ (17).

But on the reduplication hypothesis held by Trow, and

by the other English geneticists who do not accept the

chromosome explanation, the formula would be supposed

to result, not from the fact that crossing-over between A

and B was independent of that between B and C, but

from the fact that ‘‘reduplications’’ AB and BC were

independent, not being disturbed by any ‘‘ primary redu-

plication’? AC. Adherents of the reduplication hypothe-

sis have been much concerned as to whether or not their

results confirmed the assumptions made in Trow’s for-

mula, and have in one or two instances calculated that

they did. Let us examine for a moment the requisites

for proving such a conclusion. As above shown, the

whole matter turns on the frequency of coincidence of

separations AB and BC (i. e. on the frequency of

‘double crossing-over’’) and the question can be settled

by determining directly the amount of this coincidence.

No.593] THE MECHANISM OF CROSSING-OVER 287

If the per cent. of double cross-overs = per cent. AB

x per cent. BC (formula 1), then the assumption that

separation frequencies AB and BC are independent is

correct. As offspring from a back-cross all show what

factors they received from the hybrid parent, a back-

cross involving the three factors A, B, and C at the same

time will answer the question at once, for all the cases

of coincident separation (double cross-overs) that occur

can be counted. But where the hybrids, instead of being

back-erossed, are inbred—a practice followed by adher-

ents of the reduplication hypothesis—then it is impos-

sible to tell which F, individuals come from gametes of

the classes which we may term double cross-overs, unless

one of these classes is the triple recessive, and then the

only double crossovers which can be known as such are

those very rare individuals that happen to result from

the union of two double crossover gametes. The British

workers have, therefore, not been able to find the pro-

portion of double cross-overs directly, to compare this

with formula 1, but have tried to determine the frequency

of coincidence indirectly, by using the method followed

in formula 2. That is, they determined the relations

existing between frequencies AC, AB, and BC, as calcu-

lated from their F, counts, for, as above shown, the

greater the frequency of double crossing-over, the more

will AC be cut down in proportion to AB and BC. And it

seemed evident that, if the relation of AC to AB and BC

was just that given by Trow’s formula (2), then coinci-

dence of separations must have the frequency demanded

on the assumption that separations (or * reduplications’’)

AB and BC occur independently of one another. As a

matter of fact, however, this method offers no answer to

the question, unless almost impossibly large F, counts

are obtained, for otherwise the independent random fluc-

tuations of these three values in this kind of count are so

great that any deviation in AC due to excess or deficiency

of double crossing-over would be quite lost to sight. :

The question was, however, immediately and defini-

tively answered in Drosophila, before Trow’s paper ap-

288 THE AMERICAN NATURALIST [Vou.L

peared, by examination of Sturtevant’s extensive back-

crosses, especially of those involving three pairs of fac-

tors at once. As the results did not conform to the for-

mula, it was not published, but as Trow has since raised

this question publicly and the adherents of the reduplica-

tion hypothesis are still discussing it, it may not be out of

place to have given an analysis of it here, and to recall the

fact that it had already been tried and rejected. Besides,

as will appear below, a discussion of the relations which

would exist if crossings-over were independent of one

another is a necessary preliminary for a treatment of the

relations which do exist between linkage values.

The results showed that double crossing-over does not,

as a rule, occur as frequently as would be expected if,

as the above formule assumed, it were purely a matter

of chance whether or not two cross-overs happen coinci-

dently. In a sense, then, the occurrence of one crossing-

over interferes with the coincident occurrence of another

crossing-over in the same pair of chromosomes, and I

have accordingly termed this phenomenon ‘‘interfer-

ence.’ The amount of interference is determined by

comparing the actual per cent. of double cross-overs with

the per cent. expected if crossings-over were independ-

ent, i. e., if they had a purely chance distribution with

reference to each other. Now, the per cent. which would

occur on the latter expectation has already been given by

formula 1 as per cent. AB X per cent. BC. If, then, the

observed per cent. of double cross-overs were divided by

per cent. AB X per cent. BC, we would obtain a fraction

showing what proportion of the coincidences which would

have happened on pure chance really took place. This

ratio of observed double cross-overs to the chance expec-

tation appears to me to furnish the most useful measure

of interference. The ratio is itself best expressed in per

cent., and it may be called the relative coincidence, or

simply ‘‘coincidenee.’’ If the ‘‘eoincidence’’ is low, this

means that there has been much interference, for most of

the double cross-overs expected on chance were prevented

from appearing; conversely, if coincidence is high, the

No.593] THE MECHANISM OF CROSSING-OVER 289

interference must have been very weak. Some illustra-

tions may make the meaning of this index clearer. If,

for example, coincidence is 0 per cent. no double crossing-

over is occurring; the interference between one crossing-

over and another is then complete. If coincidence is 45

per cent., this figure does not mean that 45 per cent. of

the individuals are double crossovers, but that 45 per

cent. of the number of double crossovers which would be

expected as a result of pure chance (whatever that num-

ber may have been) actually appeared, 55 per cent. hav-

ing been ‘‘interfered with,’’ or somehow prevented from

occurring. If coincidence is 100 per cent., there has been

no interference, for the same number of double cross-

overs appeared as expected on the ground that the two

crossings-over did not interfere with each other’s occur-

rence. 110 per cent. would mean that if one crossing-

over occurred, the other was 10 per cent. more likely to

occur than in cases of random distributions of crossings-

over. This would be ‘‘negative interference,’’ for as

coincidence increases interference decreases.

On Janssens’s theory that crossing-over takes place in

the strepsinema stage, when the chromosomes are twisted

in loose loops, crossing-over would very seldom take

place at two points very near together, for this would re-

quire a tight twisting of the chromosomes. Accordingly,

on this theory interference was to be expected; further-

more it would be expected that interference was very great

between crossings-over that were in neighboring regions;

but between crossings-over further apart there should be

little or no interference. The results were according to

this expectation; they indicated strongly that the inter-

ference was very great for crossings-over short distances

apart, but progressively diminished as the distances con-

sidered became greater. The conclusion drawn was that

crossing-over took place as postulated on Janssens’s

theory, when the strands were loosely twisted in strep-

sinema, although the twisting and crossing-over did not

take place in the stereotyped manner suggested as a first

Possibility, in the earlier part of this section. For there

290 THE AMERICAN NATURALIST [Vou. L

was evidence that the distances between the two points of

crossing-over in double cross-overs were variable; but

this again corresponded with the fact that the chromo-

somes of Batracoseps and other forms, as seen under the

microscope, did not always twist.in loops of the same

length. Furthermore, if it be supposed that in most

maturing eggs of the fly the homologous chromosomes

twist tightly enough to cross at least once or twice, as is

certainly the case in Batracoseps and many other forms,

it must be concluded that at not every point of crossing

does actual ‘‘crossing-over’’ (recombination of strands)

take place, for it was found that nearly half of the factor-

groups emerged without having undergone any crossing-

over at all. And this, in turn, corresponded with the

observations of Janssens and others, which showed that ~

at some at least of the points of crossing of homologous

chromosomes, the latter merely untwisted again without

having undergone the ‘‘chiasmatype’’ process. Here,

then, was a theory of crossing-over that seemed com-

plete, so far as connecting the genetic facts with the cyto-

logical observations was concerned.

B. Possible Mechanisms of Crossing-Over

There is one very unsatisfying point, however, in this

original scheme of crossing-over. That is, it postulates

that crossing-over occurs at a comparatively late stage in

synapsis, when the strands have become very much shorter

and thicker than the long delicate threads which first

came into contact with their homologues (see Fig. 6).

Now, in crossing-over the chromosomes must come into

contact, and break, at precisely homologous points, other- |

wise factors would be lost or gained by them when cross-

ing-over occurs. But presumably the factors are set very

close together in the line, judging by the fact that muta-

tions in new ‘‘loci’’ (positions in the chromosomes) are

still as numerous as ever, and that, if the whole chromo-

some is packed with factors as close together as, judging

by their linkage relations, they seem to be at certain places

in it, it must contain at the very least 200 factors. It is

No.593] THE MECHANISM OF CROSSING-OVER 291

difficult to conceive how this cleavage of ultramicroscopic

nicety can take place properly at a stage when the chro-

mosomes are so coarse and short. The observations of

Vejdovsky and others, taken in connection with the ge-

netic results from Drosophila, render it practically certain

that the factors are really disposed in an extremely fine,

Bs b

Fic. 6. Chromosomes during an early stage of synapsis (amphitene). In

some preparations the apposed threads seem parallel, as in a; in others they

seem twisted about each other, as in

long thread or ‘‘chromonema,’’ which, during the meta-

phase and anaphase of mitosis, is coiled up very closely

in more or less spiral fashion (probably within a viscous

sheath of some sort), to form the thick dense chromo-

somes, but which, in the resting period and during the

early stages of synapsis, becomes, to some extent at least,

uncoiled and drawn out again. In this state, then, the

chromosomes first pair, as shown in Fig. 6., Thus pre-

cisely homologous parts of the frail threads nmay become

apposed to each other, so that this stage, which is called

the ‘‘amphitene’’ stage, would seem to be the one best

‘‘adapted’’ for the occurrence of crossing-over. Later,

when each chromosome becomes, presumably, a thick

spiral, there would seem to be much greater mechanical

difficulties in the way of exact apposition and breakage of

parts.

On any possible theory of crossing-over, however, the

known facts concerning interference should be capable of

292 THE AMERICAN NATURALIST [ Vou. L

interpretation. If crossing-over occurred during the

‘‘amphitene’’ stage, or not long after, would there be any

possible explanation of the fact that one point of crossing-

over is generally far removed from another? The ex-

planation might be found simply in the fact that each of

the ‘‘leptotene’’ chromosomes—i. e., the finely drawn out

chromosomes which are just about to undergo synapsis—

pursued a general course that had few close turns in it.

(For possibly it maintains the same general direction as

it had when it was short and thick; the reader will recall

that Boveri found that chromosomes preserve their ap-

proximate shape and position from one cell division to the

next.) When, therefore, the leptotene chromosomes are

being brought together by the synaptic attraction which

homologous loci then bear for each other, the threads are

usually crossed only at a few points, and these are gen-

erally far apart. If these initial points of crossing—

which, it will be observed, have been determined by the

original positions of the threads, and not by any twisting

—are the points of crossing-over, interference would be

accounted for, and would, in effect, be of the same general

nature as on the mechanism of crossing-over postulated

by Janssens.

It might at first seem hard to imagine why, on this

second scheme of crossing-over, recombination—?. e.,

‘‘crossing-over’’—should occur where the threads cross,

but it should be remembered that the two threads, while

coming together, often lie in about the same plane both

above and below the point of crossing. If they keep to this

original plane as they draw together, they will come to

have the same plane of apposition just above and just

below the crossing point,—although the sides of the fila-

ments that face each other will be just the opposite in the

two cases; consequently, the threads at the crossing point

must undergo a very sharp twist, and if, as we must sup-

pose, they are somewhat viscous, this may result in their

breakage and recombination, or, perhaps, first in their

fusion, and, later, when the pieces of the same chromo-

some above and below the point of crossing are wrenched

No.593] THE MECHANISM OF CROSSING-OVER 293

apart in opposite directions by mutual repulsion of the

strands or by pulling of spindle fibres, in breakage of

parts originally together. (So perhaps fusion might

occur during the amphitene and breakage in the strepsi-

nema stage; this would be a combination of schemes 1

and 2 which would account both for the exact apposition

of parts and for the phenomena observed by Jannsens.)

Be this as it may, at any rate, the negative argument may

be given that it is just as hard to account for recombina-

tion at a later stage in synapsis as at this stage, even

overlooking the objection of the thickness of the threads.

There is a serious objection to the scheme just given,

however, in that, as the threads come together, they seem,

in many preparations, not to keep their original plane of

apposition, but to twist tightly about each other, like the

strands of a rope, throughout their entire length (see

Fig. 6). It is possible that the twisting of one thread

about the other is merely apparent, however, and that the

threads lie parallel but are simply coiling up in a spiral,

in the process of forming the shorter, thicker prophase

chromosomes; for, unless the spiral were very delicately

preserved by the fixing agent, there would be apparent

knots in it as though there were a twisting of two strands

about each other. Moreover, there is evidence indicating

that this tight twisting occurs only in certain species of

animals. But let us assume for the moment that this very

tight twisting really takes place during the amphitene

stage in flies, and that crossing-over takes place at this

period (this we may call scheme of crossing-over number

three). Would there then be any way of explaining why

one crossing-over should interfere with another near by,

in view of the fact that the loops are of such small di-

mensions? In seeking an answer to this question, it will

be helpful to bear in mind that crossing-over can be di-

vided into just three essential processes—a bending of the

chromosomes across each other, a breaking of the threads,

and then a fusion of adjoining pieces (or, perhaps, the

fusion of the homologous chromosomes comes first, and

then the breaking of the original chromosomes at that

294 THE AMERICAN NATURALIST [Vou. L

point). It follows from this that interference must in

any case be due to one of the following three general

causes: (1) Hither the chromosomes are not likely to bend

across each other twice at points near together (i. e., the

_loop tends to be long), or (2) breakage at one point for

some reason interferes with another breakage nearby

(even though the threads are crossed at both of these

points), or (3) fusion of chromosomes at one point in

some way interferes with fusion of threads which are

crossed in a neighboring region. That fusion at one point

could interfere with fusion at another point can scarcely

be imagined. And if crossing-over occurs according to

scheme number three, the ‘‘loop explanation’’ must also

be thrown out. Consequently, if crossing-over occurs at

a stage of tight twisting the breakage of the threads at one

point must somehow be considered to prevent another

break near by. In explanation of this, breakage might be

thought of as resulting from the tightness of the twisting,

for then a breakage of the threads at one point would re-

lieve the tension of the filaments for some distance along

the line and so tend to prevent another breakage from

occurring near by. (Later, when threads reunited at the

point of breakage, pieces from homologous chromosomes

would be as apt, or more apt, to lie end to end, and there-

fore to join, than pieces of the same chromosome. As a

partial explanation of why the fragments should join again

at all, it might be supposed that only the chromonemas

break, the fused sheath which envelops the pair still hold-

ing the pieces together.)

It is fully realized that the above discussion is highly

speculative. It is intended, however, not as a presenta-

tion of conclusions, but as a tentative suggestion of pos-

sibilities, in order to obtain some system of ideas that may

furnish a temporary basis for a real attack—experimental

and observational—upon the subject.

Tests for These Alternatives

Is there any way of obtaining evidence as to which of

these three schemes of crossing-over is the more probable

No.593] THE MECHANISM OF CROSSING-OVER 295

one? Light might perhaps be thrown on the question by

a closer study of interference, and it was largely for this

reason that the experiment described in section V was

undertaken. If, for example, interference was a result

of length of loop (as would be true in schemes I and II),

and the length of the loop tended to vary more or less in

both directions, about a given mode, then coincidence

_ would be relatively higher between crossings-over which

were that distance apart, than between crossings-over

nearer together or still further apart. In other words, as

may be seen from Fig. 7, for small distances, the relative

Fic. 7. Diagram to illustrate the second scheme suggested for crossing-over.

The amphitene threads become sharply crossed at a particular point.

coincidence would be very small (interference high), for

longer distances much greater, and with still longer dis-

tances coincidence would fall again (interference would

rise). For distances double or triple the length of the

loop—if the chromosomes were as long as that—coinci-

dence would rise once more. Secondly, on the ‘‘loop ex-

planation’’ of interference just outlined, coincidence

should, at the modal distance, rise above the 100 per cent.

level, for crossing-over would occur at a given point (K)

more often in those cases when there is crossing-over at

another point (I) lying at the modal distance from K, than

in the average case. Of course it might be, however, that

there was no modal length of loop—that although short

loops were infrequent, all loops above a certain size were

equally frequent, or that the longer the loop, the more

296 THE AMERICAN NATURALIST [ Vou. L

frequent it tended to be. In the former case coincidence

would rise to a certain level, as distance between the points

of crossing-over considered increased, and would after

that remain constant; in the latter case it would rise pro-

gressively, and might or might not reach or pass the 100

per cent. level.

On the other hand, if crossing-over is due to a breakage

of tightly twisted threads, not so many different kinds of

variation of coincidence, with increase in distance, would

be theoretically possible, but a condition something like

the one last mentioned must always obtain. For, on

scheme 3, the interference of a breakage with the tightness

of twisting and consequent chance for another breakage

must decrease progressively at greater and greater dis-

tances from that breakage; coincidence would thus rise

until finally it reached the 100 per cent. level expected on

chance. It would never rise much? beyond this, as one

break could never make another more likely to occur;

neither could coincidence fall once more, with a still

greater distance (as it could on the loop scheme, after a

‘‘modal distance’’ had been reached). If, therefore, it

should be found that, for certain (modal) distances be-

tween two points of crossing-over, coincidence ran well

above 100 per cent., or that, beyond certain distances, co-

incidence fell again, there would be good evidence that

crossing-over did not occur at a stage of tight twisting.

If, on the contrary, it were found that crossing-over coin-

cidence rose progressively with distance, until it reached

the 100 per cent. mark, but neither went much? beyond

2 Even on scheme III, coincidence could finally rise slightly above 100

per cent., for although one break (I) could not help another (K) to occur,

no matter how far away the latter (K) might be, still it might, by pre-

venting the occurrence of other breaks (J), in between these two, give more

chance for the occurrence of the break farther off (K), since in this way

the interference of breaks J with K (which is stronger than the interference

of the more distant I with K) is removed. Thus break K might occur more

often when I also occurs than in the average case, and so coincidence would

rise above 100 per cent. However, it would be easy to distinguish between

the slight rise in coincidence above 100 per cent., due to this cause, and the

rise which would exist on the loop explanation of interferences if I and K were

separated by a distance about equal to the modal length. For, in the first case,

considering only gametes in which no crossing-over at all took place in between

No.593] THE MECHANISM OF CROSSING-OVER 297

this, nor fell again later, and if cytological measurements

should then substantiate the judgment, based on inspec-

tion, that the loops did have a modal length during the

strepsinema stage, there would be good evidence that

erossing-over must occur at an early stage of synapsis.

Other peculiarities of coincidence also might be found

which would permit of explanation on one scheme and

not on another. In groups II and III, for example, there

seem to be peculiarities in the coincidence relations in

cases where the chromosomes differ in regard to the fac-

tor C, or a similar factor. And a comparison of coinci-

dence in different regions of the chromosome in any given

case or in the same region of the chromosome in cases of

linkage variation, might very well reveal relations that

lend evidence to one scheme of crossing-over or another.

Even a determination, not of coincidence, but merely of

linkage variation itself, in different parts of the chromo-

some, might in some way shed light on the subject. In

the case of the third chromosome, experiments of this sort

are now under way with multiple stocks which I have

made up for this purpose, and Sturtevant is conducting

similar experiments with group II. The first require-

ment, however, is obviously an accurate study of normal

coincidence, and it therefore became necessary to deter-

mine the coincidence for points various distances apart,

preferably in the same experiment. But to work with a

great many factors in a group at once introduced new diffi-

culties, which made special methods necessary, as will be

explained later. Before considering this experiment, it

will be desirable to consider other lines of evidence and

modes of attacking the problem of crossing-over.

The cytological evidence which Janssens presents for

crossing-over is entirely directed towards proving that

crossing-over occurs during strepsinema or later. In

strepsinema the chromosomes, as already mentioned,

(at J), it is easily seen that the proportion of breaks at K would be lower

when breakage occured at I than when there was no breakage at I, whereas

in the second case, the proportion of breaks at K would in such gametes be

higher when there was breakage at I than when there was no breakage at I.

298 THE AMERICAN NATURALIST [Vou. L

become much shorter and thicker than in the amphitene

stage, and each chromosome in the pair can in many prep-

arations be seen to have split lengthwise, i. e., the ‘‘tet-

rads” have formed preparatory to the two maturation

divisions. Janssens often finds the four threads placed

somewhat as shown in Fig. 8a, two of the threads crossing

at one or two points, but otherwise being rather widely

separated, and the other two threads rarely crossing but

lying close to whichever one of the two threads first men-

tioned happens to be on the same side, and merely bend-

ing inwards and then back again where the first two

threads cross. The peculiar crossing of two of the

threads and the bend in the other two, as shown at point

L, he interprets, in the way shown in Fig. 8b, as meaning

k p ------ ----------

E Q

Fig. 8. Diagram to show a coincidence relations on schemes I and

esented

would have to twist in loops much smaller — the modal length. But it will

often coincide with one at C, seldom with one at D, and often again with one

at E. The relative coincidence of hetnet -over at various points on this

chromosome with crossing-over at A is shown in the curve on the right.

that both pairs of threads originally were twisted across

each other, but that the two homologous threads which

were originally on the inner side, and so touched each

other, underwent recombination, i. e., ‘‘erossed over,’’ at

the point of contact; each of the new chromosomes thus

formed, therefore, would lie entirely on one side or the

No.593] THE MECHANISM OF CROSSING-OVER 299

other; the other two threads, on the contrary, are sup-

posed not to have undergone recombination (‘‘crossing-

over’’) and therefore would still lie across each other.

It would seem equally possible, however, to interpret

these figures as meaning that (as shown in Fig. 9c and

9d) when the four threads began to separate into two

pairs, separation happened to start at some points (A and

C) between the identical halves and at other points (B)

between the homologous chromosomes, it being merely a

matter of chance in which way the separation started to

take place. It will be seen that this would result in the

formation of just such cross-figures, between two regions

where separation took place in opposite ways, as Janssens

finds.

L a b

Fa Va O pe

A ye SOE

Fie, 9. p The chiasmatype described by Janssens. (b) His epi

f it. (c) and (d) A suggested alternative interpretation

Another point in Janssens’s evidence is that the pro-

phase chromosomes of maturation divisions not only show

the strands crossing, at points, but often bending in

towards each other near the middle, as though they had

formerly crossed there, and later undergone crossing-over.

It would seem possible, however, that this figure is merely

due to the chromosomes remaining in contact more closely

at the point where the spindle fiber is attached, and spread-

ing apart elsewhere,—a relation which figures of Bridges

and others show to exist between the two identical halves

of chromosomes in the prophases of oogonial mitoses.

300 THE AMERICAN NATURALIST [Vou. L

Finally, Janssens says that crossing-over is indicated by

the fact that the chromosomes often seem to have sunken

into one another at the crossing point. This detail, which

would be very difficult to establish, might, of course (if

it had any significance at all), merely mean that the chro-

mosomes were still closely attached at the point where

they had previously crossed-over. It seems incautious,

therefore, to regard the cytological evidence as showing

more than the possible means of the crossing-over which

the evidence from factor and chromosome distribution

demonstrates to occur.

Some of the Orthopteran material which gives such

clear-cut chromosome figures might perhaps settle the

point whether crossing-over occurs at the stage of four

threads, as Janssens believed. For it is reported by

Wenrich (19) that homologous chromosomes can some-

times be distinguished from one another in prophases by .

differences in the size or shape of contained granules, that

are constant for the particular individual. Moreover, the

four threads are clearly distinguishable in the prophase

of the first maturation division. If a female could be

found (there is some reason to believe that crossing-over

does not occur in the male) which showed a difference in

respect to two granules at different points in the same pair

of chromosomes, then, if Janssens’s theory is right, it

would happen that, in some of the oocytes, of the four

post-synaptic threads two would have a new combination

of granules and the other two would not show any inter-

change. But on the view that crossing-over occurs earlier

—the identical halves being formed after interchange has

taken place—all four threads would be of a new combina-

tion in those cases where crossing-over of chromosomes in

the region between the two pairs of granules had oc-

curred at all.

There is an essentially similar possibility of finding out

the same thing genetically. For if two threads may cross

over and not the other two, then, if non-disjunction of

X-chromosomes should occur in that maturation division

No.593] THE MECHANISM OF CROSSING-OVER 301

when the threads that crossed would normally have sep-

arated from those that did not, the egg would come to

contain two X-chromosomes, one of which was a cross-

over but not the other. In the usual type of non-disjunc-

tion, the X’s never cross over—presumably because they

paired with the Y (which was present in these cases as an

extra chromosome), so this type of non-disjunction could

not afford a test of the theory. But it is to be expected

that non-disjunction should sometimes occur by mere ac-

cident without the interference of a Y, and since in these

cases the X’s could have crossed over, such cases of non-

disjunction might furnish a test of Janssens’s theory. In

1913, in an experiment designed for this purpose, I ob-

tained a fly which had received two maternal X-chromo-

somes by reason of non-disjunction in its mother, and in

which one of these X-chromosomes proved to be a cross-

over but not the other! The fly resulted from a cross of a

female which contained in one X-chromosome bifid and

vermilion, and in the other chromosome eosin and bar, by

a normal male. It itself contained in one of its chromo-

somes bifid and vermilion, and in the bifid, vermilion and

bar. Since then Bridges has obtained other exceptions

of the same general sort. But on further consideration

it appears that this result really proves nothing, for the

non-disjunction may just as well have taken place in an

oogonial division. In this way an oocyte would result

that contained three X-chromosomes. At synapsis two

of these could cross over with one another, and the egg

could then receive a cross-over chromosome and also an

X that had not crossed over. To prove that the non-dis-

junction was not of this type, but really occurred in a

maturation division, i. e., that the two threads originated

from one tetrad, it would be necessary to obtain individ-

uals in which both of the X-chromosomes received by non-

disjunction had crossed over, but each at a different point

(or one of them at two points).

In a case of the latter sort the fact that both chromo-

somes had crossed over at some point would prove either

that both of them had been in the synaptic tetrad, and so

302 THE AMERICAN NATURALIST [Vou. L

that the non-disjunction had occurred in the maturation

division in the mother fly, or else that both were derived

from the two halves of a single (cross-over) X-chromo-

some which underwent non-disjunction in an embryonic

cell division of the individual itself. But the fact that

the two chromosomes are not identical would rule out the

second possibility. The result, therefore, would mean

that in the same tetrad one strand may have crossed over

at a certain point and not another strand, i. e., that Jans-

sens’s theory is correct and crossing-over takes place at a

stage when there are four threads, two of which may

cross over at a certain point while the others retain their

original composition.

Up to the present, however, no exceptions of this type

have been found, although Bridges has obtained not a few

exceptions of the type that may as well be explained by

non-disjunction in an oogonial division (i. e., in which

one X had crossed over—but not the other), and also one

other exception, which had received two similar double

cross-over chromosomes. The latter peculiar circum-

stance must have resulted either from a non-disjunction, -

at the maturation division in the mother, of two strands

of a tetrad, both of which had crossed over in the same

two places, or from a non-disjunction, in an embryonic

cell division of the individual itself, of the two halves of

the single (double cross-over) X-chromosome, which, on

this view, was originally present. But the latter explana-

tion is very improbable, for, unless the non-disjunction

occurred in the first cleavage, only a small part of the fly

would be composed of cells descended from the one into

which the 2 X’s entered; most of the cells, therefore,

would contain only one X and these would necessarily be

male; thus the fly would be a gynandromorph.. More-

over, all the cells derived from the one which, in the non-

disjunctional division, failed to receive either half of the

X-chromosome, would probably die. Hence the evidence

is fairly good that in this case the two double cross-over

X-chromosomes represent two strands of a tetrad. Since

these two strands, although both double cross-overs, were

No.593] THE MECHANISM OF CROSSING-OVER 303

both just alike, we must conclude either that they were

both derived from the same strand, after it had already

crossed over—in which case crossing-over must occur at a

stage in synapsis before the homologous chromosomes

split to form tetrads—or else that the tetrads were formed

first, and that then crossing-over occurred at two points

coincidently in the case of both pairs of threads, and at

identical points in both. It is not probable, however,

that, if crossing-over occurs at the stage of four threads,

these two pairs of threads would both cross over at the

same points, for according to the observations on which

Jannsens bases the idea that crossing-over occurs at this

stage, a crossing-over of both pairs of threads at the same

place rarely happens. The evidence thus far gained from

non-disjunction is, therefore, rather in support of the

theory that crossing-over occurs at an early stage in

synapsis. ©

D. A Case of Crossing-Over in an Embryonic Cell

It may not be out of place here to record an exceptional

case of crossing-over in the male, which has not been ex-

plained. No other case of crossing-over has hitherto been

found in the male Drosophila. It had been established by

Altenburg and the author that the factor causing truncate

wings is in the second chromosome, and further that the

truncate factor is dominant under certain conditions, but

it does not usually express itself unless certain intensi-

fying factors—one in the first chromosome and one in the

third—are present; even then, the character sometimes

fails to develop. Thus, if a hybrid truncate male is pro-

duced by a cross of a truncate female to a black pink male

(black is in chromosome II and pink in IIT), when this

hybrid is back-crossed again to black pink females, only

the gray flies will carry the factor for truncate, since in

the male truncate can not cross over with the black in the

homologous second chromosome. But few of the gray

flies from such a cross except the gray, red-eyed females

will show the truncate character, for the others will not

contain both of the intensifying factors; and even in the

304 THE AMERICAN NATURALIST [Vou. L

gray-red females the character will not always develop.

A typical count for such a cross was as follows:

Gray Red Gray Pink | Black Red Black Pink

Inter. Norm.

Trune, | Inter. Norm. | Trunc. linter.

Norm. | Trunc. inter Norm. | Trunc.

mel soe eas tae | oe woo poet 0 |o] 82

Bo ae ee el eee 1g es es a

(The count of females is shown on the upper line, the

count of males on the lower.) A brother of the above

male, however, when similarly back-crossed, gave the fol-

lowing count:

Gray Red Gray Pink Black Red Black Pink

öl o (22 | 0 pa 19 16 |20 il o 8 7

0 Ui to 0 0 23 0 6 11 0 3 22

The sex-linked intensifier and the third chromosome in-

tensifier are inherited normally as before, for the females

have wings much more truncated than the males and the

reds are more truncated than the pinks. But, although

the truncate parent of this male contained gray in the

same chromosome as truncate, and the long-winged par-

ent contained black with long, all the truncate has

crossed over, away from the gray factor and into the

chr me with black! Not a single fly has the old com-

bination, gray truncate. It is next to impossible to im-

agine that the chromosomes of the second pair crossed

over in the synapsis period of all the spermatocytes, and

in all of them, between just these particular loci, when -

normally there is no crossing-over at all in the male and

only 30 per cent. of crossing-over between these loci even

in the female. Itis, therefore, necessary to conclude that

crossing-over took place once for all in a cell of the em-

bryo, and that, as usual, it did not occur at all during 3

spermatogenesis, although all the spermatocytes, of

course, inherited the cross-over combination. It is im-

possible to tell whether or not the chromosomes under-

went the regular process of synapsis at this early stage,

No.593] THE MECHANISM OF CROSSING-OVER 305

and whether they crossed over when long drawn out or

when short and thick, but at least the fact remains that

crossing-over may, in abnormal cases, take place in a cell

before the definitive growth period is reached, and even

in an individual (Drosophila male) in which no crossing-

over is the established rule. This fact is not utterly sur-

prising, inasmuch as even in somatic and gonial cells of

Diptera homologous chromosomes show a marked tend-

ency to lie near together (i. e., to attract each other), and

in Metz’s preparations they may not infrequently be

found even twisted about each other somewhat.

The fact that crossing-over occurs only in the female

Drosophila is naturally of great interest, although it is of

unknown significance. In the silkworms, on the other

hand, Tanaka has discovered that crossing-over takes

place in the male, but not in the female. Curiously

enough, although these seem at first sight to be opposite

cases, in both it is true that crossing-over takes place in

the homozygous sex, but not in the heterozygous, for in

Drosophila the female is homozygous for sex, the male

heterozygous, and in the moth these relations are re-

versed. Recently, however, Castle and Wright have pub-

lished data for the rat which, if sufficiently extensive, show

that crossing-over happens in both sexes. The plants in

which crossing-over has so far been studied have all been

hermaphrodites, and crossing-over takes . place in both |

their spermato- and oo-genesis. There is, therefore, at

present no general rule which can be stated, in regard to

which sex crossing-over occurs in. This fact should be

taken into account in weighing the cytological evidence

in regard to crossing-over, obtained in forms in which

the occurrence of crossing-over has not been studied

genetically. For in such cases there is always the possi-

bility that the cytological studies are being conducted on

individuals in which crossing-over does not occur and

which would consequently give results quite irrelevant to

the subjeci.

k (To be continued)

FASCIATION IN MAIZE KERNELS?!

T. K. WOLFE

VIRGINIA AGRICULTURAL EXPERIMENT STATION

In the summer of 1914 a number of different varieties

of corn were crossed for the purpose of studying the

effect of hybridization on the weight of hybrid and pure

seed produced. One of the crosses made was between

Improved Leaming as the seed parent and Boone County

Special as the pollen parent, the pollen of the two varie-

ties being mixed and applied to the same ear. The

former variety is a yellow dent and the latter a white

dent. On this ear was found two kernels, each of which

-had two embryos. The description of the kernels and

their progeny will be given in this paper.

DESCRIPTION OF KERNELS

In corn, the embryo is normally on the side of the ker-

nel toward the tip of the ear. These kernels had an em-

bryo on both sides. The kernels seemed to be normal

with the exception of the extra embryo and a slight prom-

inence or line of demarkation which extended around each

kernel parallel to the embryos.

Kernel No. 1 was yellow in one half, while the other

half was a paler yellow (diluted with white). Kernel

No. 2 was yellow in both halves. Although there was a

variation in the degree of color, the results of the F,

generation proved that both halves of each kernel were

hybrid.

PROGENY FROM KERNELS

The kernels were planted in pots in the greenhouse in

April in greenhouse soil and in due time each kernel pro-

1Paper 2 from department of agronomy, Virginia Agricnltarst Experi-

ment Station, Blacksburg, Virginia

306

No. 593]

Ta ©

fasciated kernels. Stalks on the left,

en n 2.

of kernel No. 1

one fifteenth natural size.)

(About

At first the tassels and silks

were bagged to prevent for-

eign pollination. All the

pistillate flowers were self-

pollinated, the pollen being

applied by hand at this time.

Later, paper tubes were

fastened to the tassel and

carried to the uppermost

ear shoot, the lower ear

FASCINATION IN MAIZE KERNELS

Fı generation progeny of

307.

duced two stalks. In May,

after danger of frost was

over, the contents of each

pot were removed from the

greenhouse and placed in

the field. The time of tas-

seling and silking and other

data were recorded during

the season as shown in

Table I. Each stalk pro-

duced two ear shoots; how-

ever, only one ear shoot on

each stalk produced an ear.

A sk Pome paige from

baw kernel No. oe e left, and

‘ fasciated rie No. on eye right,

showing root systems. ine one fifth

natural size.)

308 THE AMERICAN NATURALIST [Von L

shoots being covered with bags and hand pollinated as

was done at first.

DESCRIPTION OF F, GENERATION STALKS

After the total growth had been made, data were re-

corded as to the height and diameter of stalks, length,

width, and number of leaves, while the dates of tasseling

and silking had been obtained previously.

Fig. 1 shows picture of entire stalks after harvesting.

TABLE I

HEIGHT AND ea oF STALKS, LENGTH AND WIDTH OF LEAVES, IN

. INCHES. NuMBER OF LEAVES AND DATES OF TASSELING AND ee

oF F, ahame Hii STALKS FROM FASCIATED KERNELS OF MAIZ

Date of

Aran Height (Length of L Vidth Le

ey ngth of Leaves Width of aves Diameter| Tassel-| Bilk-

Kernel | Stalk

No. No howl fan Fea of Stalk ing

| 8th 9th | 8th | 9th July

1 1 11 99 337 317 313 343 H 12 26

2 13 102 364 334 is 32 3 16 24

2 3 14 921 | 36ł 343 4H 43 1$ 12 19

4 14 1082 | 402 372 475 435 ; 14 21

At maturity, the entire plants were removed from the

ground in such a way as to retain as many of the roots

as possible. The soil was removed and a photograph

was taken of the roots (Fig. 2) to especially emphasize

the fact that each stalk was separate and distinct from

the other and could not be classed as a tiller from the

other stalk, although both were united at the radicle.

DESCRIPTION oF F, GENERATION KERNELS

Fig. 3 is a photograph of the four ears produced.

All of them show Mendelian splitting. The number and

ratio of yellow and white kernels will be given in Table

II. None of the kernels possessed two embryos like their

parents.

No. 593] FASCINATION IN MAIZE KERNELS 309

a i r

b ae

bey eee Lee

4 “SLLIEErPTY

Ida ad 7

i} Baa j

` generation ears produced by the fasciated kernels. Beginning

2; stalks.

Fic. 3. F

at the left, the first and second ears were produced by kernel No. 2;

numbers 1 and 2, respectively. The third and fourth ears were produced by

Kernel No, 1; stalks 3 and 4, respectively. (About one half natural size.)

TABLE II

NUMBER AND RATIO oF WHITE AND YELLOW KERNELS IN THE F, GENERATION

Kernel Stalk Ear Number Yellow | Number White Ratio of Yellow to White

No. No. No. Kernels | Kernels Kernels

1 1 1 218 | 76 2.86:1

2 2 377 | 60 6.28:1

2 3 3 393 | 182 2.15:1

4 4 408 | 130 3.14:1

! Average ratio, 3.61:1 _

This generation seed will be grown next season in order

to discover whether any fasciated kernels appear. After

these results are obtained, a discussion of the kernels re-

ported in this paper will be presented.

SHORTER ARTICLES AND DISCUSSION

THE INHERITANCE OF SEASONAL POLYMORPHISM

IN BUTTERFLIES

ARE seasonal variations inherited, and may they play a part in

evolutionary change? ‘These are questions which Punnett in

his recent book on ‘‘Mimicry in Butterflies’ answers in the neg-

ative.

In no ease are they known to be inherited, and in no case conse-

quently could variation of this nature play any part in evolutionary

change.

Variations to be of significance in evolution, he tells us, must

be ‘‘transmissible and independent of climatic and other con-

ditions.” +

It would seem to require no demonstration that well-estab-

lished seasonal variations like those of Araschnia levana-prorsa

of Europe in which, it will be remembered, the ground color of

the spring brood (levana) is red-brown, that of the summer

brood (prorsa) black, are transmissible. Under summer condi-

tions in Europe prorsa appears with the regularity of a mono-

typic species, true to type. Monotypic species likewise require

a certain degree of temperature and amount of moisture to pro-

duce their characteristic adult coloration. A. prorsa is by n9

means peculiar in this respect. It has the definitive adult colora-

tion of the species. That which is peculiar is the hereditary

rhythmic tendency to swing from prorsa back to levana, which is

so strong that experimental control can not wholly cope with it.

Summer conditions artificially prolonged result in the appear-

ance of some prorsa in prorsa’s immediate offspring, but some-

times the intermediate, porima, is the outcome. A far larger

number of individuals of the lot under experimentation, how-

ever, refuse to be forced out of the chrysalis by artificial heat, —

hibernate, and become levana. These color variations, therefore,

are not subject wholly to the environment, nor wholly to

heredity. “

A common hereditary basis evidently underlies both of the

1 Pp. 131, 132.

310

No.593] SHORTER ARTICLES AND DISCUSSION git

color patterns. Like produces like, but under natural conditions

only by skipping a generation. Except for the innate tendency

for the types to alternate, the case is similar to that of the red °

primrose described by Baur? which, growing at 15°-20° C. pro-

duces red flowers, at 30°-35° C., white. Or it is like the mutant

stock of Drosophila described by Miss Hoge,’ which, bred in win-

ter or in an ice chest, gives a large proportion of flies with super-

numerary legs, though in summer or in moderate temperature

the stock appears to be normal. The same set of factors under

varying conditions produces different results. An analysis of

the factors underlying another similar case in Drosophila, ‘‘ab-

normal abdomen,’’ has been worked out by Morgan.‘ This re-

markable mutant was shown to behave as a dominant sex-linked

character. It manifests itself, however, only when the food in

which the flies are bred is kept moist.

The rhythmic tendency of prorsa to produce levana, notwith-

standing artificial raising of the temperature, shows that this

is a sort of alternation of generations in which the definitive sex-

ual generation, prorsa, alternates with another apparently more

primitive, levana, which is also sexual. This seasonal alternation

of sexual forms in its hereditary basis is comparable to typical

alternation of asexual and sexual types.

Weismann, in discussing the case cited, assumed the presence

simultaneously in the germ plasm of prorsa-determinants and

levana-determinants.

But these prorsa-ids were at the same time so arranged that they be-

came active under the action of a higher temperature, if this is acting

at the beginning of the pupal period, while the levana-ids become active

at a lower temperature: Heat, therefore, is only the excitant which

sets free the prorsa-determinants, while cold sets free the levana-de-

terminants,

Modernizing Wéismann’s hypothesis, may we suppose that

distinct Mendelian factors underlie each of these two discontinu-

_ ous types of coloration? The idea is attractive, but all the evi-

dence at hand indicates that the determinants or factors of both

types are borne by all the gametes. Intermediates, showing a

2,é ‘ Einführung in die kea ”* pp. 4-6.

8 Jour. Exper. Zool., 18, 1915

4‘‘ Mechanism of Mendelian Heredity,?? pp. 39-41.

5‘*New Experiments on the Seasonal Dimorphism of Lepidoptera.’’

Translation by Nicholson in the Entomologist, 1896.

312 THE AMERICAN NATURALIST [Vou. L

combination of the two patterns, called porima, occur under cer-

tain temperature conditions. The two patterns can not be Men-

delian allelomorphs of each other, though the possibility remains

they may have undiscovered allelomorphs. Too little is now

known of inheritance in this species for us to judge whether

Weismann’s hypothesis in modern form is tenable, or whether a

single set of factors, or single factor, reacting differently to dif-

ferent environments, is sufficient to account for the two types.

In a preliminary analysis of the problem the two color phases

seem like distinct ontogenetic stages. Levana possibly is prorsa

with immature colors, arrested in their development through the

action of cold. Prorsa in the chrysalis may pass rapidly through

the levana stage into its final, complete condition. Its offspring,

however, independently of the environment, hereditarily tend to

hibernate in the chrysalis and become levana. This interpreta-

tion of the two color phases is in line with the facts of dichro-

matism in beetles, as described by MeCracken® and others. Gas-

troidea dissimilis, when it emerges from the pupal ease, is black,

and certain individuals permanently retain this color, others,

however, pass on to a permanent bright green phase. Lina lap-

ponica (Melosoma scripta) has a spotted-brown phase which is

either permanent or is replaced by black. The two color phases

in each of these forms, however, are Mendelian allelomorphs of

each other, the dominant color being that appearing first in ontog-

eny, the recessive last.

The cold weather varieties of Colias eurytheme, about to be

discussed, certainly may be regarded as being produced in large

part by the arrested development of pigmentation. In this most

remarkable seasonally polymorphic butterfly of western and cen-

tral North America, Colias ewrytheme, the writer has found that

the flaming orange coloration of the summer form (usually called

the typical eurytheme) and the paler orange-yellow of the spring

and autumn broods (ariadne and keewaydin) are variations also -

due to differences in the reaction to the environment of definite

Mendelian factors. This has been shown by crossing the orange

eurytheme of the central and western states, with the clear yellow

species of the eastern and central states, Colias philodice, the

yellow of which segregates cleanly from the orange in F,, as a

recessive. The hybrids, as well as the eurytheme stock, show

seasonal polymorphism. The F, hybrids, for example, are of a

6 Jour. Exper. Zool., 3, 1906,

lage

aes

No.593] SHORTER ARTICLES AND DISCUSSION 313

dilute orange. Orange is therefore incompletely dominant. The

heterozygote is an intermediate. The amount of orange pigmen-

tation, or the degree of its dilution, in the F, hybrids, however,

varies prodigiously with the season. The summer-bred hybrid

is of dilute orange (‘‘apricot yellow’’) somewhat evenly dis-

tributed over the wings, but in the small winter-bred individ-

uals, suchas emerge in the greenhouse in December, the orange

is restricted to a faint flush near the posterior (inner) margin of

the fore wings.

But even though the underlying hereditary basis supports a

superstructure that varies widely, are these variations as such

inherited? The variety of ewrytheme called ariadne is small and

of a pale orange hue. This form appears under cold weather con-

ditions only. Its dwarfness is due to the failure of the cater-

pillar to feed during the late fall to full size, though food is

abundantly supplied. The shortness of the day evidently is a

factor in checking the feeding. The caterpillar forages actively

at mid-day, but becomes sluggish before nightfall, yet it matures

even while it is not feeding, and hence produces a dwarfed pupa.

The pale color may be readily explained by the supposition that

the elaboration of chromogenic substances in the blood of the

pupa is checked by the cold so that these materials ripen in the

scales of the wings merely into faint orange and yellow.

Ariadne’s progeny in June are not ariadne, but a large and

brilliantly orange insect. Are ariadne’s size and hue, therefore,

not inheritable, but dependent wholly upon the environment?

At first thought this would seem to be the fact, and this was evi-

dently the view of the matter presenting itself to Punnett when

he stated that such variations are not inherited. We have seen,

however, that they are hereditary in the sense that this organ-

ism must react in this particular way under these particular con-

ditions. Its inherited organization compels, determines, thi

reaction.

These seasonal variations are therefore transmissible, though

they are by no means ‘‘independent of climatic and other condi-

tions.” May they, therefore, play no part in evolutionary

change? We should not yet be dogmatic as to this. The under-

lying hereditary basis in all probability is as susceptible to muta-

tion, to disturbances in the arrangement and nature of the

chromosomal elements, as any other germ-plasm. It would seem

by no means impossible that the alternating phenotypes of A.

314 THE AMERICAN NATURALIST [Vou. L

levana-prorsa, for example, might in suitable diverse climates,

subarctic and tropical, respectively, be fixed as separate species.

A. levana bred in Labrador, for example, where it could pro-

duce only one brood, probably would not show its prorsa-produc-

ing tendency at all. This supposition is confirmed by Trybom’s

observation (quoted by Weismann) that in Siberia, where a single

brood occurs yearly, it is levana only. Conversely, prorsa in the

tropics would perhaps eliminate all traces of levana, though of

this we can not be so confident.

How much practically identical germ-plasm in different parts

of the world is masquerading as different species, because of the

diverse ways in which it reacts to different environments in which

it happens to be placed, has not been adequately investigated.

An interesting example of the sort among tropical reef fishes was

recently cited by Longley.’

Bodianus fulvus and B, punctatus are two color phases of ‘one species

of which one may almost instantaneously replace the other.

The rapid interchange of reproductive habits between Sala-

mandra maculosa and the alpine atra when transferred respec-

tively to lowland or highland conditions, as described by Kam-

merer, is probably also a case in point, due to fundamentally

similar germ-plasm in both forms. Such an assumption would

account for the inheritance of these readily acquired characters.

If the facts are correct, S. maculosa, by cold and drought, was

` forced to assume the reproductive habits of the salamander of

the neighboring alpine regions named atra, producing two adult

larve viviparously, rather than many (14-72) immature embryos

laid in water as is its habit in the warm, moist lowlands. Con-

versely, the alpine form was forced by heat and an ample water

supply to increase its fecundity from two to nine larve at a birth.

In both cases the ‘‘acquired characters’? were inherited, as we

would expect them to be if the two kinds of salamanders,

regards reproductive mechanism at least, have an identical

similar genotype. The fact that the lowland form living

higher altitudes has fewer young, and that the alpine form

the lower regions of its range produces an abnormally large

number (viz., four) also points to the same conclusion.

It is a possibility worth considering that somatic modification

accompanied by little germinal change may partially explain

7 Carnegie Institution of Washington, Year Book, No. 14, ses p. 209.

8 Archiv f. Entwicklungsmechanik, 25, 1907.

BAS &

No.593] SHORTER ARTICLES AND DISCUSSION 315

the remarkable ‘‘mimicry-rings’’ of South America that have

been so interestingly discussed by Punnett. In each of several

great regions of that continent a characteristic color pattern is

exhibited by unrelated species belonging to different genera and

even to different families. The color pattern followed in Central

America differs slightly from that adopted by members of the

Same genera in eastern Brazil, a single genus of Pierids only

dropping out of the ring in the latter region. In western Brazil

and the upper Amazons the pattern is somewhat more mottled

and the ground color darker, but the same genera are represented

almost without exception. Finally, in Ecuador, Peru and Bolivia

the pattern common to the different genera is still darker and

greatly simplified. The Pierids here have left the ring, and a

Papilio, an Acrwa and two species of the Satyrid genus Pedali-

odes have entered it. The point to be emphasized, however, is

that the same genera, e. g., Heliconius, Mechanitis, have repre-

sentatives in each local color group.

Let us now assume, with Punnett, that a set of similar or iden-

tical color factors is common to all the structurally diverse mem-

bers of each ring, and add the further hypothesis that the color

pattern resulting from these particular factors is to a large ex-

tent influenced by climatice conditions, as in seasonally poly-

morphic insects. It then follows that certain members of a

‘‘ring’’ migrating from the tropical climate of Brazil to the

temperate zone farther south, even before they should become

changed genotypically, would react to the new environment by

assuming a new color pattern such as that now characteristic of

the south temperate zone. If the genotype were identical

throughout the migrating group and a single member of the

group should so react, all naturally would respond in the same

manner. Seasonal polymorphism thus may furnish an addi-

tional clue to the explanation of this most interesting case of

convergence and parallelism in evolution.

This discussion leads to the conclusion that seasonal variations

have a hereditary basis more sensitive than that of other color

characters to temperature and other climatic conditions. A sea-

sonal variation that is constant in its recurrence is transmissible.

Its hereditary basis invariably reacts in a certain definite way to

a certain narrow range of external conditions, whereas the hered-

itary basis of other characters, e. g., eye color in vertebrates,

316 THE AMERICAN NATURALIST [Vou. L

reacts in a precise way to a far wider range of external conditions.

Seasonal variations, as was pointed out by Weismann, show a

hereditary tendency to alternate which, in some cases, is inde-

pendent of external conditions.

Seasonal varieties are in some eases (e. g., Colias, and possibly

Araschnia) to be regarded as distinct ontogenetic stages. Cold

arrests development at an early phase in color metabolism, and

the mature insect emerges with pale colors (Colias ewrytheme

var. ariadne), or with a color pattern different from the defini-

tive coloration of the species (Araschnia levana).

The suggestion is made that local color varieties, passing it

may be for distinct species, are probably in some eases the equiv-

alents of seasonal variations. That is, they are the product of a

genotype sensitive to environmental changes expressing itseif

under a particular set of local climatic conditions; elsewhere the

same genotype may respond quite differently. Such phenomena,

though not of profound evolutionary significance, may play a

rather conspicuous rôle in the evolution and diversification of

the colors of animals and plants.

Joun H. GEROULD

VARIATIONS IN THE VERMILION-SPOTTED NEWT,

D. VIRIDESCENS

WHILE carrying on some experiments with the spotted newt,

Diemyctylus viridescens, I was struck with the variation in the

size, number and arrangement of the black-bordered vermilion —

spots so characteristic of this beautiful little salamander. .

It is now generally recognized that this species exhibits two

phases which were formerly described as distinct varieties or even

species. As described by Gage! the young animal, which is ter-

restrial in habits, is red in color and was formerly called D.

miniatus; later it becomes aquatic and its ground-color becomes

olivaceous—permanently so, according to Gage. Against this

dark ground-color (which is subject to considerable variation

under different conditions even in the same individual) the bright

red spots with their black borders stand out very strikingly.

It was with the olivaceous phase that I was experimenting, and

it is upon this phase that the following observations are based.

1 Gage, S. H., ‘‘The Life-History of the Vermilion-Spotted Newt,’’

AMER. Nart., December, 1891, pp, 1084-1103.

No. 593] SHORTER ARTICLES AND DISCUSSION 317

All the drawings were made from preserved material in which

the vermilion spots had mostly faded to a white or pale pink

color.

The first twenty figures were made from about three dozen

Specimens, probably all from the neighborhood of Morgantown.

The last four figures are from animals that had been obtained

from the Marine Biological Supply Company, Woods Hole, Mass.,

318 THE AMERICAN NATURALIST [Vou. L

and had died, from time to time, in the laboratory aquaria. The

mid-dorsal ridge is indicated in the figures by the dotted line.

Only the black-bordered vermilion spots were noted, the small

black spots being too numerous and irregular to make it worth

while to study them.

It will be noticed that in the animals from Woods Hole, shown

in figures 21 to 24, the red spots were much smaller than most of

No.593] SHORTER ARTICLES AND DISCUSSION 319

those on the animals from Morgantown. This was true of nearly

but not saad all of the animals obtained from the north.

Cope says :?

On each side of the vertebral line is a row of from three to six small

round red spots, each with a black border. The rest of the surface is

marked with small black points, which are smaller but more distinct

on the lower surface.

Among all of the animals examined no two were spotted alike.

They were sorted into groups according to the total number of

red spots. The smallest number of red spots found was six; they

were all of large size and arranged as shown in Fig. 1; only one

animal with this number of spots was found.

Four animals were found that had seven red spots; Figs. 2 and

3 show the arrangement of the spots on two of these animals; all

of the spots were large and of about the same size.

Four animals exhibited eight red spots, mostly large and of

uniform size; two arrangements are shown in Figs. 4 and 5.

Three animals had nine red spots each, mostly large and of

uniform size; Figs. 6 and 7 show two arrangements of these spots.

Seven animals had ten red spots each, this being the largest

number of animals found in any group. The spots were mostly

large and uniform in size; two arrangements are shown in Figs.

8 and 9. It will be noticed that in Fig. 8 the spots are arranged

in fairly regular pairs.

Five animals had eleven red spots of somewhat more variable .

size than in the preceding. Figs. 10 and 11 show two arrange-

ments of these spots; and Fig. 10, especially, shows wide varia-

tions in the size of the spots.

Three animals exhibited twelve red spots of variable size, two

arrangements of which are shown in Figs. 12 and 13.

Two animals, shown in Figs. 14 and 15, exhibited thirteen red

spots of various sizes.

Two animals had ENE i red spots; one of these animals is

shown in Fig. 16.

Figs. 17, 18, 19 and 20 show the arrangements of red spots on

four animals that had 15, 24, 29 and 39 spots, respectively. It

will be noticed in these animals, especially in the last, that the

large number of red spots is due to an increase in the number of

very small spots, the number of large red spots being no greater

than in the earlier individuals. Thirty-nine was the largest num-

2 Cope, E. D., ‘*The Batrachia of North America,’’ Bull. U. S. Nat.

Mus., No. 34, 1889, p. 210.

320 THE AMERICAN NATURALIST [Vou.L

ber of red spots found on any single animal. Only one animal in

each of these last five groups was found. nM

Figs. 21 to 24, as noted above, represent animals obtained from

Woods Hole; they have 11, 16, 19 and 20 spots, respectively, and

it will be noted that all of the spots are small and of fairly uni-

form size.

CONCLUSION

- It would seem from this hurried survey that the number, size _

and arrangement of the vermilion spots, so characteristic of D.

viridescens, are quite variable, probably two animals very seldom

being even approximately alike.

3 ALBERT M. REESE

West VIRGINIA UNIVERSITY,

MORGANTOWN i

VOL. L, NO. 594

1AE

AMERICAN

Special Reference to the Factors of Evolution .

The American Naturalist

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

sent to the Editor of THE AMERICAN NATURALIST, Garrison- -on-Hucson, New York.

Short articles containing summaries of research work bearin ng on the

proies s of — siete are especially welcome, and will be given preference

in publicat

One ouia ea reprints of ne preside are supplied to authors free of charge.

Further reprints will be supplied atc

riptions and advertisements should be sent to the „publishers. The

subscription price es four dollars a year. Foreign postage is fifty cents and

Canadian postage twenty-five cents additional. The charge for single copies is

forty cents. The advertising rates are Four Dollars for a page.

THE SCIENCE PRESS

Lancaster, Pa. Garrison, N. Y.

_ NEW YORK: Sub-Station 84

Entered , April 2, 1908, at the Post Office at Lancaster, Pa., under the Act ot

Congress of March 3, 1879.

NATURAL HISTORY OF MAN | JAPAN NATURAL HISTORY SPECIMENS

18,000 Specimens illustrating ne daily life of ect Condition and Lowest Prices.

Specialty: Bird Skins, Oology, Entomology, Marine

Seon vage Races, many now extinct. pons,

ts, Deities, Currency, Carvings, Tools, etc. | Animals and others. Catalogue free. C

patio pae eepe a: and age. Type Collections for ence solicited.

— a a T. FUKAI, Naturalist

OLDMAN, 77 Brixton Hill, London, England Konosu, Saitama, Japan

| PUBLISHED BY THE CAMBRIDGE UNIVERSITY PRESS

| MIMICRY IN BUTTERFLIES

Bound in Buckram. Royal Svo. Pp. viii+ 188. „With 16 plates (12 of which

| colored). Price $4. 4

D a CRUNDALL Poynert, F. R.S., Fellow of Gonville and Caius College, |

z Arthur Bailor Professor of Genetics i in the University of tee eRe he

_ EXTRACT FROM THE PREFACE

: i. ; : oe has been written in the kapes that it may peel! to i oe

= kesi Re,

THE

AMERICAN NATURALIST

VoL. L. June, 1916 No. 594

NEW LIGHT ON BLENDING AND MENDELIAN

INHERITANCE

Proressor W. E. CASTLE

Bussey INSTITUTION, HARVARD UNIVERSITY

Taer question whether blending or intermediate inheri-

tance involves Mendelian principles is of prime interest

to every student of genetics. The further question

whether Mendelian characters are constant or variable is

not less important. New light is shed on both these ques-

tions by a recent and valuable investigation by Y. Ho-

shino, ‘‘On the Inheritance of the Flowering Time in Peas

and Rice.’’

The investigation is to be most highly commended for

its thoroughness. It has involved the raising of over 30,-

000 plants extending over a period of 8 years. F, and F,

generations have been raised three times and F, twice.

The publication contains a careful summary and analysis

of the facts observed. Above all it contains facts with a

minimum of speculation, not facts so marshalled as to

prove a particular theory and otherwise useless, but facts

which the reader may study from any angle he chooses.

I propose to exercise the privilege which our author’s

commendable method of publication makes possible, of

utilizing his facts for testing a slightly different theo-

retical interpretation from that which he adopts, for I am

acquainted with no material equally valuable in rela-

- tion to the two questions stated above, though I have been

engaged for many years in studying these same questions.

Hoshino crossed two varieties of peas which differed in

321

322 THE AMERICAN NATURALIST [ Vou. L

flowering time by about three weeks. The early variety

was white-flowered, the late variety red-flowered. The F,

hybrids were only a little earlier than the late variety in

time of flowering and of course had red flowers, this

being a dominant character. The F, hybrids varied in

time of flowering throughout the range of both parent

varieties and all intermediate points, but not beyond the

range of the parents. Hoshino divides the F, plants into

two main groups, an early flowering and a late flowering

group, the two being separated naturally by a class of

‘‘ minimal frequency’’ usually recognizable. A majority

of the plants in the early flowering group had white flow-

ers, a majority in the late flowering group had red flowers.

This result shows that early and late flowering have a

tendency to segregate from each other as Mendelian alle-

lomorphs and that time of flowering is coupled with

flower color. This fact of coupling, first observed by

Lock, alone would suffice to show that time of flowering

is inherited as a Mendelian character. The fact that the

F, individuals apparently fall into two main groups, not

into four or eight, indicates that one important Men-

delian factor is concerned, not two or three. But Hoshino

found that in F, four (not two) groups of relative con-

stancy were obtained and this has led him to suppose that

the inheritance of time of flowering depends upon two

independent Mendelian factors, one of which (A, a) de-

termines late or early flowering, while the second (B, b)

supplements or modifies the action of A. He regards A

as the principal factor, with which alone flower color

shows coupling. The action of B is regarded as subsidi-

ary. A characterizes late-flowering plants ; a, early-flower-

ing plants. B makes the flowering time later than it

would be otherwise; b of course makes it earlier. The

four supposedly homozygous sorts which are recognized

are of the following formule:

aabb, early; aaBB, intermediate early; AAbb, inter-

mediate late; AABB, late. Since red color of flowers is

by hypothesis coupled with A, red flowers will predomi-

nate in the last two groups, white flowers in the first two

No.594] BLENDING AND MENDELIAN INHERITANCE 323

as observed, and the dividing line between them will cor-

respond with Hoshino’s class of minimal frequency in F..

But the plants which Hoshino classified, on the basis of Fy

and F, tests, as extracted early (aabb) and extracted late

(A ABB) are usually not quite so early or late, respect-

ively, as the uncrossed races. Hence he assumes the

occurrence of ‘‘gametic contamination,’’ and recognizes

classes of ‘‘would be’’ or ‘‘pseudo’’-early and ‘‘ pseudo’’-

late. He also notes the occurrence of ‘‘qualitative’’ varia-

tion within the groups classed as ‘‘constant’’ early inter-

mediate and ‘‘constant’’ late intermediate. That is one

family supposed to be of the same genetic formula as

another may throughout its entire range produce plants

slightly earlier or later than those of the other family.

This behavior is not ascribed to any difference in genetic

formula, but to a slightly different value of thé same gene

in the two families.

The late race employed was also found to vary in late-

ness, one ‘‘pure line” derived from it being later than

another. Crosses made with these two lines are reported

separately. No factorial difference is recognized between

them. Each is AABB, but one flowers a little later than

the other and transmits this property to its descendants.

Thus ‘‘qualitative’’ variation of a gene, i. e., variation in

its potency, is recognized by Hoshino. Aside from the

occurrence of two pure lines in the late race, Hoshino con-

siders ‘‘the flowering time quite fixed and unchangeable in

the parent varieties,’’ and cites his Tables IV-VI in sup-

port of this idea. Table VI is of particular interest in this

connection because in this case seeds were planted of the

earliest flowering and latest flowering individuals of the

same pure line and descended from the same individual

grandparent. These plantings constitute a test of the

existence of genetic variation within the pure line. The

progenies of the same grandparent plant (but of different

parents) are so obviously alike and so little variable in

flowering time that Hoshino has not considered it neces-

sary to calculate their mean flowering time. But if this is

done it affords unmistakable evidence that genetic varia-

324 THE AMERICAN NATURALIST [ Vou. L

tion occurs within these ‘‘pure lines’’ (see Table I). For

in all but the last two of the thirteen cases tested the

earlier parent has the earlier progeny. From long ex-

perience in studies of rats with such small differences as

are here indicated I have no hesitation in concluding that

fluctuating variation of genetic significance is here in

evidence.

To recapitulate, as regards genetic variation in flower-

ing time, Hoshino (1) recognizes that gametic contami-

nation results from crossing early and late flowering

varieties; (2) recognizes also that variation may occur

among the cross-bred families, as well as in different

pure lines of the uncrossed races, as regards the ‘‘qual-

ity,’’ value, or potency of the same gene. (3) Although

Hoshino does not refer to the fact, his observations show

clearly that genetic variation of a graded or fluctuating

sort occurs im at least one of the varieties which he

crossed. It is probable that those varieties were as pure

as are obtainable, but almost certain that their flowering

time fluctuates slightly from genetic causes.

What I want to suggest is that in these several agen-

cies we have a sufficient explanation of the variations

observed in Hoshino’s F,, F, and F, generations, with-

out invoking a two-factor hypothesis, one factor being

enough. Hoshino has shown that a three-factor, or

multi-factor hypothesis will not fit the facts observed.

Will not one factor fit them quite as well as two, pro-

vided gametic contamination occurs, which he admits? The

‘‘pseudo-early’’ and ‘‘pseudo-late’’ classes Hoshino ex-

plains plausibly as due to gametic contamination. Could

not the ‘‘intermediate early’’ and ‘‘intermediate late’’

be reasonbly explained as due to further contamination?

For they intergrade with the pseudo-early and pseudo-

late classes, respectively, and also with each other. From

Hoshino’s carefully controlled results, it is perfectly

clear that early and late flowering are allelomorphs, and

that segregation of early and late types occurs in F, but

attended by gametic contamination. It is perfectly clear

that the contamination is not uniform in amount. Some-

No. 594] BLENDING AND MENDELIAN INHERITANCE 325

times little or no contamination is observable; sometimes

it is considerable.

TABLE I

COMPARISON OF THE MEANS OF FAMILIES DESCENDED FROM THE SAME PURE-

LINE GRANDPARENT, BUT FROM PARENTS OF UNLIKE CHARACTER

Based on Hoshino’s Takle 6.

Í |

Designation of Grandparent Moan EN pra gg Yay Fog td ‘as | Difference

FLEUR TE ae ee S 59.96 60.11 15

OVE ore. ce about 59.94 60.65 71

Gp RECS ot a cod ie eGo 60.16 60.50

p DS sie aed PS E Ae 63.65 63.9 33

APs VER io Ohta son ane eee T 63.86 64.21

EO EE rarer | poke Wray, oS 60.25 60.58 33

Oi Re isis Ci eee eee 59.43 60.05 62

C OL sd ta Vay ewes: ie ena 62.91 64.18 1.27

Ee aka raS S te een .23 64.56 3

ETAP A A E A PR AE 66.45 66.63 18

COBB SE ea EE YE T, 67.30 1.08

Gp VILIS: o a aaa 71.75 71.20 -.55

GOERE; cx ceyhanaadn cae 71.25 -—.75

Contamination sufficiently great would account for the

intermediate early as a modified early class and the inter-

mediate late as a modified late class. The matter of

coupling is unaffected by this hypothesis, since coupling

is shown with only one factor, Hoshino’s factor A. The

observed variability and intergradation of the two inter-

mediate classes favors a hypothesis of contamination

rather than one of an independent modifying factor.

If we suppose modification due to gametie contamina-

tion to occur in half of the gametes formed by F, indi-

viduals and that this contamination is definite in amount

(say equivalent to 5 days) the F, expectation would be

exactly the same as from a two-factor system such as

Hoshino adopts. As a matter of fact neither supposition

is exactly correct. If we adopt contamination as a suffi-

cient explanatory hypothesis, we must suppose the

amount of contamination to be variable; if we adopt a

definite modifying factor, we must suppose the amount of

modification to be variable.

To make the matter clear, let us suppose the early race

to be stable at 35 days between sprouting and flowering

326 THE AMERICAN NATURALIST [Vou. L

and the late race to be stable at 55 days. According to

Hoshino’s hypothesis aabb stands for a 35-day period,

and AABB for a 55-day period. If we assign to the

assumed modifying factor B a delaying effect of five days,

then the class aaBB (early intermediate) will have a

value of 40 days, and the other homozygous class AAbb

(late intermediate) will have a value of 50 days. Hetero-

zygous classes will be intermediate as follows:

TABLE II

' COMPOSITION OF F, ON HoSHINO’s Two-FACTOR HYPOTHESIS, WITH EQUIVA-

LENTS IN DAYS FROM SPROUTING TO FLOWERING

Heterozygous

Designation and Expected Frequency Homozygous a Ee

Monohybrid Dihybrid

SBA ce Cos GRE a eee 35

MORENO ig ie ee ee eS 37.5

ROAD ie Ror ee a 40

MGS Siok sob vin’ GO Paes ese 42.5

MPs Gao ve Oo So ee ies 45

2 BABS ets poo a a 47.5

ERAO iss aSa a 50 5

RABE. Ve. See iGee beers ee 52.5

1 AABB Wied E EE E R EE E A 55

- If, however, we replace the modifying factor B in this

scheme by a modification in A amounting to 5 days, then

we can dispense with B and yet obtain exactly the same

classes and in the same numerical proportions and with

nearly the same expectations as regards their breeding

capacity. Let us assume that a stands for a 35-day-

period, A for a 55-day period, and that modified a, which

we will call a’, stands for a 40-day period, and modified

A, which we will call A’, stands for a 50-day period and

that all are allelomorphs of each other. Then the F,

gametes will be a+ a’+ A’ +A and F, will contain the

classes shown in Table III.

It is evident that both schemes fit the observed facts

fairly well. Either one will explain the decreased varia-

bility of F, as compared with F, and the production of

several different types of F, families differing in the

amount of their variability, some of which are relatively

constant. But the former scheme will not answer with-

No. 594] BLENDING AND MENDELIAN INHERITANCE 327

out the further assumption of gametic contamination

(which Hoshino makes) and the latter must invoke con-

tamination of different degrees in order to explain the

pseudo-early and pseudo-late classes. The former scheme

involves two explanatory principles, the latter only one.

Other things being equal, the simpler hypothesis is to be

preferred.

TABLE III

COMPOSITION OF F, ON A ONE-Factor HYPOTHESIS WITH CONTAMINATION

o DEFINITE AMOUNT (5 Days) IN HALF THE GAMETES

Heterozygous, All Ménohybrid

‘Designation and| Homozy-

ETTE as Coat, prose 5 | Difference 10 Difference 15 Difference 20

BE ETET 35

e E AED ewan 37.5 (35:40)

ae on 40

DAA E 42.5 (35:50)

DBA eS rs 45 (35:55)

TWA or Tg 45 (40:50)

y E E AASA 47.5 (40:55)

LEA es, 50

r r. ENEE 52.5 (50:55)

ERA ye 55

Experiments decisive between the two hypotheses are

difficult to devise, but certain tests are possible. On the

hypothesis of Hoshino one would not expect to obtain a

class splitting into homozygous early intermediate and

homozygous late intermediate. On the alternative hy-

pothesis such a class (40-50) should be obtainable. A

“constant” class exactly intermediate between the parent

varieties (say 45) would be impossible on Hoshino’s

hypothesis, unless he is willing to admit an indefinite

amount of contamination, which, however, would render

the two-factor hypothesis superfluous. On the alterna-

tive hypothesis such a constant intermediate class should

be obtainable after a sufficient number of inbred genera-

tions. In reality Hoshino’s observations show that it is

obtained in F, and is then more abundant than any other

‘“‘constant’’ class. The largest group of hybrid offspring

in Hoshino’s experiments and that belonging to the latest

inbred generation is his F, generation raised in 1914 (see

328 THE AMERICAN NATURALIST [ Vou. L

his Tables G-U). This includes 231 families classed by

Hoshino as ‘‘constant.’? They are descended from 15

different F, individuals ranging in flowering time (in

1912) from 48 to 65 days, this being practically the entire

F, range. The combined range of the F, families raised

in 1913 extended from 45 to 71 days. From certain indi-

viduals, selected at intervals throughout the ranges of the

fifteen F, families so as to represent their complete vari-

ability seeds were planted which produced the F, genera-

tion. 421 such F, families were reared and studied and

of these Hoshino regards 231 as ‘‘constant’’ because of

the limited oa of variation in flowering time of each.

The others are regarded as still heterozygous. The mean

flowering time of each of the ‘‘constant’’ families has been

calculated by Hoshino and these means have the distribu-

tion shown in Table IV (omitting fractions from the class

magnitudes which would make them .5 greater than those

given in the table, but would not affect their distribution).

It should be noted that on account of the peculiar weather

conditions of 1914, the flowering time came about 10 days

earlier than in the two previous years, the range of the F,

means extending from 34 to 59 days, whereas the F, range

of flowering time was from 45 to 71 days. Since both

upper and lower limits of the range are displaced by like

amounts and in the same direction, the general character

of the distribution is not affected thereby.

TABLE IV

CLASSIFICATION OF MEANS oF F, ‘‘ CONSTANT’? FAMILIES FROM HosHINO’s

TABLES G-U

Classes AQ) 41| 49| 423 44 4E AR 47 4R ADO EO 51| 59) 52) BA! BR 57/58 59

Z1 24| 29 24 49 40 4/ 45

White..| 1/13152 3 |1812 15/14/13.12 3

Red or

mixed Birt 1} 2} 1) 4/13/17) 606/2 5/3/3|2/11/16/3 (10 8}1

[m |

Totals .| 1 115/6 |3 4 9|2 513/312 /11|16) 3 110 8 11

Table IV shows unmistakably that the F, constant

families fall into three natural groups, not four as

Hoshino’s hypothesis would lead one to expect. The

No. 594] BLENDING AND MENDELIAN INHERITANCE 329

modes of these three groups fall at 35, 44 and 54, respect-

ively, with frequencies of 15, 30 and 16, respectively.

Each group is separated from the next adjacent by a gap

(a class of 0 frequency). The modes of the early and

late groups correspond closely with those of the uncrossed

early and late parent varieties, which in this season had

modes at 35 and 56, respectively. The third and largest

group has its mode almost exactly midway between the

modes of the early and late groups, with 8 intervening

classes below it and 9 above. It would be difficult to imag-

ine a finer example of a stable intermediate class produced

by hybridization. For it will be remembered that every

part of this population is stable, since it includes only

families shown by Hoshino’s breeding tests to be reason-

ably constant, those which he actually pronounced ‘‘con-

stant.’’ From it, therefore, one would need only to choose

families of the desired flowering time, in order to have a

complete succession of varieties from very early to very

late. |

But it may be asked, is the middle group possibly an

‘‘early intermediate’’ group of Hoshing’s formula aaBB

separated from the later groups by a class of minimal (0)

frequency, as in the F, distribution? If so it should

contain very few red-flowering families, no more indeed

than the early group itself, since each would obtain red-

flowered families only by cross-overs. Inspection of

Table IV shows that this hypothesis is untenable, The

truly hybrid origin of the middle group is shown by the

large number of red-flowered or mixed families which it

contains. Nine out of 12 of the F, families which con-

tributed to the production of the middle group contained

red-flowered or mixed red and white families. The middle

F, group itself contained 90 white-flowered to 52 red-

flowered or mixed families, whereas the early group con-

tained 21 white: 4 red, and the late group contained only

red-flowered or mixed families.

Hoshino observed that the flowering time of F, plants

was close to that of the late parent, being only 2 or 3 days

330 THE AMERICAN NATURALIST [ Von. L

earlier, and considers this to be evidence of dominance,

but I am inclined to think it should be interpreted differ-

ently, for F, plants having the genetic properties of F,

plants are in some cases at least (Hoshino’s Tables T and

V) much earlier in flowering time, being in fact almost

exactly intermediate between the parent races, although in

one family (S) the F, plant was late like F,, but its F, and

F, descendants covered the entire range, as did those of

T and U. I am inclined to interpret the inheritance as

truly intermediate and to explain the lateness of F, and

of an occasional F, individual as due to physiological non-

genetic causes. Recent observations made on size in-

heritance in guinea-pigs together with certain observa-

tions recorded by Hoshino lead me to this conclusion.

When the small Cavia cutleri is crossed with the rela-

tively large guinea-pig, the F, hybrids are larger than

either parent, but the F, hybrids as a group are close to

intermediate and only a little more variable than F,. A

stimulus due to crossing makes F, larger than its genetic

constitution would otherwise make it, but the added size

due to this stimulus does not persist to any great extent

beyond F,. Hoshino’s F, peas probably possess a simi-

lar vigor due to crossing, which quickly disappears in

later inbred generations. If this vigor due to hybridiza-

tions causes extra growth it may delay flowering time,

for Hoshino, confirming Keeble and Pellew, has shown

that late-flowering plants have longer internodes than

early-flowering ones.

Hoshino, as we have noted, divides his F, families

into two portions, early-flowering (chiefly white) and

late-flowering (chiefly red). But we have seen reason

to think that the F, families fall naturally into three

portions, early, medium and late, and it is possible

to divide the F, families in a similar way, though of

course somewhat arbitrarily, classing as early those fall-

ing within the range of the early parent or a little beyond

it, and as late those which fall within the range of the

late parent, while those which lie between are placed in

No. 594] BLENDING AND MENDELIAN INHERITANCE 331

the middle class. When the F, families are treated in

this way we find the distribution of white-flowering and

red-flowering plants shown in Table 5.

TABLE V

DISTRIBUTION OF WHITE- AND OF RED-FLOWERING AMONG THE F, PLANTS

Early Medium Late

Hoshino’s Table y Total

White | Red | White | Red | White | Red

Table 2... EE N E eae | 49 13 20 114 7 168 371

Table 8. Ass cin Soe eons 10 3 10 62 1 11 97

TADOS, As. I AANA | 17 2 2 20 3 25 69

Pahl 8B oie vic re cx ose 4 0 6 25 0 7 42

ORNS Ba oe oe ne 17 0 7 37 5 30 9

Thee, CL | 26 | 24 eit cafe ate a

out OLE oe es 123 | 42 | 48 | 295 | 21 | 286 | 815

Fér Gent. White. 5 oes. as | 74 14 6.8 ace

Hoshino uses his F, tables as a basis for calculating the

strength of the coupling between earliness and flower

color, and concludes that the coupling is approximately

7:1. If, however, F, is divided as in Table 5, the coup-

ling appears to be less strong, probably about as 4:1 or

5:1. The percentages of white-flowered plants expected

in each group on various integral coupling ratios are as

follows:

Coupling Early Group Medium Group Late Group

oo ae ae Py avec P ERT] 64% 16 % 4%

GI As See ee 69% 14 % 2.8%

S E E A AAE AE 73% 12 % %

Rar Be ROMER YAN TE 76% 10.9% 1.5%

Observedi i. on eds 74% 14 % 68%

It will be observed that the percentage of white-flowered

plants in the early group indicates about a 6:1 coupling

ratio, but in the medium group, it indicates a 5:1 ratio,

while in the late group it would indicate a 3:1 ratio.

Much uncertainty exists as to the classification of many

of the F, plants as regards flowering time, because of

irregular and delayed germination. Undoubtedly the ~

classes early, medium and late oo so that not much :

332 . THE AMERICAN NATURALIST [Vou. L

reliance can be placed on the categories adopted. But

the figures given indicate that Hoshino has estimated too

high the coupling strength, and that more probably it

does not exceed 5:1. This is not due to any inaccuracy in

Hoshino’s calculations, but would follow only if the

hypothesis suggested in this paper is substituted for

Hoshino’s hypothesis.

If I have correctly interpreted Hoshino’s observations,

flowering time in peas is clearly a Mendelian unit-char-

acter, entirely devoid of dominance, so that a strictly

intermediate hybrid form is the commonest end-product

of a single cross between early and late varieties. Fur-

ther, segregation is imperfect so that blending results,

which becomes more and more complete with each genera-

tion of inbreeding. From the incompleteness of the blend-

ing in the F, zygote and so the imperfection of the segre-

- gation in the F, gametes, it follows that many different

types of F, zygotes are produced, some of which are

practically constant (homozygous) particularly those at

either extreme of the series (the ‘‘early constant,’’

‘‘ pseudo-early constant,’’ ‘‘late constant’? and ‘‘pseudo-

late constant’’) and also at two intermediate points (‘“in-

termediate early’’ and ‘‘intermediate late’’).

Other F, zygotes, resulting from the union of gametes

quite dissimilar, produce a highly variable F, progeny,

but one which will give rise to F, families individually

less variable for two reasons: (1) because the process

of blending continues and so gametes produced by the

same zygote become more like each other than were the

parent gametes of that zygote, and (2) because hetero-

zygotes under self-fertilization tend to produce about

50 per cent. of homozygous offspring, while homozygotes

produce only homozygous offspring.

The entire population therefore will in accordance with

recognized Mendelian principles gradually resolve itself

into relatively constant self-fertilizing lines. But because

of the slow but continuous blending which occurs, these

pure lines will in a very few generations form a complete

No.594] BLENDING AND MENDELIAN INHERITANCE 8333

gradation of forms connecting one parental mode with the

other. The most numerous of these intermediate forms

will in F, and later generations be that which is midway

between the modes of the respective parents.

This case throws a flood of light on the nature of blend-

ing and of Mendelian inheritance and of their relations to

each other. In typical Mendelian inheritance determiners

of allelomorphic characters may meet each other genera-

tion after generation in a common zygote, separating ©

again in gametogenesis without apparent modification of

either in consequence of their conjugation in a hetero-

zygote. This is well illustrated in the color inheritance

of animals and plants.

In typical blending inheritance the determiners of con-

trasted parental conditions apparently blend into a deter-

miner of intermediate character, the gametes formed by

an F, individual being practically as uniform in char-

acter as those of either parent individual. Blending is

illustrated in the inheritance of ordinary size differences

in birds and mammals (Castle, 1916).

A third type of inheritance must now be recognized

which is a compromise between these two, for it exhibits

Mendelian segregation of the contrasted parental condi-

tions but with modification due to partial blending of the

unlike determiners in the F, zygote. The blending in-

creases and evidences of segregation decrease with every

generation during which the contrasted characters re-

main in conjugation. Consequently with every genera-

tion of inbreeding or self-fertilization following a cross

of this sort, a stable intermediate class is more and more

closely approached until its realization is complete. See

Marshall (1916) on the Corriedale breed of sheep. The

existence of this third type of inheritance was pointed out

by Castle and Forbes (1906) in.the case of hair-length in

guinea-pigs and by Castle (1906) in the case of polydactyl-

ism in guinea-pigs. The opinion was then expressed that

‘more characters fall in this category than in any other.”’

Hoshino’s observations on flowering time in peas (if I

334 THE AMERICAN NATURALIST [ Vou. L

rightly interpret them) fully establish the existence of

this third type of inheritance. Incidentally they indicate

that the Nilsson-Ehle principle of multiple. determiners

to explain blending inheritance is quite superfluous.

PAPERS CITED

Castle, W. E.

1906. The Origin of a ii ae Wad, Race of Guinea-pigs. Publ.

N of Washington

1906. Heredity of Hair-length in Guinea-pigs and Its Bearing on

h eory of Pure Gametes. Ibid.

Hoshino, Yuzo.

1915. On a Inheritance of the Flowering Time in Peas and Rice.

Jour. Col. ohoku Imp. Univ., Sapporo, Japan, 6, Pt. 9,

pp. poe 88

Marshall, F. R.

1916. Corriedale Sheep. Journ. of Heredity, 7, pp. 88-95.

THE OCCURRENCE OF THREE RECOGNIZED

COLOR MUTATIONS IN MICE

Dr. C. C. LITTLE

HARVARD MEDICAL SCHOOL

THE common wild mouse, Mus musculus, has a type of —

coat-color pattern well known to geneticists as being

characteristic of almost all wild species of rodents. This

pattern is commonly called the ‘‘agouti’’ pattern. It

consists in a ‘‘ticking’’ or ‘‘banding”’ of the hairs of both

dorsal and ventral surfaces. On the dorsal surface each

hair has a subapical band containing yellow pigment.

The tip of the hair is black pigmented, while the half

proximal to the yellow band is dark, containing both

black and brown pigment granules. The general effect

of the dorsal surface, when considered as a whole, is dull,

brownish gray. The ventral surface is distinctly duller

and paler than the dorsal. The distal half of the hair is

lightly pigmented with black and occasionally with some

yellow pigment, while the proximal half is much more

heavily pigmented with black and brown granules. The

general impression, conveyed by the ventral surface, is

dull faded gray. This fact has led to the adoption by

various investigators of the term ‘‘gray-bellied agouti’’

to describe the wild mouse color pattern.

In addition to this gray-bellied agouti another type of

agouti house mouse has been reported as occurring wild

in certain localities. This type has been used in genetic

investigations by Cuénot and by Morgan.

The chief difference between the gray-bellied agouti and

the aberrant type is that the latter has, generally speaking,

more yellow and brown and less black pigment. This is

especially true of a band of hairs which runs laterally along

1 Morgan, noticing that Cuénot’s mice of this variety had a small patch

of reddish brown hair between the front legs while his own mice had not,

has suggested the possibility that the variety with which he worked was

not identical with that described by Cuénot. I have at present animals of

this color variety with and without the spot and therefore hope soon to

have a definite answer to this question:

336 THE AMERICAN NATURALIST [Vou. L

the boundary of the lighter ventral hairs between the front

and hind legs. This streak is very reddish in color and

adds to the lighter and warmer coloring of the dorsal sur-

face. Ventrally, the tips of the hairs are to all extents

and purposes unpigmented and this produces a white or

_ nearly white ventral surface. The proximal half of the

hair is dark pigmented, a distinction between this type

of white area and that found in spotted animals where the

hair is white throughout its length. This type of agouti

has been given the name of white-bellied agouti (gris a _

ventre blanc; Cuénot).

In crosses with gray-bellied agouti the white-bellied

form is epistatic. It can be obtained in a homozygous

condition and is, as stated by Cuénot and by Morgan, one

of a series of four allelomorphic forms of coat pigmen-

tation.

It is the appearance of this epistatic ‘‘white-bellied’’

agouti in two experiments involving gray-bellied agouti

and non-agouti mice that I wish to record.

| EXPERIMENT A

In 1913, a cross was made between wild gray-bellied

agouti mice and a race of dilute brown mice which I had

started in the laboratory of the Bussey Institution.

The agouti race used was directly descended from wild

animals without any out crossing with any domesticated

mice. It therefore represented a stock of wild mice

raised for several generations in captivity.

From this stock of agoutis I have recorded 531 young.

All of them were gray-bellied agoutis? of the type that

one might catch wild in any house. A certain amount of

variation in intensity of pigmentation occurred among

these mice. This variation, however, was not different

so far as could be observed from that shown by a consider-

able number of specimens of this species from any single

locality, judging from skins in the museum of comparative

— of Harvard University. Among these skins, and

e brown agoutis recorded later in this paper are of the dark or

ein perches type.

No. 594] COLOR MUTATIONS IN MICE 337

among the mice raised from the gray-bellied agouti stock,

there was not even an approach to the white-bellied agouti

type.

The dilute brown animals which were used belonged to

a race descended from a single pair of dilute brown mice

which had been tested by suitable matings in order to

establish their homozygosity. These animals in common

with other ‘‘non-agouti’’ varieties of mice lack all visible

traces of the agouti pattern and are with the exception

of a slightly lighter ventral than dorsal surface uniformly

pigmented throughout.

The cross between these two races, then, was one be-

tween a gray-bellied agouti and a ‘‘non-agouti’’ race.

Both races used as parents were therefore color varieties

which were hypostatic to the white-bellied agouti type

and which for this reason could not carry that pattern as

a recessive.

Four females of the dilute brown race were crossed

with a gray-bellied agouti male 131 and produced 26 gray-

bellied agouti young as shown in Table I.

TABLE I

Gray-BELLIED AGOUTI

Mating

DES SABE eS ee ee weno s eee 6

OBR S38T E ce vee kd ce ee 8

ODS So E E S A EE E PE dunes vec enon 6

PA Se PIB i Se we E E Phi :

The F, gray-bellied agoutis were more intensely pig-

mented than their gray-bellied agouti parents. This

deepening of color has frequently been observed in cross-

ing wild agouti varieties with tame races and is probably

due to modifying factors introduced by the particular

tame race used or else to a general acceleration of pig-

ment production due to heterozygosis.

The F, animals were crossed inter se and 285 F, young

were recto Of these 14 died at too early an age to have

their color recorded.

338 THE AMERICAN NATURALIST [Vou. L

In an earlier paper (1913), I have reported a cross sim-

ilar to the one here recorded and have shown that three

pairs of alternative mendelizing factors are involved.

These are:

A—gray-bellied agouti. a—non-agouti.

B—black pigment. b—no black pigment.

D—intense pigmentation. d—dilute pigmentation.

The gray-bellied agouti parent was homozygous for

the factors for gray-bellied agouti, black and intensity

being AABBDD in zygotic formula. The dilute brown

parent was homozygous for the hypostatic conditions,

non-agouti, no black pigment, and dilute pigmentation,

and was, therefore, aabbdd in formula.

The F, gray-bellied agoutis would of course be hetero-

zygous for all three pairs of factors and would be AaBbDd

in formula. When such F, hybrids are crossed together,

eight color classes of young are expected in a 27, 9, 9, 9,

3, 3, 3, 1 ratio. As will be seen from Table II, this result

was approximated with the addition of a ninth and un-

expected color variety, the white-bellied agouti.

TABLE II

Fz Generation

3> lga | (3 | | S181 8

on 33083 a egia el alee:

“2 us a BSis3| gi 2184) 2

35 |38 B äjä ia

a” ie A AIA |A

BOBS KIB Se S E NA 4 | —|— j2 1 2 | —| 1 |{—]i1

BHX. o a —|— |-| 1 |—]| 1 2 |— 2

XW.. es oe —| 6 3 3 1 I i—i rle

PIXI ona 4 |; —|— |4 a e r

E yo Fen ee a 2 }—|;—}|—}—|}—}—;—)}-| 5

POUROU Foo eee. 15 12 4 3 3 3 2 | 1 ==

PMOL E R A a 10 |—| 4 8 4 —| 3 2 1 |=

H TB a 3 |—| 4 1 2 | — uf — | a

E OXIS or ee 16 '—| 3 3 3 2 3 40) — i

Posh o a a 9 —| 3 3 2 | — — | — | oI

MOU MIO A oe 24 — 6 4 5 J 2 2 | 2|—

GII ooa an 15 |—| 3 3 4 1 — | — | |

Moxa o o 12 1), 4 2 3 1 2 Li e

Obtained numbers............. 124 | 2 3235 37 32- In i1 4 |5414

Expected on three factor basis. . .| 116.1| — | 38.7|38.7| 38.7| 12.9! 12.9| 12.9) | 4.3) —

No. 594] COLOR MUTATIONS IN MICE 339

It will be noticed that the approximation to the expect-

ancy on the basis of three pairs of Mendelian factors

is extremely close. The occurrence of the two ‘‘white-

bellied’’ agouti young is the matter of especial interest.

These two young, one a male the other a female, are ap-

parently identical with the gris à ventre blanc of Cuénot,

judging from his description.

The question may quite naturally arise as to the possi-

bility that the white-bellied agouti animals in F, had arisen

by an accidental cross with some white-bellied agouti

male in an adjoining cage. This possibility, however, is

obviated by the fact that the F, generation white-bellied

agoutis appeared in different cages at the Harvard Med-

ical School where there had been no other white-bellied

agoutis for more than a year before the appearance of the

mutant animals. The two F, white-bellied agoutis must

then be considered as true mutants epistatic to the type

from which they originated.

Further evidence is provided by the F, generation, a

tabulation of which is given below.

It will be observed that there is one ‘‘white-bellied’’

agouti among the 624 animals comprising the F, genera-

tion. This one mutant occurred as a descendant of a

family of F, animals shown in Table II, mating J. This

mating in F, gave 9 black agouti and 3 brown agouti

young, none of which were ‘‘white-bellied’’ and the mu-

tation therefore originated in a gamete formed by one of

the F, generation animals while the other two white-bel-

lied agoutis originated as mutations in the gametes of F,

animals.

This recurrence of the mutation is a matter of consider-

able interest and indicates a germinal change, the occur-

rence of which is largely irregular, having some under- —

lying cause or causes which become acpi in animals `

of this particular hybrid race.

That hybridization between the same wild agouti and

dilute brown races does not always produce this or any

other recognizable mutation from the eight expected color

340 THE AMERICAN NATURALIST [ Vou. L

TABLE III

Fs Generation

s.\s g | lg z

3 x FEE EA c| Alil ER

2 3 <7| <3 S| </As) maal a | $2| ga

| E S | |

5 z 4202/7 | EE EEE

a g ao | a KIA] Ala | al al A

M 164 X161..... a 2 —|/—|/—|—|-|—-—

169101... 3i—| 3} —I|—| 1/—/|—|—|-|-

73x76... ae a 90) a oe

163-164 X161..... 4|/—| iuj 2}—|—}|—-|-| >

L 331 X335..... —|—{—}| 2/—}—}—}—|—}-—| —

TRIOS. 222 w= 10} 5] 2} 2}/—; 1/—j;-|—

165X168... .: 2;/—| 1;—/—}]—| 1)/—!—|—|=—

89X86... Tope a

BA X86.5..54 16; — — OD ba, be Kaen a ree

neat PRES 30; — |—|}—}| s3/—|—!—}|—!|—-| =

165-167 x163 eee Ti —| 6) 92) 2) 1)/— ayers =

K 280 X320..... tii i=l

ixi... ai- 2) 2l 2l r

BT EG. eues Si 2?) he be 8 hee

BB X56... 3i H llaa — PR

LDE. a ala el 2}—] aam paa

51-57 X56...... ee | A Ge Ces oe oe ee

J 145 X142..... hee Pt Se Le eS Bp ee

144 X142..... silile ei

oxo o 10 | — |17| —| 2} 1| 3; — |=|- |>

90 X93...... —|—|35|—|—| 8} 6|/—}| 5|—};—

144-145 X142 s] =l 3) —i— 1 1 —

F 136X137... 9 —| 1} 3|— i ae a —|—

136-140 X137..... 6—=] 121i itani d l

G at es sar e Capel a e SA ea oc age E

aoo TOX poe be Blt 8] 81S) 2S) ae

102-103 x 101-106. 3;—| 1;—|—|—/| 4—32

I | 172-174X171.....| 16 | — | — | 10 | 17 | —|—| 14) —|—|—

BOXI- ou. eit Ge ei ah, E a 8 a

70x18... ao Fie bt a ee

79-80-82 X78...... 16 | — | 20| — | — ee 4|}—|—|—|—

79-80 X78...... orgy Oh 2 ee ed ee ee

200 | 1 |137| 93 |72 |35 | 40|32| 9| 31 2

varieties shown in Table II, is proved by the following

experiments.

In 1911 a parallel cross was made between wild agouti

and dilute brown races. The making of this cross I have

recorded (1913; p. 55, Cross 20a; p. 52, Cross 10a, ete.) in

a previous paper. The first hybrid generation consisted

of gray-bellied black agouti animals, 8 in number. The

second hybrid generation consisted of 55 animals which

were of the expected color varieties. In the third hybrid

and ensuing generations, including back crosses, I have

recorded more than 4,500 young. None of these animals

No. 594] COLOR MUTATIONS IN MICE Sate. 3

were white-bellied agoutis or indeed of any but the ex-

pected eight color varieties. This proves conclusively

that the hybridization of these two races does not neces-

sarily cause the white-bellied mutation to appear.

The fact that the white-bellied mutation has occurred

in wild races of Mus musculus as shown by the capture of

white-bellied mice would indicate that this mutation does

not necessarily depend for its origin upon the presence in

the germ cell of any of the recognized color factors in a

heterozygous condition.

The test by breeding of the white-bellied agoutis ob-

tained in the F, and F, generations has been attempted.

One of F, generation mutants has not yet been bred, but

the other F, and F, white-bellied agoutis have shown that

their color pattern was epistatic to gray-bellied agouti

and non-agouti. In these respects, the mutants have be-

haved in a manner identical with the gris 4 ventre blanc

mice of Cuénot or the white-bellied agoutis of Morgan.

The observed facts therefore prove clearly the origin

of this epistatic white-bellied agouti mutation at three

different times in the course of this experiment.

It is interesting to note that there has been no selection

in the direction in which the mutation occurred as was

the case in the experiments of Castle and Phillips (1914)

on hooded rats.

Although the white-bellied agouti may easily be con-

sidered merely an increased state of activity of the agouti

factor beyond the gray-bellied stage, it is certain that it

arises, not as the product of a series of small gradual

changes, but suddenly and distinctly, without warning, and

that after its appearance it behaves at once as a mendeliz-

ing character.

Taking into consideration the facts now recorded con-

cerning this particular mutation we can say that its origin

is apparently not solely dependent upon any of the known

genetic processes. Inbreeding, hybridization, selection,

none of them is indispensable to the occurrence of the

mutation. The underlying cause or causes are at present

clearly outside the bounds of analysis.

342 THE AMERICAN NATURALIST [Vou. L

EXPERIMENT Ð

In 1912 I started, on a small scale, an experiment which

had for its object the modification of the agouti pattern

by repeated crossing with black (non-agouti) animals.

It was thought that if the gametes of agouti animals

showed quantitative variation in the factor underlying

the agouti pattern repeated crossing with non-agoutis

would succeed in decreasing the amount of the agouti

character. It follows that if this was accomplished by

contamination between the agouti gamete and the non-

agouti (black) gamete, the non-agouti animals should

show increasing traces of the agouti character as the

agouti animals showed a diminution of that character.

The method followed was to cross black agouti animals

derived from a sooty yellow stock with blacks from the

same stock and selecting the blackest agoutis from each

litter repeat the process. All the agoutis used were thus

heterozygous for the agouti factor. This process was re-

_ peated for more than seven generations during which

approximately 400 young were recorded. From the out-

set two facts were evident. First, the agoutis grew dis-

tinctly blacker: second, there was no corresponding sign

of contamination on the part of the non-agouti animals,

but they too grew blacker. It appears, therefore, that

there was no sign of modification of the factor underly-

ing the agouti pattern; but that a race was isolated by

selection which showed a distinct increase in depth of pig-

mentation. In this connection it is interesting to note that

while the agouti animals grew blacker and the yellow

areas in the hair decreased in extent, the yellow pigment

itself was a deeper richer yellow than that of the ordi-

nary black agoutis. The yellow pigment had been in-

creased in depth while the black pigment was being in-

creased in depth and in extent, two entirely distinct

processes.

All the agoutis up to the third generation were gray-

bellied agoutis usually having more dark pigment on the

ventral surface than is found in wild mice. In the third

generation there occurred an agouti with a distinctly yel-

No. 594] COLOR MUTATIONS IN MICE 343

lowish tinge to the under surface. The dorsal appear-

ance of this animal was apparently the same as the other

agoutis of this generation, a deep rich agouti with much

black pigment. The ventral surface was, however, dis-

tinctly lighter in color than the other agoutis and showed

a decrease in dark pigment as compared to them.

The yellow-bellied variation reappeared in the imme-

diate progeny of the original yellow-bellied female and

a race of these animals was established. When single

mating tests were made to determine its behavior it was

found that when yellow-bellied agoutis one of whose par-

ents was a non-agouti were crossed with non-agouti, only

yellow-bellied agouti and non-agouti young resulted.

This fact indicated that yellow-bellied agouti fell some-

where in the series of multiple allelomorphs recorded by

Cuénot and by Morgan.

At about this time English ‘‘black and tan’’ mice?

crossed with white-bellied agoutis were found to give

among their progeny agouti young almost identical with

the yellow-bellied agoutis described above.

It at once suggested itself that the yellow-bellied

agoutis occurring in my selection experiment were really

white-bellied agoutis with one or more modifying factors

which encouraged a higher degree of pigmentation than

is normally found. This increase in an oxidation process

would account for yellow pigment appearing in the tips

of the ventral hairs which in ordinary white-bellied

agoutis are unpigmented.

This supposition was further favored by the fact that

as certain of the yellow-bellied agoutis grew to be ol

mice they showed a diminishing depth of pigmentation,

and developed typical white-bellied agouti coat color. In

old age they were not visibly different from some of the

darkest white-bellied agoutis descended from a male of

this variety kindly sent me by Professor Morgan.

8 Black and tan mice are a very dark type of yellow. Only a small amount

of yellow appears on the sides, while the ventral surface is a deep rich

reddish yellow. Black animals descended from black and tans are coal

black with deep black ears, feet and tail.

344 THE AMERICAN NATURALIST [Von L

One breeding experiment is interesting in showing the

probable identical nature of the factor underlying the

yellow-bellied and white-bellied agouti types. Two yel-

low-bellied agouti females known to carry non-agouti

were mated with a white-bellied agouti male known also

to carry non-agouti. On the supposition that the white-

bellied and yellow-bellied mutations are the same, this mat-

ing should produce two types of young, agouti (either

yellow or white-bellied, not both) and non-agouti in a 3:1

ratio. If on the other hand yellow-bellied agouti depended

on a distinct factor falling in the multiple allelomorph

series hypostatic to white-bellied agouti, the mating

should produce three types of young, white-bellied agouti,

yellow-bellied agouti, and non-agouti in a 2:1:1 ratio.

Finally, if yellow-bellied agouti depended upon an inde-

pendent modifying factor which was acting upon a gray-

bellied agouti animal, one of two results should be ob-

tained. If the factor was epistatic we should have white-

bellied agoutis (plus the modifier), yellow-bellied agoutis

and non-agoutis in a 2:1:1 ratio. If the modifier was

_ hypostatic in the cross, white-bellied agouti, gray-bellied

agouti and non-agouti young should occur in a 2:1:1

ratio. The three litters obtained have totaled 19 young;

11 white-bellied agouti and 8 non-agouti. The two classes

of young seem to indicate strongly the correctness of the

view that yellow-bellied agoutis are in reality white-bel-

lied agoutis plus a darkening modifying factor or group

of darkening modifiers the nature of which is not yet

sufficiently clear to allow further description.

The point of extreme interest is that the white-bellied

agouti mutation has occurred in a race which was being

selected in directly the opposite direction from that taken

by the variation. White-bellied agoutis represent a

stronger agouti factor than the gray-bellied agouti factor,

while every effort was being made in this experiment to

weaken the gray-bellied agouti factor. It appears that

the occurrence of the white-bellied mutation in experi- -

ment A, above recorded where no selection was being

No. 594] COLOR MUTATIONS IN MICE 345

exercised, and the independent oceurrence of the same

mutation in experiment B against the course of selection

are evidences that’ the direction of mutation is largely if

not entirely independent of selection and that the occur-

rence of the plus mutant rat in the plus selection series of

Castle and Phillips before referred to, is in all probability

a matter of coincidence rather than the result of selection

as they have hinted.

PINK-EYED MUTATION

It will be noticed in the table showing the F, generation

of the previous experiment (A) that among the 624 young

recorded, 5 are pink-eyed. The pink-eyed mutation is

hypostatie to dark-eye color and has been known for

some time (see especially Castle and Little, 1909, Dur-

ham, 1911). These young all appeared in a single pen

in which were two females and two males all of a single

litter. Hight dark-eyed and five pink-eyed young were

produced by these mice.

Up to the appearance of the pink-eyed young it was not

_ suspected that the pink-eye factor in any way entered into

the experiment. It was certain that one parent stock

consisted of pure wild mice and that the pink-eye factor,

which is a recessive, was not brought into the cross by

this parent.

The pink-eye mutation, being recessive, could easily

have been latent in the cross for some time without the

combination of gametes necessary for its manifestation in

a zygote having been realized. The dilute brown animals

from which the second parent race was descended had

originally been tested by breeding and had been found to

be entirely free from the recessive pink-eye factor. Since,

however, this mutation appeared in only one F, family

it seemed distinctly unlikely that the original wild male,

131, which had been used as the male parent of all fam-

ilies, was the animal through whose gametes the mutation

came into the cross. On the other hand, as there Pe

four dilute brown females (Nos. bl, b2, b3, b4,) used in

346 THE AMERICAN NATURALIST (Vou. L

producing the F, generation, it seemed quite likely that

one of these first introduced the pink-eye factor.

If such was the case the animal in question would be

female No. b2. This mouse was the dilute brown ancestor

of the particular group of F, animals which in turn were

the parents of the pink-eyed mutant individuals.

The suggestion that the dilute brown parent is the ani-

mal which introduced the pink-eye factor is supported by

the fact that at about the same time at which the mutation

appeared in the F, generation of this experiment it also

appeared within the ‘‘pure’’ dilute brown race. This

makes it extremely probable that the mutation had already

- occurred within the dilute brown race and was brought

into the cross by a single dilute brown female which be-

cause of the fact that it was dark-eyed concealed the pres-

ence of the recessive pink-eyed factor which in all prob-

ability existed in approximately half its gametes.

Unfortunately at the time that this mutation appeared

in F, the dilute brown great grandparent had died. A

breeding test was therefore impossible in order to ascer-

tain whether she actually carried the pink-eye factor in —

one half of her gametes.

EXPERIMENT C

Another mutation, this time of the black producing fac-

tor, has occurred in a stock of pure wild mice, the original

individuals of which were caught either in Wenham,

Mass., or in Forest Hills, Mass., in 1912.

The particular family in which this mutation occurred

is shown in Table IV. As will be seen, two wild mice

both caught at Forest Hills, Mass., were bred together,

these are female 1 and male 4. They and all their de-

scendants, unless especial mention of the fact is made,

will be considered black agouti in color. That is to say

they were in appearance ordinary wild house mice. Fe-

male 1 and male 4 gave among their progeny male 52,

who was crossed back to his mother and thus gave rise to

female 90. She in turn was crossed back to her father,

male 52, and from this mating male 131 was obtained.

No. 594] COLOR MUTATIONS IN MICE 347

TABLE IV

ANCESTRY OF BROWN-AGOUTI MUTANTS

91 X of

op Pure Wild Wenham Stock.

Q g5

Pal7 Q z15 Brother and sister mating. 1st generation.

423 Geli oF ne

g g34 Q g38 " aE ae $ 3d

g g55 RROD E T A ee 4th

g 77, 76, 78

8 Black agouti.

3 Brown Agouti.

Male 131 was used in many crosses, one of which was

with female g5. This female was a pure wild mouse

taken from a stock pen of wild mice which for several

generations had been reared in captivity by Dr. J. C. Phil-

lips at Wenham, Mass. From this mating came 6 black

agouti young in a single litter. Two of these young, fe-

male g15 and male g17, brother and sister, were mated

together and produced 11 black agouti young, two of

which, male g23 and female g21, were mated together.

This brother and sister mating gave 16 black agouti

young from which another brother and sister mating,

male g34 and female g38, was made. From this pair was

obtained 13 black agouti young, among which were male

g55 and female 258, the parents of the mutants. These

two mice male g55 and female g58 have had three litters

of young. The first born in December, 1915, consisted of

three animals, male g77 Brown Agouti, female g78 Brown

Agouti and female g76 Black Agouti. The appearance

of the brown agoutis was entirely unexpected and was

thought to be possibly due to an error in records. A sec-

348 THE AMERICAN NATURALIST [Vón L

ond litter however, born in February 1916 after the par-

ents had been under careful observation, consisted of two

black agoutis, a male and a female, and one brown agouti

afemale. The third litter of five is all black agouti.

This is a clear case of mutation within a closely inbred

race, and is interesting to contrast with Experiment A

already referred to, in which a mutation occurred in

hybrids.

There is one fact of possible interest in connection with

the mutations recorded in Experiment A and in this ex-

periment. Male 131, black agouti, is a common ancestor

of all the races in which the mutations occurred. It has

been shown that the evidence is against his having intro-

duced the pink-eyed mutation and that this probably came

from the dilute brown race.

For the other two mutations, however, the white-bellied

agouti and the brown agouti types, it is theoretically pos-

sible that male 131 possessed or transmitted an instability

of germplasm which has manifested itself in the crop-

ping out of these mutations among his descendants. For-

tunately the stock within which the brown agouti muta-

tion arose is being carried on in single pair, brother and

sister, matings. By this method we should be able to

recognize mutations at the earliest possible moment after

their occurrence.

SUMMARY

To sum up the facts above recorded it may be stated

that:

1. A previously recorded mutation of the gray-bellied

agouti pattern, known as white-bellied agouti, has arisen

in two experiments on color inheritance in mice.

2. In experiment A it has arisen independently three

times in a hybrid race of mice.

3. In this experiment there has been no selection in the :

direction of the mutation.

4. In experiment B it has arisen once in an inbred race

in which selection was being carried on.

5. In this race the mutation represents a variation in

No. 594] COLOR MUTATIONS IN MICE 349

exactly the opposite direction from that in which the

selection was being made.

6. A recessive pink-eyed mutation has occurred in a

closely inbred dilute brown race. A similar mutation

has appeared in a hybrid race into which one animal from

the dilute brown race probably introduced the mutation.

7. A mutation involving loss or suppression of the

black producing factor has arisen in a stock of inbred

wild mice. This has caused the appearance of brown

agouti young.

. The wild race in which this occurred is related to the

hybrid race (see conclusion 2) in which the white-bellied

agouti mutation appeared three times. The suggestion

is offered that a tendency to germinal instability may

have been transmitted by male 131 a common ancestor of

both races.

LITERATURE

Castle, W. E., and Little, C. ¢.

1909, ii N. S., Vol. 30, pp. 813-14.

Castle, W. E., and Phillips, J. C.

1914. "Publ. Carnegie Inst. of Wash., No. 195.

1907. Arch. de Zool. Exp. et Gén., Notes et revue (4), Vol. 5, page 1.

1908. ae de Zool. Exp. et Gén., Notes et revue (4), Vol. 6, pp.

T-

1911. Areh, de Zool. Exp. et Gén., Notes et revue (4), Vol. 8, pp.

Durham, F. $a

1911. Journal of Genetics, Vol. 1, pp. 159-178.

Little, C. C.

1913, Publ. Carnegie Inst. of Wash., No. 179.

Morgan, T. H.

1908. Science, N. S., Vol. 27, Ae 493.

1909. Am. Nart., Vol. = sds

1911, Annals, N. Y. Acad. of yg Vol. 21, pp. 87-117.

1914. Am. Nart., Vol. i, pp. 449—458.

THE MECHANISM OF CROSSING-OVER III

HERMANN J. MULLER

CoLUMBIA UNIVERSITY

V. Aw EXPERIMENT To DETERMINE THE LINKAGE oF Many

Factors SIMULTANEOUSLY

A more exact knowledge of the interference of one

crossing-over with another required an experiment, or

series of experiments, in which the distance between the

two points of crossing-over in cases of double crossing-

over could be more accurately determined. In an experi-

ment involving only three factors—A, D and H—if a

double cross-over occurs, all that can be known is that

crossing-over has occurred at the same time somewhere be-

tween A and D, and somewhere between D and H, but

nothing can be known of the precise location and distance

apart of the two points of crossing-over, except that they

could not be further apart than A and H. On the other

hand, if the inheritance of four points could be followed—

say, A, D, F and H—then the distance between the two

points of crossing-over could be determined a little more

exactly, for a double crossing-over involving breaks be-

tween A and D, and between D and F, would cut out a

shorter segment of the chromosome than one occurring in

regions A-D and F-H. And the more numerous were

the factors that could be followed—other things being

equal—the more exact the determination would become.

At the same time, it might be possible by comparing the

results of a series of different experiments to arrive at the

desired end with the three-factor method also. For ex-

ample, the difference in frequency between the double

cross-overs obtained in an experiment involving A, B, C

and in an experiment involving A, B, D, must obviously

be due to the double cross-overs involving regions A-B

and C-D,* except in so far as these differences are due to

the random deviation of different samples from each

1 For region BD is made up of BC+ CD. Therefore double crossovers

involving AB and BD really consist of double cross-overs involving AB and

BC plus those involving AB and CD. Consequently, if we subtract the

number of double cross-overs involving AB and BC from the number involv-

ing AB and BD, we obtain the number involving AB and CD.

350

No.594] THE MECHANISM OF CROSSING-OVER 351

other, or to actual differences in the behavior of the chro-

mosomes in the two experiments. As the last two influ-

ences seemed by no means negligible, and as the experi-

ment involving many points at once gave a more direct

and graphic picture of the results, it was decided to use

this method of attack in preference. Moreover such an

experiment incidentally afforded an opportunity of attack-

ing certain other questions, such as the effect, on crossing-

over, of having the two chromosomes different in regard

to many factors. Meanwhile, the indirect method of attack

would be followed by other workers, and the two sets of

results could finally be used as checks upon each other.

The many factor method is in itself for several reasons

very laborious, but this is compensated for by the fact that

when the results are obtained they are the equivalent of

an entire series of different experiments involving in turn

the linkage of each factor with every other one, and indeed,

the results are much more than the equivalent of these,

for in the latter cases the linkages are obtained in differ-

ent experiments, so that there is much more chance for

error in determining the relation of one linkage to another.

It was evident from the outset, however, that there was

one very important obstacle to be overcome in any study

of linkage exact enough to give useful information regard-

ing coincidence, and that the difficulty was especially great

in the type of experiment contemplated. The difficulty re-

ferred to is ‘‘differential viability,’’ for it is found that

in nearly all experiments not involving the characteristics

of seeds or other structures dependent upon the maternal

organism for support, the individuals belonging to differ-

ent genetic classes may be very differently equipped in

respect to their ability to meet the struggle for existence.

Thus, since the count generally takes note only of the indi-

viduals which survive, the ratios obtained may be very

different from the ratios of the different classes of gam-

etes. These discrepancies apply especially to forms like

flies, the larval life of which can not be well controlled, and

they are, of course, particularly great in crosses involving

many factors at once.

352 THE AMERICAN NATURALIST [Vou. L

Before considering the means by which differential

viability may be reduced in crosses of multiple stocks, it

may not be out of place to explain two methods I devised

for getting a more correct estimate of the gametic ratio in

back-crosses involving only two pairs of linked factors.

Let us say that the gametic ratio is r(AB):r(ab):

s(Ab):s(aB). Assume that when A is present the viabil-

ity of the flies is reduced so that only A’ per cent. of

those which would otherwise survive, now come to ma-

turity, and assume that factor B lessens the output to B’

per cent. of what it otherwise would be; similarly a and b,

when present, lower the output to a’ per cent. and b’ per

cent., respectively. Then the relative number of AB indi-

viduals which survive will be rA’B’ (per cent. marks are

omitted for brevity); the relative number of Ab will be

sA’b’, ete. The actual, observed, numbers will be some

multiple (k) of these relative numbers; thus the number

of AB individuals actually found will be krA’B’, the actual

number of Ab will be ksA’b’, ete. It can now be shown

that the true gametic ratio (r:s), which it was desired to

find, may be derived by the formula

AB X ab

Ab X Ab

(using Ab, ab, etc., to denote the number of AB observed,

of ab observed, etc.), for, substituting the above values of

AB, ab, etec., in this formula, we obtain

lkrA’B' X kra’b’ _ [Er A'B'a'b' o P

Nis AD X ksa B! Ni282A'b'a'B’ Ne

This formula should be used only when the smallest

class has not a very large probable error, for, by multiply-

ing the value of this class in the formula, -we give the

entire result a probable error proportional to that of the

smallest class. Another objection to the formula is that it

assumes that each factor produces the same specific lower-

ing of viability, independently of whatever other factor it

comes into combination with; this is not always true, since

factors often produce different effects when in different

combinations.

No. 594] THE MECHANISM OF CROSSING-OVER 353

The two difficulties encountered above are largely

avoided by the second method, which involves making two

different kinds of crosses in preparation for the linkage de-

termination: 7. e., cross AB by ab, and what may be termed

the ‘‘contrary cross,’’ Ab by aB. A back-cross of the F,

from the first cross gives the gametic ratio r (AB) :r (ab):

s(Ab):s(aB); and the other cross results in gametes

showing the proportion s(AB):s(ab):r(Ab):r(aB).

Suppose that w per cent. of AB individuals are viable,

x per cent. of ab, y per cent. of Ab, and z per cent. of aB.

Then in the first cross the observed ratio would be

rw(AB):rx(ab):sy(Ab):sz(aB), and, in the second

cross, s w (AB):sx(ab):ry (Ab):rz(aB). Thenumbers

actually observed in the crosses would be some multiple of

these ratios, but a different multiple in the two cases.

Thus we could designate the numbers actually observed in

the first cross as krw(AB) : krx(ab):ksy(Ab):ksz(aB),

and the numbers in the second cross as esw(AB):¢sx(ab) :

ery(Ab) :erz(aB). 7

In this case the ratio r: s may be obtained by the follow-

ing formula:

‘ABs x Abı

By X Ab;

(using the symbol AB, to denote number of AB observed

in the first cross, Ab, to denote number of Ab observed in

the second cross, ete.). Now, the value of AB, has already

been given as krw, of Ab, as cry, ete. Substituting these

values in the above formula, we obtain

rari krw X ery _ ra E

Nesw csw X ksy Nsckwy Na

Besides this formula involving AB and Ab, there are

three similar formulas which will also give the gametic

ratio, namely :

AB, X aB, abı X Ab. abı X aB:

AB, X aB,’ Nab, X Ab,’ ab, X aBı`

That formula should usually be chosen which contains the

largest number of individuals in its smallest class, for this

would usually have the least probable error.

354 THE AMERICAN NATURALIST (Von. L

This method makes no assumption as to an independent

action of the different factors in reducing viability. It

does assume, however, that for individuals with the same

combination of factors there is the same degree of via-

bility in the two experiments ; this is not always true, since

under different conditions of food, ete., individuals of the

same genetic type may have very different degrees of via-

bility ; moreover, there are sometimes ‘‘invisible’’ factors

present in one experiment but not in the other which influ-

ence viability and which are linked with the factors that

are being studied. The assumption, nevertheless, is un-

avoidable. But it can be shown mathematically that any

errors in the calculated values, due to assumptions made

in following the formulas of either the first or the second

method, are greatly reduced by using a combination of

the two methods; namely, by making ‘‘contrary crosses,’’

calculating the linkage value in each of them by means of

the first method, and then taking the square root of the

product of these two values.

In a cross involving three or more factors no formula

corresponding to the one first given is possible, and before

it is possible to use a formula corresponding to the second

method, an increasingly large number of different kinds

of crosses must be made, according to the number of fac-

tors involved. Still another method is, therefore, neces-

sary in order to obtain fairly accurate results from crosses

involving many factors, except in the rare case that these

factors have very little differential effect on viability.

The method devised is as follows:

The female, heterozygous for many factors, whose

gametic output it is desired to study, is back-crossed, not

to a multiple recessive male, but to one homozygous for all,

or nearly all, the dominant factors (these are, in the case

of flies, mostly the normal allelomorphs). All the off-

spring appear alike, then, in that they all show the domi-

nant characters of their father (except in the case of sex-

linked factors, which are transmitted by the father to his

daughters only), and so all should be of the same viability,

except for the insignificant effect of the recessive factors

No. 594] THE MECHANISM OF CROSSING-OVER 305

present in heterozygous condition (and the effect of the

one or two characters wherein the father may not have

been dominant). Thus error due to differential viability

may be held within safe bounds.

It may be objected, however, that we have, as it were,

killed the patient in curing the disease—that there is no

use in overcoming the discrepancies in the count due to

differential viability, if we thereby eliminate the possi-

bility of making any count at all, by making all the off-

spring appear alike! It is true that, in such an experi-

ment, it is impossible to tell by inspection of any offspring,

what maternal factors were present in the ova from which

they sprang, since these factors are made invisible, so to

speak, by the dominant factors brought in by the sperm.

But the factorial composition of each of these offspring

(which we will for convenience call ‘‘F',’’) can be deter-

mined by breeding tests. The plan which was followed

was to mate the F, flies, each in a separate bottle, to indi-

viduals containing the recessive factors. Thus whatever

recessive factors were present in the eggs of the original

heterozygous female (‘‘F',’’), whose output it was desired

to test, would become visible the generation after (in

“F,”). Whereas, in an ordinary linkage determination,

each bottle produces a large number of flies, which need

merely be classified according to their appearance, and

counted—in this case, each of the offspring themselves re-

quires to be mated and given a whole bottle to itself, and

its progeny in turn (‘‘F,’’) must be examined. In other

words, in ordinary cases, there is only one bottle necessary

for a count of many flies, but in this case one bottle repre-

sents one fly of the count. The numerical relations exist-

ing between the flies (‘‘F,’’) hatching in one of these final

testing-out bottles need not be determined, however ; that

is, these flies need not be counted; all that is necessary 1s a

‘‘qualitative’’ determination of what recessive characters

appear among them, in order to judge of the composition

of their parent (F,), which is the fly recorded in the count. —

Thus far 1008 of these test bottles have been recorded.

In preparation for this experiment the main task was to

356 THE AMERICAN NATURALIST [Vou. L

secure stock that contained many mutant, linked factors at

. the same time. But, as was explained in the account of

experiments with the third chromosome, it is necessary,

in dealing with linked factors, to make the crosses in a

particular way to secure a ‘‘multiple stock.” Thus, it

may be pointed out again here, a stock containing factors

A, B, C cannot be obtained ordinarily by crossing stock

A to stock C, and then crossing the double stock A C (pro-

duced in F., F,, or F, from the first cross) to stock B; be-

cause it would require double crossing-over for the hybrid

fly, containing A and C in one chromosome, and B in the

other, to produce a gamete with A, B and C all in the same

chromosome (assuming the factors to be linked in this

order). If the linkage is tight such double crossing-over

will never occur. But by first obtaining stock AB and

then crossing this to C, stock ABC may be secured; for

in the hybrid fly that contains A B in one chromosome and `

C in the other, a crossing-over between B and C will result

in a chromosome that contains A, B and C, the link be-

tween A and B not having been broken.

In other words, the factors can only be added together

in a certain order, owing to their position in the linkage

chain. Just as in adding links to a chain, one or more

factors cannot be wedged in between factors in another

collection (except by double crossing-over) ; but if they lie

beyond this collection, they may be added on, either singly

or in a group. The information that had already been

gained by Sturtevant, Morgan and Bridges concerning

the order in which various factors lay, was therefore of

great service in determining how the crosses should be

made, to get the factors together, and besides this several

double stocks of a sort that could be used in the present

experiment had already been synthesized by them. But

the progress of the experiment was very considerably re-

tarded by the fact that the position of a number of the

factors which it was desired to use had not yet been deter-

mined. These comprised bifid and forked in chromosome

I and dachs, jaunty, curved, are and balloon in II. (The

exact position of jaunty with respect to black, and of

No. 594] THE MECHANISM OF CROSSING-OVER 357

balloon with respect to speck is still unknown.) Various

‘‘trial and error’’ matings were therefore made in the

hope of getting these unplaced factors in suitable combina-

tions, and crosses were also undertaken to secure such

data in regard to their position as would be useful for the

purpose in view. These attempts were often cut short,

owing to the information which was meanwhile being

accumulated by the other workers, but before the latter

information was obtained the positions of bifid, forked and

dachs had been determined, and several multiple stocks

that were later used had been made up.

We may now consider specifically what combinations of

factors were actually employed in the experiment and

what special methods were used for securing and main-

taining these combinations.

In the case of the first chromosome, it was desired, for

the final linkage determination, that the heterozygous flies,

whose gametic output was to be tested, should contain

most of the recessive factors (which are usually the mu-

tant ones) in one chromosome, and the dominant (usually

normal) factors in the other, for it was considered worth

while to test the possibility that a chromosome containing

so many mutant factors might behave abnormally. Fur-

thermore, if the commonly accepted belief were correct,

that recessive factors are ‘‘absences,’’ it was possible that

the chromosome with many recessive factors might be

shorter than the other, and that linkage disturbances

might arise for this reason. Two stocks were, therefore,

made up, of the factors in the first chromosome, to supply

_these two kinds of chromosomes for the heterozygous

females to be tested. One stock contained the mutant

factors for yellow body color, white eyes, abnormal ab-

domen, bifid wings, vermilion eyes, miniature wings, sable

body color, rudimentary wings, and forked spines. All of

these factors are recessive except abnormal abdomen,

which is only partially and irregularly dominant. The

other stock contained only the mutant factors cherry eye,

club wing, and bar eye. Cherry is an allelomorph to the

factor for white eyes carried by the other chromosome,

358 THE AMERICAN NATURALIST [Vou. L

and is dominant to it, though not completely; club is

recessive ; bar is dominant (somewhat incompletely).

The reason that club was not put into the series with the

other recessives is that it was discovered (by Bridges)

after this series had already been put together, and so it

would have required taking the stock apart again, or else

obtaining a rare double cross-over, to wedge club into this

series. It was a valuable factor to have in the experiment,

however, since it lay in a region of the chromosome where

there were no other mutant factors to give data as to

crossing-over. Accordingly, it was inserted in the other

series. It will be observed, however, that, in spite of this,

one chromosome contains 7 more dominants than the other.

The order of the above factors is y, w or c, A, H, Co N,

m, s, T, f, Br. In making up the first stock the factors were

put together as follows (omitting from consideration all

trial or discarded combinations) :

yw A me oS a g

DA N NY

ywA b vm

S bori d

ywAb vmserf

ywAbvmsrf

Of course the putting together of factors from two stocks,

although shown above as only one step in each case, always

requires several generations. Moreover, as will be seen.

below, these steps do not usually consist in getting the

ordinary F, or back-cross, in the case of the complicated

combinations. This is partly because of the serious ob-

stacle which the poor viability of flies having many mu-

tant characters presents to the making up of multiple

stock, just as it does to the securing of counts from it;

moreover, in the making up of stock, the sterility of such

flies is an equally important difficulty.

These difficulties were overcome here in much the same

way as they were in making the counts—namely, by keep-

ing the stock, so far as possible, heterozygous. For ex-

2 From Morgan.

3 From Bridges.

No.594] THE MECHANISM OF CROSSING-OVER 359

ample, in the last step shown in the preceding diagram,

where factors ywAb and vmsrf are to be put together, it

was found that females of both of these kinds were ex-

tremely difficult to keep alive. It was, therefore, decided

to mate a vmsrf male by a female which contained ywAb

in one chromosome and normal factors in the other. Such

a female would be easy to breed from, as the normal fac-

tors dominate. About half the daughters (let us call them

F,) would be of composition meo (representing

the mutant factors in the maternally derived chromosome

on the upper line, those from the father on the lower line).

All the daughters (F,) would, however, appear normal,

but if these F, females were bred in separate bottles, those

; ENS bad | ee

of the desired composition ieee would be distin

guishable from the others by their offspring (F,). All

bottles in which the parents (F,) had not been of the de-

sired composition could then be discarded. Next, among

the offspring (F,) of those females which proved to be of

composition Aub ri it. was necessary to select the

ones which, by reason of crossing-over between b and v,

contained all nine factors in the same chromosome (7. e.,

ywAbvmsrf). But such individuals, if homozygous, never

live long enough to mate, so great is the lowering of via-

bility produced by all these mutant factors at once. Con-

sequently, some method must be used of obtaining in this

cross heterozygous individuals (F,) which received this

cross-over ‘‘nontuple’’? chromosome from their mother,

and of distinguishing these from other individuals pro-

duced by the cross. The natural suggestion would then

be that the F, females should be mated by normal males,

and the F, which receive this cross-over chromosome could

then be distinguished by breeding tests as their mother

had been. The crossing-over desired, however, does not

occur in more than one eighth of the flies, and so breeding

tests designed to be certain of securing at least one indi-

vidual of the required composition would have to be

rather extensive. In this case, however, the desired F,

360 THE AMERICAN NATURALIST [Vou. L

flies can be ‘‘spotted’’ in another way, without breeding

tests, and yet without making them homozygous for many

mutant factors and thus inviable. The method used was

to mate the F, females to bv males, which had been made

up for this special purpose. The bv daughters (F) must

be cross-overs, since in the F, mother b and v were in dif-

ferent chromosomes; moreover, a glance at the formula of

the F, females will show that these cross-over chromo-

somes must have been formed of the left-hand end of the

upper chromosome and the right-hand end of the lower.

Thus these bv females contain a chromosome with all nine

mutant factors (except in the case of the few double cross-

overs). Since, however, they were homozygous in only

two mutant factors, they could easily be bred.

A similar scheme was used in many of the other steps

shown in the diagram representing combinations made in

group I, and was also used frequently in group II. Owing

to the fact that rudimentary winged females (group I)

are practically sterile, devices of this sort had to be used

in dealing with flies containing this factor from the very

start, and the same may be said of flies with dachs legs

(group II), since these also were found very hard to

handle. In most of the other cases, however, it was not

necessary to use such a method before several factors had

been combined together, as flies homozygous for just two

or three mutant factors were generally viable enough to

handle. There would be no object in wearying the reader

with a description of the exact way in which each of the

steps was taken; it is the author’s purpose only to explain

the nature of methods used, giving only sufficient examples

to make clear the details of any devices never previously

employed that might be capable of application to other

cases.

From the example of the cross involving bv, previously

given, we may now generalize, and establish the rule that

in making up, and also in keeping stocks containing many

linked recessive factors, if the latter cause a marked less-

ening of fertility or viability, it is best to follow the prac-

tise of keeping the stocks heterozygous, by back-crossing

them to stocks containing only the few recessive (or par-

No.594] THE MECHANISM OF CROSSING-OVER 361

tially recessive) factors necessary to show which offspring

contain the desired cross-over or non-cross-over chromo-

some. In the example, this method was used in combining

two stocks to make up a recombination stock. The same

means is employed in maintaining the multiple stock after

it has been synthesized. Thus, in the case of group I, the

females containing in one chromosome the combination

ywAbvmsrf (the ‘‘F.’’ obtained above), were crossed to

eciB; males. In this way some daughters (F,) are pro-

duced (which these were was determined by breeding

tests) that received from their mother ywAbvmsrf, and

from their father cc,B;. These F, females having the com-

ywAbvmsrf

Ce; By’

males again, in order to maintain the stock. Since all the

daughters (F,) received cc:By from their father, those

which do not show these characters fully developed must

have received from their mother factors near both ends

of the chromosome containing the nine mutant factors.

Therefore, except for the very few flies in which crossing-

over occurred between w and y, which is at the very end,

or in which double crossing-over occurred, all the light

cherry, normal winged, partially bar eyed flies will have a

composition like that of their mother, and may be bred in

the same way, again to the cc,B; males, which now hatch

from the same bottle. This then is a cross exactly like the

preceding one, except for the few cross-over flies above

mentioned. The latter may be detected, however, and

their offspring discarded if the females are bred in sepa-

rate bottles. This same cycle may be repeated generation

after generation. Thus a continual supply is maintained

of flies heterozygous for all these factors.

In making the linkage determinations, such flies are bred

to normal or to bar males, and the female offspring, which

are all alike in appearance except in respect to the par-

tially dominant factors A and B», and which should, there-

fore, have had approximately equal chances for surviving,

are individually tested for their contained characters.

For the tests, the female need not be virgin, since, what-

ever kind of male is employed, the sons will show only

position were then back-crossed to ec, B:

362 THE AMERICAN NATURALIST [ Von. L

those sex-linked characters that their mother contained

and they may therefore be used to determine the composi-

tion of their mother. As a matter of fact, however, males

containing v were generally employed, so that v, if it had

heen present in the tested female, would appear in her

daughters as well as her sons. This additional test for v

was desirable because it is a factor which in a white,

cherry, or bar eye it is difficult or impossible to detect.

Stock of the second chromosome was obtained, and is

maintained, in an essentially similar way. Here the at-

tempt was not made to put most of the mutant factors into

one of the two chromosomes of the heterozygous females

to be tested. This was partly because an experiment of

this sort with one chromosome would seem sufficient.

Moreover, it was harder to make up multiple stocks of the

second chromosome, since the order of certain factors had

not at first been well determined, and since, besides, it takes

a greater number of generations to put non-sex-linked

factors together into the same stock than it does to put

sex-linked factors together. For, if two recessive stocks

of chromosomes II are crossed, the F, males, in which

crossing-over never occurs, transmit the recessive factors

of only one stock to each son and daughter. The latter

then can not be homozygous for both sets of recessive fac-

tors, and so it is impossible to pick out, except by further

breeding, those that received both sets from the mother.

But as in the case of the bv illustration given, if a male

with C D is available to cross with the ‘‘F,’’ hybrid female

ABC the FP indivi : :

AP” the F,” individuals showing both characters

C and D must have the composition a , and so the

desired cross-overs may be picked out immediately in F».

A somewhat similar scheme, often especially useful for

obtaining desired combinations in a non-sex-linked group,

involved making use of cross-overs that break combina-

tions already obtained. This too may be shown by an

example. It was desired to obtain a stock containing the

second chromosome factors dachs legs, jaunty wings,

curved wings, and balloon wings. Dachs black stock al-

No.594] THE MECHANISM OF CROSSING-OVER 363

ready had been made up, as had j cba. These two stocks

were crossed together, and the F, was back-crossed to

da bı. Inthe F, generation all the dachs that are not black

are cross-overs in the region between these two factors,

and so must contain, in the same chromosome with dachs,

instead of black, the factors j cv and ba, for black is in such

a position in the chromosome that a cross-over between

da and b, must nearly always be a cross-over between da

and j ¢y ba.

In the case of the second chromosome the individuals

tested for linkage contained, in one chromosome, the fac-

tors streaked thorax, black body color, purple eyes, ves-

tigial wings, are wings, and specked thorax, and in the

other chromosome, the factors dachs, jaunty, curved and

balloon. The order of these factors is as follows:

Str da bi pu Ve Cv Ar Sp ba. The way in which they were com-

bined is as follows:

Pe Bs ad big jet b

Pu Vg ar Sp da j Cy ba

bi Pu‘ Pu Vg âr Sp daj Cy ba

Set bi pu Vg Ar Sp

Str bi pu Vg ar Sp

As daj Cba is very inviable it is kept in heterozygous

condition by back-crossing, in each generation, normal

appearing males of the composition ele (called 4%

for short) to bi pu Ve ar Sp females (called 5). Since there

is no crossing-over in the male, all the offspring are either

apparently normal, 4%, or the homozygous quintuple reces-

sive, ‘“5.’? The same process can then be repeated in

every generation, by crossing the normal-appearing sons

to their recessive sisters. It is evident that the ‘‘5’’ fe-

males which are used need not be virgin, as they could

have been fertilized only by % or by 5 males. When %

males are crossed by 5 females which have been made up

so as to contain in addition the dominant factor streak

(these 5 females need only be heterozygous for streak ;

homozygous streaks are hard to handle), the daughters

4From Bridges. ; ene sue

5 From Morgan.

364 THE AMERICAN NATURALIST [Vou. L

which appear streaked but otherwise normal, must have

‘:_™ ‘These are the “F,”

¢ Pa Ve Ae s

females whose gametic output is to be tested. Accord-

ingly, they are crossed to normal males. All the offspring

(‘‘F.’’) appear normal (except for the dominant, streak),

but the factors they received from their mother may be

determined by mating them, individually, to % males,

for the latter contain (heterozygous) all recessive charac-

ters possible in the former.

It was at first thought that labor might be saved, and

certain points in addition determined, by conducting the

linkage determinations on flies heterozygous for the fac-

tors used in both chromosomes I and II at the same time,

instead of making determinations of the linkage in the

two chromosomes in separate experiments. The multiple

stocks of the two chromosomes were, therefore, crossed

together, and females were finally obtained that had the

composition:

Pa daj

the composition: Beh

ywAbvmsrf _ Str by Pu Ve ar Sp

ee, By dj Cy Da

These females, heterozygous for 22 mutant factors, were

then crossed to normal males, and the composition of their

female offspring was tested by mating these in separate

bottles to % males. The maintenance of the double-

multiple stocks proved to be extremely difficult, however,

and so, after obtaining determinations for 166 offspring

from such females, the two groups of mutant factors were

again separated. The data obtained in this part of the

experiment show that there is no linkage of any of the

twelve factors studied in group I with any of the ten

studied in group IT; this is of course in marked contrast

to the relations shown between factors in the same group.

The conclusions of previous workers that no factor in one

group was linked with any factor in another group were

based on results obtained with comparatively few com-

binations of factors, which were chosen as samples, so

to speak. It will be seen that in the present work these

conclusions have been confirmed by a study of 132 differ-

No. 594] THE MECHANISM OF CROSSING-OVER 365

ent combinations of factors in group I with factors in

group II.

CLASSIFICATION OF FACTOR COMBINATIONS TRANSMITTED BY FEMALES

ywAbvmsrf

HAVING THE COMPOSITION: p K

$J 1 r

| Yellows | Grays | Totals

Non-cross-overs

| ywAbomsrf 186 | é: gi. Be- 200 | 386

Between Single Cross-overs

VAR Wey Aacs aes | ye Cl Bic wAbumsrf 5 7

WBN Beh ua i ieee Fictie Byss cAbumsrf 5 8

Oa": 3 Re En ywA cı Bye 4 c bumsrf 11 15

Beane Gc on | Ra | ywAb cy Be N c vmsrf 27

CBN Woe ies ee ywAb B, 46 c ci vmsrf 51

VADAMI e eC | ywAbo Beg cc, msrf 9 16

m ADA Go issn ou oi Sees | ywAbom B, 18 aes ae. 3d 19 37

ROUGE Be aa ee. | ywAboms B, 28 ta Y S8 66

rand Fe Tas | ywAbumsr B, 0 cci J 5 5

Rm Bes a ET T EA | ywAbumsrf B, 0 CCl 1 1

Between Double Cross-overs

yandw;ciandv..... ye cwmsrf LACS OR eA ee: 1

y and w; m and s.. .... O e twee wåbom B, 1 1

yand w:sandr...... pope ee ae 1 wAbrms B, 1 2

y and w;randf......| ye 5 OE E tre an ae T 1

wand A;:candv..... | yw cwmsrf Bo ee pein a eae 1

and A; rand £27202; | yw Pe Sie a T 1

A gud bro and v.oe. Aa ee we ves p% B, 1 1

Aand b;sandr...... | ywA c Pe ee E be ee 1

band ci; mands...... | ywAbe; srf 1 co m B, 1 2

and cı; s andr. | Aba rf 4 c ms B, 3 T

advi Viandiin. 205.0) 2 es caot By 1 1

aand v;sandr...... | ywAb rf 7 ccyums B, 1 8

and v;randf...... | ywAb J D Wie Ghee a ee 2

e1 and v; f and B:..... | ywAb Lay ese. PA Gs oe Sa 1

Total Double and Single Crossing-over

Observ Observed | Per Cent. of

— Namber -| orveiag-e¥er Between Number ver

Vand wo. 13 2 Ny ADA Tings es 17 2

wand Aj. cf; 10 1.5 manda... sste 40

Aand b.. foe, 17 2 s TA ae 84 11.5

band on. a: 52 7.5 radi ine 9 1.2

ci and v 112 16. f and Br. 2 0.3

Not only the independence of the factors in the two

groups was shown by thi¢ experiment in which the two

groups were followed at once, but also the independence of-

the crossings-over. In the total of 166 cases, there were —

81 in which chromosome I underwent crossing-over (either

366 THE AMERICAN NATURALIST [Von. L

single or double), and 101 in which chromosome II crossed

over. If it was a matter of pure chance whether or not

erossings-over occurred in I and II at the same time,

coincident crossing-over should have happened in 166 xX

101 = 49 -+ cases. The actual number of cases in which

crossing-over occurred in both chromosomes at once was

52. Thus there is neither interference nor synchronism of

these crossings-over, and this result too is strikingly dis-

similar to the relations found between two crossings-over

in the same chromosome.

Since the results in the two chromosomes were found to

be independent in all respects, it is deemed unnecessary to

list here all the different combinations which were found

of factors in group I with factors in group II, and their

observed frequencies. The results for the two chromo-

somes may more advantageously be separated and added

to the other results, obtained when groups I and II were

followed in different crosses.

The data for the first chromosome are given in the tables

which follow. In all, 712 offspring of females heterozy-

gous for the 12 mutant factors in group I have been tested.

The above results give a direct demonstration of. the

fact that the factors behave as though they are joined in

a chain; when interchange takes place, the factors stick

together in sections according to their place in line and are

not interchanged singly. The fact is shown so patently

as to require no further comment.

Non-crossing-over occurs in this chromosome in 54.4 per

cent. of cases, single crossing-over in 41.7 per cent., double

crossing-over in 4.2 per cent. No triple crossing-over was

obtained in this count, although one, which will be de-.

scribed later, was obtained in the next generation, in one

of the ‘‘testing out” bottles.

(To be concluded)

SHORTER ARTICLES AND DISCUSSION

THE CAUSE OF THE BELIEF IN USE INHERITANCE

THIS note expresses an effort to view the old and recurring

problem of use inheritance from the aspect of the underlying

motives of thought involved instead of through a consideration

of the evidence directly bearing upon it.

The heredity of acquired traits is, theoretically, biological

heresy. But the interminable cropping out of the belief even in

professional circles indicates- a strong psychological impulse

toward the conviction. The mainspring of this impulse thus

becomes a matter of some importance to the student of heredity.

To begin with, it is well known that the lay public almost with-

out exception takes use inheritance for granted. Even evolution,

in the real mental workings of most educated but unprofessional

people, is more generally explained, unconsciously and in con-

crete cases, by appeal to the machinery of use inheritance than to

that of selection. The phrases struggle for existence and sur-

vival of the fittest have indeed evoked a wide popular response on

account of their picturesqueness, but their concepts are still but

little employed, even in the intelligent and studied folk mind, as

a real means of understanding or explaining evolution.

Those sporadic but in the aggregate numerous biologists who

adhere to the doctrine of use inheritance, revert to it, or evince

Symptoms of a leaning toward it, may be divided into two types.

The first class, probably because they think more penetratingly

than the average, long ago perceived the inadequacy of selection

alone as an explanation of organic evolution; and more lately

perceive also the insufficiency of selection with mutations and

Mendelian phenomena superadded. To students of this type, use

inheritance is therefore merely a last resort, a hypothesis on

which they fall back in default of any other to stop a logical gap.

The only methodological criticism that can be made of this school

is that it would undoubtedly be more stimulating of new dis-

covery if we were frankly to avow the limits of our knowledge

and leave certain things unexplained, than to complete the mental

‘structure of evolution by piecing in a principle which admittedly

rests only on contested facts and has opposed to it about as large

a body of evidence as can be assembled on behalf of any rat

and therefore logieay PTE proposition. —

367

368 THE AMERICAN NATURALIST [ Von. L

The second class consists of biologists and utilizers of biological

material whose keenness of thought is below the average. This

school introduces use inheritance into the conception of evolution

because it has failed to comprehend adequately the essential prob-

lems of evolution, and approaches them substantially in the atti-

tude of the layman.

The latter class is therefore merely unscientific and popular in

its thought processes; the former, having exhausted scientific

means and found them inadequate, returns, more or less frankly

in despair, to current folk opinion. The problem accordingly is

to discover the basis of the deeply rooted popular notion that is

involved in both cases.

_ While never formulated into a definite working principle until

Lamarck, because of the world’s lack of specific scientific interest

in organic phenomena, the principle of use inheritance has never-

theless been tacitly assumed by civilized nations of all periods,

and is taken as self-evident even by savages. It must therefore

rest on a large mass of common experience interpreted by an

elementary process of thought. Such an elementary process—in

fact the only elementary process of wide scope—is analogy.

The question then becomes what may be the basis—real enough

though unscientifically employed—for the analogizing that has

resulted in the conviction that use heredity exists. There must

evidently be a broad group of phenomena in human experience

that bear some resemblance to the hereditary transmission of the

acquired.

These phenomena are the exceedingly common ones of social

inheritance or cultural transmission and growth. We do “‘in-

herit’’ a name, or property, or knowledge of a language, or the

practice of an art, or belief in a particular form of religion.

Biologically such ‘‘inheritance’’ is of course absolutely distinct.

rom ‘‘heredity’’ because the mechanism of transmission is dif-

ferent. The source of social inheritance is not restricted to

parents and actual ancestors in the line of descent, but embraces

a multitude of individuals, consanguineous and unrelated, dead,

living, and sometimes even junior to the inheritors; in other

words, the totality of the social environment, past and present,

of an individual. We can and do ‘‘inherit’’ property from an

uncle, our ‘‘mother tongue’’ from a nurse, the arithmetic evolved

by past ages from a schoolmaster, our dogmas and philosophy |

from a prophet, our political and moral beliefs from the whole

cireumambient public opinion.

No.594] SHORTER ARTICLES AND DISCUSSION 369

As this social or cultural transmission concerns human beings,

it is of more immediate interest to the normal unschooled mind

than the transmission which gives organs, instincts and peculi-

‘arities to animals and plants. It is therefore recognized much

sooner than the processes which guide biological or organic trans-

mission. It needs no proof that in his development man was

concerned far earlier with himself than with animals or other

parts of nature. It is well known, for instance, that the animism

which is accepted as the basis of all religion, anthropomorphizes

not only its gods and the vaguer forces of nature, but especially

animals, plants and objects.

It is only recently, accordingly, that the world has paid any

true attention to organic heredity, whereas since the beginning

of human existence there has been recognition of social inher-

itance. History, the science of human society, is, even in a

relatively advanced form, several thousand years old, and as a

rudiment has enough interest to appeal to savages. Biology, the

science of the organic, has an age of barely two centuries.

It is significant that the first theory of organic evolution, that

of Lamarck, resorted wholly to the explanation of use inheritance

borrowed from social inheritance. A second stage was reached

when Darwin introduced the organie factor of selection, though

refusing to break with the older explanation. A last phase was

inaugurated when Weismann insisted that organic phenomena

must be interpreted solely by organic processes.

The priority of reasoning by analogy over reasoning by means

of a specifie mechanism is a world-wide historical phenomenon.

The two modern views of evolution and creation are found as

crude cosmic philosophies in the mythologies of the most primi-

tive savages, as well as in the thinking of Hindus, Semites,

Greeks, and Romans. But they occur, one as an analogy with the

familiar phenomenon of manufacture or making of objects by

hand, the other as an analogy with the equally familiar phenom-

ena of birth and growth. What modern science has done is to

adopt these age-old and erude ideas, as it has adopted the half-

mythologie concepts of the atom and ether, and put them to new

use. Only the uneducated think of Darwin as the originator of

the doctrine of evolution. What he originated was an organic

and in his day new mechanism, by which the old concept of evo-

lution could be explained and therefore supported.

The distinction between the social and the organic is far from

370 THE AMERICAN NATURALIST [Vou. L

a novel one. But the two groups of phenomena, and the proc-

esses involved in each, are still very frequently confounded in

other domains than that of use inheritance. The whole eugenics

movement, for instance, so far as it is a constructive program

and not a mere matter of ordinary practical prophylactic social

hygiene, rests upon the assumption that social progress can best

be accomplished by organic means. ‘It may be rash to deny

wholly that such an end can be achieved in this way or that it

would be useful. But the orthodox eugenist, from the time of

the founder Galton, has consistently and complacently made this

assumption without any inquiry as to its justification. Lamarck

erected a false doctrine of evolution through explaining the

organic in terms of the social, or in terms derived by mere an-

alogy from the social. The eugenists of to-day, it may fairly be

suggested, bid fair to vitiate a movement that springs from the

most sincere of motives, by resting its basis on an interpretation

of the social as merely organic.

In summary, the doctrine of the hereditary transmission of

acquired characters is no more disprovable than it is provable by

accumulation and analysis of evidence. It springs from a naive,

unscientific, and even primitive method of reasoning by analogy,

which in this case works to a confusion of the long-distinguished

and necessarily distinct concepts of the organic and the social.

The doctrine must therefore be dismissed on purely methodolog-

ical grounds. It is possible that when the missing factor or ele-

ment of evolution is discovered that neither Darwin nor the

mutationists have been able to find, this factor will prove to be

something superficially similar to use inheritance. But it will

differ from the present only half-discredited but discreditable

factor of heredity by acquirement, in containing an organic

mechanism, and will therefore be essentially different from this

crude and confused assumption.

A. L. KROEBER.

UNIVERSITY OF CALIFORNIA

TRIFOLIUM PRATENSE QUINQUEFOLIUM

Hugo De Veris in his mutation theory tells us in detail about

his production, by means of selection from two mutant forms,

of a five-leaved race of red clover. This race he called Trifolium |

pratense quinquefolium. The two plants obtained for starting

No.594] SHORTER ARTICLES AND DISCUSSION 371

his selection were found, according to the author, growing near

the edge of a road that was covered with grass. He does not teil

us the exact composition of all the leaves of these two plants with

which he started, but states that they bore several tetramerous

and one pentamerous leaf. Neither of the plants, therefore,

could be called mutants of a new race, but were mutating forms

from which De Vries obtained, after a process of most rigid se-

lection, his highly variable race, Trifolium pratense quinqué-

olium

During the spring of 1914, I found growing in an old orchard

at Corvallis, Oregon, a red clover plant that showed ‘‘full-

fledged’’ all the characters of Trifolium pratense quinquefolium

about which De Vries has written, and which took him so long

to obtain by the aid of selection. This clover plant, after careful

examination, was transplanted in one of my experimental plots

for further study. The following descriptive notes of it are

given: Of medium height; good color; normal as to vigor, but

not luxuriant; seven stalks; leaves, 4 trimerous, 5 tetramerous,

12 pentamerous; total number of leaves, 21. Not only did the

pentamerous condition of so many leaves represent the mode for

leaf variation, but there were more five-leaved leaves than both

four-leaved and three-leaved leaves combined.

The magnitude of this mutation may be more fully appreciated

when we reflect that De Vries, after selecting for three genera-

tions and obtaining 300 plants, found only one that gave as high ~

a percentage as 36 for both tetra- and pentamerous leaves; while

the percentage of tetra- and pentamerous leaves for all those

counted, 8,366, was only 14.

After finding this specimen of Trifolium pratense quinque-

folium I was exceedingly desirous of obtaining another plant

with which to cross fertilize it so as to obtain a race which could

be used commercially, but repeated searches made for many days

failed to reveal any other plant suitable to cross with this one.

Thus failing to find a second plant, I decided to propagate the

discovered mutant vegetatively. This method gave some degree

of success, and a few plants were reared during the summer of

1914 from slips. When I left Oregon at the end of the summer,

four of these plants were transferred to a private lot, and a rail-

ing, supported by stakes, was put around them.

xamination on June 3 the following summer (1915) showed

two of these slips doing well, one had been trampled on and

372 THE AMERICAN NATURALIST [Vou. L

killed by a cow, and the other was dead. At this time I still had

hopes of obtaining a race of five-leaved red clover, but when I

returned to Oregon and examined these plants on June 20 all

such hopes vanished, for a neighbor’s cow had completely ruined —

them, cropping off all the stalks down to the ground. None of

the plants revived after this last injury.

Records for the leaf production of this mutant were kept, and

from them I have obtained the following: On May 11, 1914, a

count was made of all the leaves produced up to date. There

had been produced 6 trimerous, 7 tetramerous and 17 pentam-

erous leaves; 30 in all, over 56 per cent. pentamerous. We no-

tice, however, a slight decline in the percentage of pentamerous

leaves produced, since the plant was found. This decline, early

noticed, continued throughout the summer, and on August 23,

when I made my last leaf counts, I obtained the following record

of leaf production for this mutant clover plant:

Trimerous Tetramerous) Pentamer-

ei Whar: Noes i S Leaves Leaves ous Leaves

Leaves on plant when found .................. | 4 5 12

30 11 f7?

34 T EE

` E -*

a agit

Trimerous Tetramerous Pentamerous

Fic. 1. Frequency polygon showing leaf variations of a mutant indi-

vidual of Trifolium pratense quinquefolium. Light solid’ line shows variations

of leaves on plant when discovered: dotted line, variations in leaves produced

No.594] SHORTER ARTICLES AND DISCUSSION 373

How are we to interpret these results? Why should there be

such a preponderance of pentamerous leaves produced during

the early growth period of the plant, then a preponderance of

trimerous leaves during the latter part of the season?

The records obtained for 1915 for the slip plants added but

little to the 1914 records. The leaf production of slip plant No.

3, however, is interesting. In 1914 this slip plant produced 4

trimerous, 2 tetramerous and 4 pentamerous leaves. In 1915,

however, it produced 0 trimerous, 0 tetramerous and 4 pentam-

erous leaves—showing strongly the inherited tendency to pro-

duce the pentamerous leaves during the second season. Clover

heads were produced during the summer of 1914, but no seed

was found in them.

No leaves were produced by this plant having more than five

leaflets, a condition that obtained during the first three genera-

tions of De Vries’s race, yet later he obtained both 6- and 7-

merous leaves in abundance. A frequency polygon is plotted

(see figure) for the leaf variations of this red clover mutant.

H. E. Ewine

IOWA STATE COLLEGE

NOTES AND LITERATURE

FAUNAL DISPERSAL?

THE widely different conclusions drawn by different students

from the same facts is well illustrated by Matthew’s recent paper

on the geographic distribution of animals.? Although these dif-

ferences are of many and varied sorts a fundamental one seems

to be concerned with the question as to whether the peripheral

part or the central part of the range of a group contains the

more progressive members of that group.

The idea which is held, either consciously or unconsciously, by

many is that after a group has arisen it spreads; and the mi-

grants, meeting new conditions, develop new characters. These

new forms then spread still further and develop other charac-

ters. Thus the more primitive members remain at or near the

point of origin of the group; the successively more progressive

members will be found at respectively greater distances from

this point. According to this theory we may trace backward the

dispersal of the group by following the distributions of the sev-

eral members from the most progressive to the most primitive.

We may not be able to follow this line to the actual point of

origin of the group for the most primitive members may not be

alive now and we may be unable to find their fossil remains.

. However, as organisms usually extend their range in more than

one direction, we may be able to trace several lines of dispersal

and deduce that the point of origin was near the place at which

these lines tend to converge.

The other theory is as nearly the opposite as can be. Accord-

ing to it the first progressive makes its appearance well within

the range of the primitive members of the group. If the new

form is not an improvement over the primitive one, it dies out.

If it be an improvement, it crowds the primitive form which is

forced to leave its place of origin and migrate. It is sometimes

made a part of this theory (but it is not a necessary part) that

the reason a new form appeared among the old was that the new

climatic conditions appeared there and that if the old form could,

in its migration, keep in climatic conditions suited to it, well and

good ; if it could not, it either died out or-became adapted to the

TAS Read at the February 15, 1916, meeting of the N. Y. Entomological

ety.

2 W. D. Matthew, 1915, ‘‘Climate and Evolution,’’ Annals, N. Y. Acad. 2

of Sciences, XXIV, pp. 171-318,

374

No. 594] NOTES AND LITERATURE 375

climatic conditions into which it was forced without, however,

losing all the earmarks of its primitiveness. Later still, newer

forms appeared at the old stand and forced their predecessors

still further away. Therefore, if we follow the distribution of

the several members of the group from the most progressive to

the most primitive, we will tra¢e the dispersal of the group

forward, not backward.

It would probably be impossible to decide, either by logomachie

methods or by watching present-day movements of species, which

of these two theories is correct. However, we are not left with-

out hope because paleontologists are daily digging up evidence

which, when sufficiently complete and properly translated, will

leave us in no doubt as to the history of the dispersal of certain

groups, and there is little doubt that the same general principles

which hold for those groups will apply to others concerning

which we have not and probably never will have fossil record.

Of course details may differ from group to group, but it is not

probable that there is one set of laws for mammals and another

for reptiles, one for birds and another for insects, that nature

constructed a trans-oceanic bridge for one group, but stationed a

guardian angel on it to prevent the passage of others. The chief

differences probably were such as the differing responses of the

different groups to changes of the environment and the differing

powers of different groups in overcoming given barriers to dis-

persal.

Next to that brought about by the holding of the diametrically

opposing theories just discussed, perhaps the most important

source of confusion is in not keeping clearly distinct ‘‘center of

origin,’’ ‘‘center of dispersal’’ and ‘‘center of greatest develop-

ment.’’ They may all be in the same region or they may be as

far removed from each other as the earth’s surface will permit

but they are not the same. A group may arise at A and move to

B, from which point there are easy paths of migration to C, D,

E and F. It may, and doubtless would, go to all of these but it

might find that E alone furnishes good conditions for its future

development. Its center of origin would then be at A, its cen-

ter of dispersal would be at B, and its center of greatest saad

ment at E. According to the first theory discussed abov

more primitive forms would be found at A or B and the sabes

developments at C, D, E and F, but chiefly at E. According to

the second theory, C, D, E and F would have the primitive forms

(although they might be much modified, especially at E) and the

progressive members of the aap would be at A or-B.

376 THE AMERICAN NATURALIST [ Vou. L

Adams? did good service in bringing together a number of sug-

gested ‘‘criteria for the determination of centers of dispersal.’’

It was not to be expected, and doubtless Adams did not expect,

that all of them would stand up under a test. Without attempt-

ing to exhaust the subject, present anything new or review all

that has been said about them, certain notes may be made in

connection with this discussion, taking up the criteria seriatim.

1. Location of the Greatest Differentiation of the Type.—lI be-

lieve it is more than a mere question of definition to say that the

offering of this criterion is an instance of the confusion of ‘‘cen-

ter of dispersal’’ with ‘‘center of greatest development.’’ Two of

the stock illustrations of great differentiation are those of mar-

supials in Australia and lemurs in Madagascar. Since these

groups are so greatly developed on the respective islands they

should, according to this criterion, have spread out from these

islands to the rest of the world. Unless the paleontological evi-

dence, as brought together and interpreted by Matthew, is false,

that was not the history of these cases. On the other hand, wher-

ever their points of origin were, their ancestors got into the

Holarctic region and then spread in various directions. Now, if

this be true, their center of dispersal would be Holarctica, al-

though their greatest differentiation at the present time is toward

the other end of the world. Lest the quibble should be raised by

others, it should be stated that the real fundamental center of

dispersal of a group is its center of origin and that there are

3C. C. Adams, 1902, ‘‘Southeastern United States as a Center of Dis-

tribution of Flora and Fauna,’’ Biol. Bull., VII, p. 122. Aftor the pres- `

ent paper had gone to the press I soe: through the kindness of Pro-

fessor Adams, a portion of ‘‘An Eco logical Survey of Isle Royale, Lake

Superior’’ prepared under his direction and published (1909) as a re

of the Report of [Michigan] Board of the Geological Survey for 1908.

this he takes up, again, these criteria. He — ‘‘It should be clearly em-

phasized that it is the convergence of evidence from many criteria which

ust be the final test in the te ates of origins rather than the de-

pendence upon any supposedly absolute criterion.’? He discusses the vari-

ous criteria in greater detail than was done in his 1902 paper and adds, as

another criterion, ‘‘Direction indicated ee seasonal appearance; vern

suggesting boreal or montane origin and aestival as austral or lowland

derivation.’’ I regret that I did not have earlier aecess to this paper but

cussion in the 1909 paper by Adams. This is especially true since I in-

tentionally omitted, for the sake of brevity, reference to certain papers,

such as the one by Tower (1906), which bear on the same subject.

No. 594] NOTES AND LITERATURE 377

other centers of dispersal wherever there are branches in the

forward movement of the group; but I understand the usual

meaning of ‘‘center of dispersal’’ to be the point or points from

which the principal lines went out which resulted in the wide

spread of the group.

2. Location of Dominance or Great Abundance of Individuals.

—This criterion seems futile. A group may have moved in all

directions from a given region and died out entirely in the re-

gion from which it moved. Nothing is more clearly established

than the changeableness of geologic climate and hence of all en-

vironmental factors and we know of many instances where noth-

ing is left of a group of organisms but fossil remains in the re-

gions of their former abundance and a few living remnants in

far-away, protected spots. If, as seems very clear, we can not be-

- lieve that even the most populous of these havens marks the cen-

ter of dispersal of such a group, neither can we apply this eri-

terion with safety to any other group. The area of present domi-

nance is merely that area, of all those now inhabited by a group,

which is at the present time most suited to the group—unless, of

course, it has arrived in a more suitable area so recently that it

has not had time to develop its dominance.

3. Location of Synthetic or Closely Related Forms.—This cri-

terion will be considered more in detail later, but it may be re-

marked in passing that the location of closely related forms is of

little help in arriving at the center of dispersal unless we know

whether these forms are more primitive or the reverse and un-

less, furthermore, we have selected the right one of the two oppos-

ing theories which were mentioned in the beginning of this dis-

cussion.

4. Location of Maximum Size of Individuals.—It is difficult to

see why individuals should be larger at the center of dispersal

than elsewhere. Possibly they may be larger the nearer they are

to present-day, optimum environmental conditions, but this is

probably not often, and odes not necessarily, anywhere near

the ancient center of dispe

5. Location of Greatest PA and its Relative Sta-

bility, in Crops—Adam says this criterion is very closely re- _

lated to the second one. If so, it fails for the same reason. Also,

it would seem that the latter part of this criterion as with

criterion number one— ‘greatest differentiation of type.”

6. Continuity and Convergence of Lines of Dispersal. _This

certainly ought to work, provided we follow the lines in the right

direction. The difficulty is that north and south lines on a more —

Ei deo

378 THE AMERICAN NATURALIST Von: b

or less spherical world converge either way we go. Furthermore,

if we place confidence in oceanic bridges which have been washed

away, piers and all, the lines of dispersal are apt to be fre-

quently discontinuous.

7. Location of Least Dependence upon a Restricted Habitat.—

For example, certain plants and animals which, in the region of

New York City, are found only in sphagnum bogs, such as those

of the Jersey pine barrens, occur more widely distributed

farther north. According to this criterion their center of dis-

persal would be in the north and this may be true. Again, liz-

ards occur, in the region of New York City, only or chiefly in

these same pine barrens, while farther south they run about

wherever they can get sunshine. According to this criterion the

ancient center of dispersal of lizards was somewhere near the

equator, but this may not be true.

8. Continuity and Directness of Individual Variations or

Modifications Radiating from the Center of Origin along the

Highways of Dispersal.—This criterion is probably stated rather

more fully than was meant. If we knew all this, the problem

would be solved. Perhaps Adams meant that continuity and di-

rectness of modifications (“‘individual variations’’ gets us into

the biometric-mutation discussion and that is another story)

point out the highways of dispersal from the point of origin.

His reference on the next page to Osborn’s law of adaptive

radiation indicates that this is the proper interpretation of this

criterion and, if so, the present author has no quarrel with it

except to point out once more that we are left in doubt as to

which way to follow the lines.

9. Direction indicated by Biogeographical Affinities—I am

not certain as to what this means. I suppose a given group

which is neotropical at the present time has biogeographical affin-

ities with other present-day neotropical groups. If we know

the centers of dispersal of the other groups we have a working

hypothesis concerning the center of dispersal of the group in

question. If this be what is meant, it seems to be probable.

10. Direction indicated by the Annual Migration Routes, in

Birds.—This criterion is meant to apply only to birds and I fear

we know too little concerning the intricate problems of bird mi-

gration to say whether their present-day routes of annual mi-

gration follow the route of ancestral dispersal or not. Probably

they do not, as the birds would be expected to change their

routes with changing environmental conditions. Furthermore,

although it is believed that birds return to their ancestral home __

No. 594] NOTES AND LITERATURE 379

to breed, this is not so firmly established as to leave no doubt

about which is going and which is coming.

Adams clearly stated that these criteria are for use where

‘we do not have paleontological evidence in sufficient abundance

to materially aid us,’’ but I confess to a feeling that we must

still depend on paleontology to give us the general laws of dis-

persal. It is for this reason that such papers as Matthew’s seem

to me so helpful. I regret that the organisms in which I am most

interested did not leave more marks on the sands of time, but if

the great majority of mammalian groups left records to show

that they followed a certain set of lines of dispersal and the end

result is of a given character, it seems worth while to compare

the end results of the dispersal of other forms with those of the

mammalian groups. If the comparison is close, the deduction

that the lines of dispersal also are comparable seems not unsafe.

It certainly seems unwise to construct trans-oceanie bridges

where conservative geologists say there could have been none,

especially if they must be made so tenuous that only insects,

spiders, snails, earthworms, fresh-water fishes and such small fry

can cross, mammals being forbidden.

Several of the criteria given by Adams which appear to fail in

helping us discover the ancient centers of dispersal seem to be

indicators of present or potential centers. Thus the first one: a

region where there is a ‘‘great differentiation of type’’ within a

group would seem to be a region prepared to send members of

that group into all the world. If paths exist or chance inter-

venes, this group should be able to fit some, at least, of its many

different types into the new environments which it encounters in

its spread. If all the surrounding territory is already occupied

by more successful competitors, that merely means that there are

no paths for the dispersal of this group. One of the real, but

sometimes overlooked, barriers to faunal spread is the presence

of competitors. This, however, does not negate the idea that a

region of ‘‘great differentiation of type” is a potential center of

dispersal.

In somewhat the same way the region of ‘‘dominance or great

abundance of individuals’’ is a potential center of dispersal.

The case is not so clear; but it would seem that where there is a

relatively large supply of individuals more could be spared for,

or would be forced into, colonization ater than where there :

are relatively few.

This suggests that what was the center of greatest develop-

ment (especially in variety of types) became the center of dis- :

380 THE AMERICAN NATURALIST [ Vou. L

persal, and that what is now the center of greatest development

is or will be a new center of dispersal provided there are means

of dispersal. There seems to be a genetic relationship between

these three centers and, of course, at the beginning the center

of origin is also the other two.

Little need be said here about means of dispersal except to

point out what Matthew also emphasized, namely, that time is

long and luck is real. Those of us who have been brought up on

the doctrine of evolution by selection of ‘‘chance’’ variations

should have a whole-hearted respect for that sometimes abused

word. Those who believe there has been time enough for alert

Nature to seize upon enough chances to differentiate 400,000 spe-

cies of insects, for example, need not strain unduly in swallowing

the notion that a very small proportion of these have been able

to get across relatively short stretches of water without a bridge.

Those who believe that Nature not only seized upon but made

opportunities for the differentiation of species should have no

trouble in discovering easier ways for her to help her creatures

spread their range than by raising up long narrow portions of

the ocean bed for a certain few to cross dry shod and then sink-

ing it to the discomfiture of those which are not of the elect.

Insects probably get about as easily as any creatures because

most insects fly or may be blown long distances and, further-

more, the majority have at least two distinct stages in their life

cycle during which they remain inert and without the necessity

of feeding. If mammals ean reach islands not connected by

bridges, surely insects of many kinds can. Matthew places great

stress on natural rafts as a means of transporting mammalian

fauna. After giving briefly a ‘‘series of facts and assumptions `

[which] may serve to give some idea of the degree ot probability

that attaches to the hypothesis of over-sea transportation to ac-

count for the population of oceanic islands”? he says:

If then we allow that ten such cases of natural rafts far out at sea

have been reported, we may concede that 1,000 have probably occurred

in three centuries and 30,000,000 during the Cenozoic. Of these rafts,

only 3,000,000 will have had living mammals upon them, of these only

30,000 will have reached land, and in only 300 of these cases will the

species have established a foothold.t This is quite sufficient to cover

the dozen or two cases of mammalia on the larger oceanic islands.

Few of these assumptions can be statistically verified. Yet I think

that, on the whole, they do not overstate the probabilities in each case.

* Matthew apparently changed, between pages 206 and 207 , his ideas

_ concerning the probabilities. However, it is the general notion and not

the actual figures which is important.

No. 594] NOTES AND LITERATURE 381

They are intended only as a rough index of the degree of probability

that attaches to the method, and to show that the populating of the

oceanic islands through over-sea transportation, especially upon natural

rafts, is not an explanation to be set aside as too unlikely for con-

sideration.

I confess to some haziness as to the probabilities here set forth,

but, if they are anywhere near true, entomologists need not

worry. In addition to their creatures not needing rafts as badly

as do mammals, it is certainly probable that every, not ‘‘one in a

hundred,’’ natural raft big enough to be noticed and recorded

by voyagers contained not one, but many, insects. Smaller rafts

or even single trees might contain many individuals of several

Species and, since a single fertilized female gives birth to many

offspring, the chance of a given species establishing itself on

virgin soil is much greater than it is in the case of mammals.

Furthermore, insects have been dispersing since before the Car-

boniferous. Many of the islands may not be that old, but this

simply means that insects have had a chance at such an island

since the first wavelets rippled about its uplifting peak. The one

thing which may be comparatively disadvantageous to insects is

that many of them are rather closely bound up in their food re-

lations with certain plants, but this disadvantage is somewhat

decreased by the fact that, if phytophagous insects are carried

on natural rafts, their food plant is likely to be a part of the

material which makes up the raft and both may be established

together. The pros and cons are numerous and involved. It is a

balancing of probabilities with the burden of proof on the side

which claims the right to make over major features of the earth’s

surface in the face of contrary geological evidence.

If this be true for the seant fauna of oceanic islands, what shall

we say of the suggested bridges, running this way and that,

across the oceans for the purpose of connecting continental

faunas and floras, especially in equatorial regions? Mercator

gave us a map of the world so constructed that the longitudinal

lines are parallel from the north pole to the south. Now the fact

is that a degree of longitude equals approximately 111,300

meters at the equator, 104,600 meters at 20° latitude, 85,400

meters at 40°, 55,800 meters at 60°, 19,400 meters at 80°, and no

meters at the poles. Therefore the equatorial distances on Mer-

cator’s projection are relatively far too short. On the globe or

on a proportional projection in which a meter at the equator is

as long as a meter in Alaska we see that north of the Tropic of

Cancer in the eastern hemisphere lies a huge land mass consist-

382 THE AMERICAN NATURALIST [Vou L

ing of Europe, northern Africa and most of Asia. This mass

almost touches at its northeastern corner the somewhat smaller

mass of North America. The intervening space, Bering Strait,

is only about sixty-five miles wide, very shallow, and dotted with

islands. The other gap between these masses is somewhat wider,

but still not so great as the shortest distance between South

America and Africa. Thus the arctic region is almost encircled

by land and itself contains much land. The earth’s surface is

really one huge northern land mass with three southward pro-

jections, namely America, Africa and the East Indian Islands,

including Australia. Antarctica is a small disconnected mass at

the other end. A species or group of species originating in, or

getting into, the far north could, as far as the present configura-

tion of the continents goes, populate the earth and have solid

ground under its feet most of the way. i

- Probably one of the main reasons for the little consideration

which students of geographic distribution have given to this route

of dispersal is the present climatic conditions in the far north.

It seems to have been easier to imagine the ocean’s bottom

heaved up between Africa and South America than to conceive

of a different climate in the northern regions. Yet we have defi-

nite and incontrovertible evidence of mild arctic climate in at

least several geologic periods, while the only moderately strong

evidence of a bridge across the Atlantic, for example, is the pres-

ence of related or identical forms on the opposite shores.

Suppose two of us are known to have been together on Broad-

way, but now one of us is in eastern Connecticut and the other at

the eastern end of Long Island. One theory might be that we

traveled together to the eastern end of Connecticut and then,

while one of us stayed there, the other crossed the Sound in some

way or other, even though we could not swim that far, there was

no regular boat service, and the only evidence of a bridge having

been built and then destroyed is that one of us is in eastern Con-

necticut and the other in eastern Long Island. It would seem to

me more probable that we parted company in Manhattan and,

While one of us crossed by a known bridge to Long Island and

then had good going along the southern shore of the Sound, the

other crossed known bridges and travelled through Connecticut

along the northern shore of the Sound. This seems to be the way

the respective theories concerning the biogeographical relation-

ships of South America, Africa and Australia stand except that

we have some other facts. To continue the comparison: although

there may be no direct evidence as how one of us came to be in

No. 594] NOTES AND LITERATURE 383:

eastern Connecticut and the other at the eastern end of Long

Island, yet it is known that other people have left New York and

gotten to these two places without crossing the eastern end of the

Sound and, furthermore (to make the comparison more accurate),

we must put in as a part of the argument that no one was ever

actually known to cross directly from eastern Connecticut to

Long Island or vice versa. In the face of such evidence he would

be rash indeed who would hold to the first theory as to our move-

ments after we left Broadway.

Now I know very little about geology and still less about fossil

mammals, but I am willing to take on faith the conclusions of

well-accredited students of these subjects if the conclusions seem

to have been reached by logical deduction from reasonable prem-

ises and if the assertions of fact are not too widely different from

my recollection of the assertions of fact made by students who

have reached other conclusions from what seem ‘to me to be un-

reasonable premises. It would be out of place here to give the

details of Matthew’s analysis of vertebrate, especially mam-

malian, paleontology. He takes up group after group and out-

lines their fossil record showing that they

accord fully and in detail with the principles® here set forth, and to be

impossible of explanation except upon the theory of permanence of the

ocean basins during the Cenozoic era. While the. prominence of the

Holarctic region as a center of dispersal is ascribed to its central posi-

tion and the greater area, some evidence is given to show- that climate

is also a factor in the greater progressiveness of the northern, since it

is also noticeable in the southern as compared with tropical faune.

The distribution of the Reptilia appears to be in conformity with the

principles here outlined, and exténds their application to the Mesozoic

era. The distribution of birds and fishes and of invertebrates and

plants is probably in accord with the same general principles, modified

by differences in methods of dispersal. The opposing conclusions that

have been drawn from the distribution of these groups are believed to

be due to an incorrect interpretation of the evidence. A few instances,

which have been prominently used to support opposing conclusions,

are analyzed and shown to conform to the conclusions above set forth,

if interpreted upon similar lines as the data for mammalian distribution.

As an example of the widely divergent conclusions which may

be drawn from the same data concerning present-day distribution

_ °Permanency of the abyssal oceans; climatic cycles from extremes of

cold or arid zonar climates culminating in glacial epochs, to the ext

of warm, humid, uniform climates, associated with cycles of moderate

ment of progressive groups of animals in the great northern land mass

and their dispersal southward as the result of these climatic changes. -

384 THE AMERICAN NATURALIST [ Vou. L

I take the liberty of referring to my paper on the spiders of the

Greater Antilles. On pages 139 and 140 it is shown that sixteen

Antillean genera, selected in a rather random fashion, are found

to-day in regions connected by a certain hypothetical system of

trans-oceanic bridges and not elsewhere. In some cases there

appears to be almost no specific difference, even, between spiders

whose present range is far from continuous. In fact, this set of

data and other instances mentioned in that paper seem to me to

furnish as strong evidence for such bridges as is given by the

recent distribution of any one group of organisms, and yet I felt

that this was not at all the explanation. It seemed more rea-

sonable to believe that spiders had dispersed by the way of

Holarctica on land masses which were practically the same as -

they are now and that the present discontinuous distribution of

the ancient types is brought about by the fact that they are

merely relicts which are now found far separated from each

other.

Entomologists will at once think of a number of species which

have reached the United States in historic times from Mexico and

many will use these in a contention that much of our fauna has

been derived from the south. I believe that the movement has

been largely the other way and that the ‘‘southern element’’ of

our fauna is largely made up of those things which have dropped

behind in the general southern movement. As in any stream

there are back eddies, so in the stream of dispersal we must ex-

pect back eddies (especially when man makes a channel as he did

when he planted a large number of potatoes up to the former

habitat of the Colorado potato beetle or grew great quantities of

corn and other cereals for the chinch bug) but the eddies do not

indicate the direction of the main stream. In their progress

toward the equatorial region the streams of dispersal leave pools

here and there—the stranded relicts of an ancient fauna. There

are doubtless numerous swirls and back currents, while near their —

mouths these streams of dispersal may be much subject to ‘‘tides”’

due to minor, i. e., measured by centuries, fluctuations of climate,

but their general movement is, nevertheless, from the poles oF

toward the equator. Since the northern polar regions have the

larger land masses and better facilities for such dispersal-stream

flow, the larger movements have been from the north, but simliar,

though smaller, currents are to be expected in the southern —

hemisphere, rank E. LUTZ

AMERICAN MUSEUM OF NATURAL HISTORY

° Lutz, F. E., 1915. ‘‘List of Greater Antillean Spiders with Notes on-

their Distribution.’? Annals, N. Y. Acad. Sciences, XXVI, pp. 71-148.

The American Naturalist

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

sent to the Editor of THE AMERICAN NATURALIST, Garrison-on- Huvson, New York.

Short articles containing summaries of research work bearing on the

problems of organic ee kop are especialiy welcome, and will be given preference

in peene

a reprints of contributions are supplied to authers free of charge.

furthers uiters will be supplied at cost.

S tions eae should be sent to the publishers. The

martirigi tao is four dol a year. Foreign posteer is fifty cents and

Canadian post Shea ier: yin additional. The charge for single copies is

forty cents. The advertising rates are Four Dollars for a page.

THE SCIENCE PRESS

Lancaster, Pa. Garrison, N. Y.

_ NEW YORK: Sub-Station 84

, April 2, 1908, at the Post Offi t Lancaster, Pa., under the Act of

Congress of March 3, 1879.

NATURAL HISTORY OF MAN | JAPAN NATURAL HISTORY SPECIMENS

Perfect Condition and Lowest Prices.

tna wage Racos many now extinct won aie a: of | specialty: Bird Skins, Oology, Entomology, Marine

ments, Deities, Currency, Carvings, Tools, etc. | Animals and others. Catalogue free. Correspond-

a eerste age. Type Collections for | ence solicited.

EETRI T. FUKAI, Naturalist,

OLDMAN, 77 Brixton Hill, London, England Konosu, Saitama, Japan

PUBLISHED BY THE CAMBRIDGE UNIVERSITY PRESS

MIMICRY IN BUTTERFLIES |

Bound in EEUN gece pir Pp. viiit+188. With 16 plates (12 of isin

colored). Price $4.50.

y By Tins CRUNDALL Punnett, F.R.S., Fellow of Gonville and Caius reo

ae Arthur Balfour Professor of Genetics cs in the University of Cam

EXTRACT FROM THE PREFACE ig

„This little Book h has been written in the hope that it may appeal aari z

_ Not infrequently I have been asked by friends of different callings in life to-

_ recommend them some book on mimicry which shall be short, well illus-

| been obliged to tell them that I know of nothing in answering t0 | —

į this description, and it is largely as an attempt to remedy thisd a defcieney that - re

_ present little volume has been es ;

Di land, oar also that it will be found of interest to those who live in or or visit

Rai

| trated without being very prot and not too hard to understand. I have — se ee

THE

AMERICAN NATURALIST

VoL. L. July, 1916 No. 595

SEX CONTROL AND KNOWN CORRELATIONS IN

PIGEONS!

Dr. OSCAR RIDDLE

STATION FOR EXPERIMENTAL EVOLUTION, Coup SPRING Harsor, L. IL., N. Y.

WHEN one nowadays states that he has obtained a real

control—a reversal—of the development of sex, he can

feel assured that his biological audience demands a very

large volume of rigid proofs. The first reason for this

large requirement is, as you well know, that the asser-

tion of sex control has been often made, and that in most

of these cases the data have proved disappointing; in-

adequate in one or another respect. A second reason for

present widespread skepticism as to even the possibility

of a real control of sex-development centers in the now

well-demonstrated fact that in some groups of animals,

the male—and in other cases the female—produces sex

cells of two kinds when these are considered from the

standpoint of their chromosomal numbers or character-

istics ; and further that each of these two groups of germs

normally gives rise to organisms of the sex correspond-

ing to the chromosomal constitution of these germs.

Moreover, certain linkage phenomena observed in breed-

ing such forms, unquestionably show themselves to be

normally associated with these same chromosomal dif-

ferences. |

But the experimentalist has learned through some pre-

1 Paper read ee the American Society of Naturalists, Columbus, Ohio,

December 30, 1915,

385

386 THE AMERICAN NATURALIST [ Vou. L

vious contests with ideas of fixity and causality that when

normal structural correlations have been demonstrated in

the field of development, nothing has been decided as

to causality and inflexibility; indeed it is commonly at

such a point that experiment applies the pressure of new

or unusual conditions and makes an approach toward

learning the nature of a phenomenon, by forcing the latter

to break from its normal correlations, and disclose some-

thing of its real nature through its versatility—through

its own capacity to shift from response to one set of

conditions, to response to another set of conditions.

Laws of causation, in the field of development, are not to

be deduced from studies concerning the normal associa-

tions of the structures of the cell; they may be approached

through demonstrations of the versatility and responsive-

ness under pressure of those processes native to living

matter.

We have stated that when sex is controlled an audience

like this will demand a volume of proof. It is clear that

the time limits here do not admit so extensive a presen- |

tation. I should like to note here, however, that Profes-

sor Whitman’s: complete studies on sex in doves and

pigeons have been prepared for publication, and for sev-

eral months have been in the hands of the publisher.

The results of my own studies of the past five years de-

signed to test the reality of the sex-control, and the na-

ture of sex, as exhibited in these forms, will already

doubtless fill another volume. And, since the last volume

of the posthumous works of Professor Whitman is near-

ing completion, I can promise that it will not be long be-

fore the work of preparing my own results for publica-

tion will be begun. Only when all of these data are fully

available to you, may we expect a judgment as to whether

the evidence for our thesis to-day is adequate. It is pos-

sible to give here, within the time limits, only an outline

of the kinds of study which have yielded evidence on the

question of sex-control in pigeons.

These studies were begun, and carried on for many

years by Professor Whitman. He obtained indispu-

No. 595] SEX CONTROL IN PIGEONS 387

tably—a profound modification of the sex-ratio, and iden-

tified in a general way the factors associated with the

modified ratios. Whether the modified ratios signified a

real control—a reversal—of sex could not at that time

be definitely decided. It was to help in making a deci-

sion as to whether the changed sex-ratios signified a

real—or only an apparent—reversal of sex that I pro-

posed in the winter of 1908-9 to carry out some chem-

ical studies on the ova of the doves and pigeons which in

Whitman’s hands were yielding these striking sex ratios.

The methods for the quantitative and qualitative anal-

ysis, of the very small samples to be used, were devel-

oped, and these were tested during 1909-10 on consider-

able numbers of the larger ova of jungle fowls and do-

mestic fowls. Since April, 1911, I have carried on this

and other lines of study to determine if possible whether

the changed ratios observed by Whitman involve a real

reversal of sex; this work is being actively continued.

Whitman showed that ‘‘width of cross” in doves and

pigeons is of first importance in determining sex ratios

and that the wider the cross the higher is the proportion

of males. Family crosses produce—in practically all

matings—only male offspring. Generic crosses produce

from their ‘‘stronger’’ germs—those of spring and early

summer—nearly all males. If, however, the birds of

such a generic cross be made to ‘‘overwork at egg-pro-

duction’’—that is if their eggs are taken from them as

soon as laid and given to other birds for incubation—

then the same parents which in the spring threw all or

nearly all male offspring may be made to produce all,

or nearly all, female offspring in late summer and

autumn. At the extreme end of the season eggs capable

of little, then of no development, are often found in such

a series. As the birds of such a mating grow older the

time of appearance of females, and of eggs incapable of

full development, is reached earlier and earlier in the

summer or spring.

_ In the case of a number of hybrids Whitman showed

that color is also affected by this pressure of reproduc-

388 THE AMERICAN NATURALIST [Vou. L

tive overwork and season. White color could be ob-

tained from the later, ‘‘weaker’’ germs, though this color

did not appear in birds from the ‘‘stronger’’ germs of

the earlier season. And further, that white, or whitened,

‘‘mutants’’ from pure breds were derived almost or quite

exclusively from those conditions which produce ‘‘weak-

ened germs.” Among such conditions are late season

and overwork, inbreeding and great extremes of age—

either very old or very young. This brief outline of

Whitman’s findings on sex is perhaps a more adequate,

and more accurate, one than I was able to give to one of

the societies represented here when I had only begun the

examination of this data in 1911. Two brief summaries

given on the chart (not given here) will assist in obtain-

ing a picture of the nature of the results. I may add that .

by very strongly ‘‘overworking’’ females of some species

—overworking them more strongly than Whitman did—

I have been able to obtain a high predominance of females

during autumn from a cross merely of specific? value.

This result is illustrated by Chart II, though the matings

there exhibited were prepared for the primary purpose

of illustrating results in the study of size. It will be

noted in the chart that parents overworked in a previous

year throw a high proportion of females during the whole

of the succeeding year, and most markedly in late

autumn. In this mating the ratio at the end of the sea-

son is 14 females to 1 male; in the other (not previously

overworked) there was an excess of females only after

overwork—during the latter half of 1914 (73:109); and

in the year following this overwork there were 21 or

more females, to 11 or fewer males. Such data are not

exceptional; they coincide with the usual.

Now, in the generice crosses which give all, or nearly

all, males at the beginning of the season.and all, or nearly

all, females in the autumn what is happening?—true sex

reversal? or is it selective fertilization, differential mat-

2Some data from pure breds (pure species) mated to their own kind

show also this predominance of females from late autumn under extreme

overwork; such predominance is here probably less pronounced than in the

ease of the crosses,

No. 595] SEX CONTROL IN PIGEONS 389

uration or a selective elimination of ova in the ovary?

This was from the first the whole of our own problem.

We have had no other, nor have we now, except in so

far as the entire question of the natwre of sex—in germ

and adult—is concerned.

Our method has been to study the eggs, progeny and

parents of such series as show this seasonal ‘‘reversal of

the dominance”’ of sex from as many different angles as

possible. The result till now is that we have learned

some ten kinds of facts concerning the germs, or the

prospective value of the germs, which issue from such a

series. Let us note that these ten lines of correlated

fact do not relate merely to a ‘‘normal’’ state of the

germs, but have to do with measurable changes which

occur when ova are subjected to the stress of parental

reproductive overwork, which as Whitman has shown is

accompanied by a shifting from male-production to

female-production during the progress of the season.

The diagrams of chart I will assist in making clear the

nature and significance of the several correlations. The

solid lines indicate a double correlation, i. e., for both

season and egg of clutch; the broken lines represent cor-

relations established thus far for only one of these.

The generic cross that has been most fully studied in-

volves Turtur orientalis—the Japanese turtle dove, and

Streptopelia alba—the white ring dove. These species

together with their reciprocal hybrids are shown (photo-

graphed) in another chart (not given here). Some data

for egg size, and for sex-differences in the adult size of

the several forms concerned—parents and reciprocal

hybrids—are also given in that chart. j

The first correlation that we have established for this

series results from a study of the size of ova—i. e., of

yolks freed from shell and albumen. The result clearly

establishes the fact that the yolks of late summer and

autumn—those that produce mostly, or all, females—are

larger than the yolks produced in the spring which give

rise to males. And there is no jump from the one size

to the other, but what may be better described as a

390 THE AMERICAN NATURALIST [Von. L

gradual enlargement. This fact is represented diagram-

matically on chart I, and some of the actual figures may

be had from the charts dealing with size of offspring

(II), with analyses (III), and with calorimetry (IV).

Spr. oz oy oe on oe #9 79 Fo Fo Fo lç Sq Qo on Aut.

R TNE Fee. | G27 w n ote S ae

a À

PI N me A a - % % % % %

"la a erzs, $

Ce % % n a % % A a o o % o

"3 aes 3

AA A

: È

LE oe "ERS,

<a as:

"4 ca > u ill

ip; “ty fees

N, Ed t

"5 e wees =

of a =r

gers

= N <

ae Oe aver <

"6 a= gE

ey =

ah R

*7 ay e

F NP

‘Ie aS

CHART I

COMPARATIVE SIZE or EGGS OF ALBA (a) AND ORIENTALIS (0).

At the same time that this seasonal increase in size of

yolks was learned, it also became evident that the eggs

of doves and pigeons are dimorphic: That the two eggs

forming the pigeon’s clutch bear usually—there are ex-

ceptions—a smaller yolk in the first, and a larger yolk

in the second. Since Whitman had already shown that

No. 595] SEX CONTROL IN PIGEONS 391 -

in the pure wild species with which he worked, males

predominate in hatches from first eggs of the clutch, and

females predominate in hatches from second eggs of the

clutch, it became evident that the male-producing yolk

is smaller—both in relation to season, and to egg of

clutch, than is the female-producing yolk. Correspond-

ing to the fact (commonly obtained from matings of in-

dividuals of the same species) that two males or two

females may sometimes arise from the same clutch, we

have found that a similar number of pairs of yolks of

these forms are equal in size; and too that such pairs may

be either large or small. The charts just referred to may

be consulted in this connection. We have previously

noted (1911, 1912) that in eggs laid by hybrids neither sex

nor yolk-size bears the above described relations to the

order of eggs in the clutch.

Still a third situation has yielded positive evidence

that the smaller yolks are male-producing and the larger

yolks female-producing—namely that in respect to age.

It has already been mentioned that Whitman learned

that the females which were ‘‘overworked’’ tended, when

older, to begin the production of females at earlier and

earlier stages of the season. Now a comparison of the

size of yolks derived from younger and from older birds

has conclusively shown smaller eggs for mature but

younger birds, as compared with the old birds (see Chart

2). In scores of individual cases the yolk-size has now

been followed from youth, and comparative youth, to

old age.

In even a fourth situation it has been possible to test

the relation of yolk-size to sex. Breeding data show that

from the very first egg in life, and the very first egg pro-

duced after a long period of rest or inactivity, more fre-

quently produce a female than do the first eggs of succeed-

ing pairs, or clutches. Our studies on the size of such

yolks show a wholly similar reversal of order of size of

the two eggs of the very first clutches; the size reversals :

here being more Froana than in the succeeding iub

(see Chart 4). : :

392 THE AMERICAN NATURALIST [Vou. L

CHART II

BREEDING Recorps—1914

Q St. risoria 641 (old); 1913 = 42 eggs

Series 1

= A 1/1 White 7/20 White 140

1/3 White dead eps o'T2 7/22 Dark 164

Y yA =2.066 g. 2d (4) = 2.2

U1 7/28 White 144

H1 4/4 Inf. yolk =1. me $ Q9 U2 7/30 i

H2 4/6 Inf. yolk =2.

s À QV 8/14 White 155

Q I1 4/12 White killed 4/29 | nae 8/16 Dark 169

912 4/14 | | White 158 (?2) |

W1 8/22 White 152

oJl1 4/21 Dark killed 2/25 W2 8/24 Soft at pole

9 J2 4/23 White 158

9X1. 8/30 White 161

9 K1 4/29 White 147 9 X2 9/1 White 145

Q K2 5/1 White 151

g Yıl 9/9 Dark 161

Ll 5/9 Broken 9 Y2 9/11 White killed

L2 5/11 Dark

9Z1 9/18 White ——

oMi 5/18 Dark 161 (?1) 9 Z2 9/20 White dead 10/26

Q M2 5/20 White 163

9 AA1 9/26 White 141

Q N1 5/30 White 150 9 AA2 9/28 White 146

Q N2 6/1 White killed with

ext. 9 BB1 10/7 White 150

9 BB2 10/9 . White 144

goi 6/7 Dark 150

202 6/9 White 150 ?9CC1 10/17 Dark dead 11/8

9 CC2 10/19 White dead 11/10

g Pi 8/18 Dark 149

E r 6/20 Broken 9 DD1 10/26 White 130 (?1)

$ DD2 10/28 White 162 (?2)

Q Q1 6/28 White 143 ;

2 Q2 6/30 White 137 J'EEL 11/6 Dark 152

9 EE2 11/8 White 143

Q R1 7/4 White 154

o'R2 7/6 Dark 162 OFFI 11/16 White 166

FF2 11/18 Broken

?g'S1 7/12 Dark dead 7/29

9 S2 7/14 White dead 7/31 9 GG 11/26 White 150

lst 16 =59:119 2d15=5g:109

last 15 =107:149

The relation of the order of the eggs in the clutch to

the prospective sex of the offspring is an important

point, and we wish here to make this situation clear,

since it seems that two rather brief statements made in

1911 and 1912, before the Society of Zoologists, have not

been understood by all. F

From the time of Aristotle to the present year thore

No. 595] SEX CONTROL IN PIGEONS 393

CHART II—Continued

BrEEDING REecorps—1914

Q St. risoria 647 (young); 1913 = 18 eggs

Series 2

Cl 2/8 Inf. yolk =1.445 g. 9 Pi 7/1 White 150

C2 2/10 Broken 9 P2 7/3 White 15 da. emb.

D1 3/5 Dark embr. 9 Q1 7/9 White 148

9 D2 3/77 - White FQ? -7/11 Dark 164

g'El 3/19 Dark 167 QR1 7/22 White 152

o'E2 3/21 Dark 180 gZ R2 7/24 Dark 172

QF1 3/29 White 154 S1 8/3 White 13 da. emb.

g'F2 3/31 Dark 190 $2 8/5 Broken 3 da. emb.

Gl 4/8 Dark killed 5/6 o'T1 8/12 Dark 174

9G2 4/10 White killed 5/3 Q9T2 8/14 White 164

Q@H1 4/16 White 153 U 8/20 yolk =1.490 g.

Q@H2 4/18 White 153

V1 9/6 “Blood circle”

ll 4/25 Dark 169 oV2 9/8 Dark 170

912 4/27 White 154

?W1 9/19 Dark dead 10/16

J1 5/5 3-da. embr. killed QW2 9/21 White dead 10/14

J2 5 -da. ki

SY a one a ZX1 9/30 Dark dead 10/19

ite 145

@K1 5/14 Dark 169 QX2 10/2 Whi

@L1 5/25 Dark 179 Q¥Y2 10/31 White 15 da. embr.

SEL, 37 WIWO Z1 12/27 No dev. yolk =1.870 g.

i - 92 12/29 No dev. yolk = 1.925

SMi’ 6/3 Dark 169 I wate

Q9M2 6/5 White 11 da. emb. 9641 =(170) (7'170) ` aoa

3's (5) from Ist =155 = 9’s (18) = 149

| g’s (3) from 2d=165 9's (11) =150

ON1 6/13 Dark 165

Q9N2 6/15 White 150

g'Ol 6/22 Dark killed 7/13 | g's (7) from 1st =170 Q's (5)=161

O2 6/24 Broken | g's (5) from 2d=175 9's (6) =158

| A

lst 17=99:89 2d17=70':1092 1915 =11 dark: 21 white

have appeared statements concerning a predominance—

or a lack of predominance—of males from the first egg

and of females from the second egg of the pigeon’s clutch.

It is unnecessary to outline these divergent reports. It

is only necessary to point out the reason for discordance;

though the reason we had thought to be quite obvious

since 1911. The statements hitherto made have all been

based on a general statistical method, which is a wholly

394 THE AMERICAN NATURALIST [Vou. L

CHART III

SUMMARY OF PARALLEL BREEDING AND CHEMICAL STUDIES ON sey Ea@es OF

Ọ T. orientalis No. 500X St. alba No. 410 FOR THE YEAR 1912

- Result

Date An’l’s or | Wt. of | | |

Inc. Yolk Ale. Phos- | Energy

Soluble | phatids | Protein, Ext. Ash H20 | Total

: | |

4/13 Broken when found |

4/15 Broken when found |

| |

18.32 | 25.44 | 5.28 | 4.85 | 57.01 7,405

17.54 | 25.63 | 5.25 | 2.62 | 54.82 8,990

5/26 159 2.330 | 72.65

5/28 160 2.660 TRAD |

[i [i

6/7 Inc. .... | Only one egg laid Dark g

| |

6/15 Ine. aeons Ss a ES ae Borate Cae T igre

6/17 Inc. .... | “Very large egg” ee age bib Q

1 T

6/24 Inc. aa ha aa oaa ES En Ve atoms ONA wie (No. dev.

6/26 Ine. esr bee a EA E we Lee et ies cee as .. Dark J

7/3 . 186 2.026 | 72.21 | 16.49 | 26.00 | 3.63 | 2.43 | 56.05 | 6,714

7/5 187 2.330 | 72:27 1-19.18 | 26.55 | 3.75 1 1.93 | 55. 22 | 7,881

7/15 Inc. Se oe Gy J a ed de ee ee

Tie ane: ee oe a er Poa ee

7/23 192 2.422 | 72.42 | 17.82 | 25.88 | 3.82 | 1.80 | 55.84 8,061

"7/25 193 2.720 | 72.45 | 18.88 | 25.96 | 3.86 | 1.81 | 55.33 9,296

8/2 Inc. mae ee aa a on rene D ee

8/4 Ine. De Ae Ter ee ees ee a DE

8/13 Inc. poy a ee ee ak ea Oe ee a ea

8/15 Ine. ea aye en bee Mae bn ig | in Darks

8/23 Ine. seed eea chee aa da oes | .... |No dev.

8/25 Inc. : | : | White Q

9/15 Inc. Sk T E ak Ae T e Bee | e Wha 9

9/17 Inc. Ses bos E E E E a | White Ẹ

11/29 259 2700 | T317 T2140 26084 oa as [5552 9,323

12/1 260 2.715 | 73.02 | 21.63 | 25.38* : | 55.39 9,383

* Calculated.

inadequate and useless one for a study of the problem.

It is now clear that the method that would be valuable for

this purpose must be a thoroughly analytical one. Whit-

man has properly analyzed this situation. He has shown

that normally —i. e., with effects of crossing eliminated— —

from the periods for the production of the strongest

germs an undue proportion of pairs of eggs produce u

males; and from the opposite period there arise undue —

No. 595] SEX CONTROL IN PIGEONS 395

CHART IV

STORED ENERGY oF Eces (1914) oF Streptopelia risoria (558) as DETER-

MINED BY THE BOMB CALORIMETER i

No. Date Wt. Yolk Energy Per Cent. Diff.

665 Al 6/6 1.010! 3,358!

666 A2 6/8 0.970 3,175 —5.8?

` 674 Bl 6/19 0.855 2,807

675 B2 6/21 1.000 3,245 +15.6

699 C1 7/14 1.145 3,815?

700 C2 7/16 1.463 | 5,008 +31.3?

728 D 8/30 1.395 | < es

ECER E 9/9 or 10 soft shell, broken |

P oi es erae]

ar F2 10/19 Weaken y |

770 G1 11/6 1.440 | 4,837 (?)

771 G2 11/8 1.720 | +19.8 ?

774 H1 11/20 1.590 -+sl. loss | 4,906 +

775 H2 11/22 1.780 6,015 +22.6 —

776 Ii 12/1 1.640 5,614

777 12 12/3 1.820 6,255 +11.4

781 Ji 12/12 1.535 5,302 |

782 J2 12/14 1.690 5,601 | +5.6

791 K1 12/23 1.485 ` 5,266 (7) | j

792 K2 12/25 1.718 5,880 ! +11.7 ?

1 This egg was not only the first laid during season, but first during lite

of this bird.

2 The percentage differences are based upon a value of 100 per cent. for

the smaller egg of the pair. l

numbers of pairs of eggs that produce females. To

lump these all together and to count the number of males

arising from first, and females from second eggs is plainly

to cover up or to lose the significance of the intervening

pairs of eggs which bear the significant data. Again,

many matings, because of exceptional strength or of

weakness, will yield a considerable total predominance of

males or of females, and the statistical method lumps all

these and others without thought or care of the cancella-

tions and unsatisfied cancels involved; all of which as

easily contributes to a smoothing of the results, as it does

to a smothering of them.

But Whitman has also shown that not only is the

method previously employed at fault, but that, much more

important still, the material used—in probably all of

those cases in which no correspondence of sex to the

396 THE AMERICAN NATURALIST [Vou. L

order of the eggs of the clutch was found, and where the

worker has thought it worth while to mention the kind

of birds studied—such material has been wholly unsuit-

able to leading to a decision. That is to say, the

‘‘pigeons’’ used in these cases were one or another of the

150 mongrels collectively known as domestic pigeons.

One of the clearest points of our present knowledge of

the relation of sex to egg of clutch is that the normal re-

lations are lost immediately upon hybridization—4. e., in

passing from the pure state of the species. The count-

less degradations and crossings suffered by the various

domesticated breeds since their existence as a pure

species, is therefore a sufficient index of the suitability

of this material for a study of this subject. Whitman

demonstrated the predominance of males from the first,

and of females from the second egg of the clutch when

pure species mated with pure species produce the eggs,

and also the random distribution of the sexes from the

eggs of hybrids. And as early as 1911 and 1912 I dem-

onstrated charts and lantern slides which showed that the

size of the yolks from pure species showed with consid-

erable uniformity a smaller first, and a larger second

yolk; and further, that this regularity breaks down at

once and completely in hybrids.®

et us now note the conclusions which follow upon the

demonstrated dimorphism of the ovat in the pigeons,

when this is reviewed in the light of breeding data on

these forms and in connection with the demonstrated re-

lationships of size of yolk to sex—relationships which

are continued even under the pressure brought by over-

‘work, season, and age.

It Doomen clear, first of all, that a selective fertiliza-

tion by one kind of sperm is quite impossible—the sex

3 Note that in Chart 2, already referred to, where the eggs are produced

by the female cage or blond a ring dove—in which purity of the species is

often doubtful—that a predominance of males from first, and of females

from the second egg of the clutch is indicated in both series. In series I,

where the two sexes arise from a single clutch, the first egg gave rise to

the male in 6 (or 7?) eases; to a female in 3 cases. In series os the first

egg yielded males in 9 (or 210) cases; females in o i

4 Yolk size has now been nécntataly datirati 1 in akok 10, “roe cases. —

No. 595] SEX CONTROL IN PIGEONS 397

differential residing in two kinds of eggs and not in the

sperm. We may here recall that previous to our own

studies, breeding data obtained from other birds had in-

dicated that in the birds the sexually dimorphic germs

are borne by the female—or to use Mendelian terms,

that the female bird is heterozygous for sex.

The second conclusion that must be drawn is that a

selective elimination of ova in the ovary does not occur

during ‘‘overwork,’’ while mated to a mate of another

genus, nor otherwise, since the two kinds of ova are—

from their size relations—positively known to present

themselves under these, and under all the conditions

which have been studied. In other words, the generic

cross, which produces all or nearly all males in the spring,

and all or nearly all females in the autumn, is utilizing

in the spring a number of female-producing ova for the

production of males, and in the later season is utilizing

for the production of females ova one half of which had

initial inclinations for the production of males. Note

too that the evidence for the continued production during

the season of ova of two kinds as regards sex does not

rest alone on our knowledge of the dimorphic ova. For,

from breeding data we learn that if the same female

which threw all males in the spring and all females in the

autumn, had been mated to one of her own species, then

both males and females would certainly have appeared

at all seasons, and largely or wholly in relation to the

order of the eggs of the clutch, with but slighter effects

of season to be noted. If the overwork were extreme, a

predominance of females in late autumn might be ex-

pected; but in the earlier season the sexes would surely

be found in nearly equal numbers. Several of the cor-

relations soon to be mentioned, moreover, further attest

that ova of two grades—in respect to sex—are produced

throughout the year.

The data thus far examined exclude the possibilities

of accounting for the observed sex-ratios of the generic

cross on the basis of a selective action of the sperm, or

of a selective elimination of ova in the ovary. What

light do these data shed on the possibility of acco unting :

398 THE AMERICAN NATURALIST [ Vou. L

for the seasonal difference in sex-production on the basis

of a differential maturation? The fact that the sperm is

present in the pigeon’s egg during the whole of the

second maturation division may properly raise this ques-

tion. On this point we must say that the particular data

we have just been citing are perhaps not entirely con-

clusive; these data alone, however, offer the following

RE ee points for consideration: To account for the

Pria sex ratios of the generic cross the maturation

would have to be definitely differential in (1) the elim-

ination of an X chromosome during the spring from one

half of the ova, and the retention of this same X in the

homologous? eggs of the autumn. (2) The elimination

of a Y chromosome from the other half of the eggs laid

during the autumn, and the retention of all these same

Y’s in homologous eggs of the spring; and (3) all other

chromosomes than the sex chromosomes must display

no such thing as seasonal preferences for ‘‘staying’’ or

for ‘‘going,’’ since every observable character of the

hybrids betrays the presence of both of the parental

genera. This is not all, but let us pause at this point to

note that even if the sex chromosomes were here capable

of such wholly unknown and almost unthinkable þe-

havior, that they have—after all—in this case wholly

lost the initiative in governing sex, since it is the place

in the season and the degree of the pressure of the over-

work that has been shown to prescribe the sex of the off-

spring; and further, the correlations of size, water con-

tent, energy storage, etc., which we have proved to exist

throughout the whole season—these correlations are all

established prior to the formation of even the first polar

body; this latter being formed only at the time of ovula-

tion, and the second polar body forming 1 to 14 hours

after the entrance of the egg into the oviduct.

If, however, we were inclined to set no bounds to the

5 The chromosome situation in the germ cells of female doves and pigeons

is as yet quite unknown. But whatever it may be, our statement illustrates

the difficulties of a chromosome theory in the cases under consideration.

We make use of a familiar case in which XY germs are male-producers, x

and XX germs female producers.

eI. e., in eggs of identical (original) chromosomal constitution.

No. 595] SEX CONTROL IN PIGEONS 399

marvels of selective power that may be exhibited by the

sex chromosomes, and to feel that even the above difficult

formula remains for them a possibility, we may refer to

the decisive data obtained in studies-on the sex behavior

of the birds which are hatched from such a sex-controlled

series. We shall there see that those data differentiate

several grades of females. Some are quite nearly males,

—though they lay eggs. Is it too hazardous to suggest

that in one and the same egg the Y could hardly have

‘‘gone out’’ to allow the egg to develop into a female,

and yet have ‘‘stayed in’’ in order to deliver the rela-

tive masculinity that we easily detect and measure? If

sex is directly the creature of a sex chromosome, the sex >

situation found in some of my female doves requires that

the male-producing chromosome be eliminated from, and

retained in, one and the same egg! The only alternative

that it is within my power to imagine is that in addition

to the selective elimination of the Y’s during autumn,

there be further postulated a gradual fractional elim-

ination of parts of this chromosome, larger and larger

parts being eliminated during the progress of the season.

Or, that the reverse of this occurs, namely that the Y,

during the progress of the season, gradually adds some-

thing of X quality to itself, finally becoming more X

than Y. For those who would value this interpretation

I have no evidence or word of contradiction. The fact

must always remain that our procedures have not only

produced male and female from ova of opposed initial

tendency—largely under control—but that several grades

of intermediate sex have also been produced.”

7 Three previous publications, besides several addresses before the Amer-

ican Society of Zoologists and elsewhere, have clearly stated this result.

The publications now two years since, and the citations are as vais (a)

Carnegie Year Book, No. 12, 1913 (p. 322), Report of Year’s Work. ‘‘The

results strongly indicate that the hereditary basis of sex (and, therefore,

probably all characters) is a quantitative, graduated thing; not qualitative

and alternative as rather generally believed.’’ (b) Science, N. 8., Vol. 39,

No, 1003, Mar., 1914 (p. 440), “A Quantitative Basis of Sex as as Indicated

by the pres Rehasice of Doves from a Sex Controlled Series.’’ ‘‘These .

results together with our very abundant data on the storage metabolism of

the ova of these toa and the initial fact of sex venient. itself, strongly

400 THE AMERICAN NATURALIST [Vou. L

We shall be able presently to note more closely the con-

clusive facts as to the matter of a differential matura-

tion. Continuing our examination of the further data

which we know correlate with this sort of a sex-series we

shall meet with additional and other kinds of facts which

lead toward a constructive view of the nature and basis

of sex; facts immediate and specific concerning the meas-

ured powers or capacities of these series of ova which

present us the sort of sex-series in question—facts which

reveal sex in quantitative terms.

Correlations marked (2) and (3) on Chart 1 were first

noted by Professor Whitman. I have been able every

year to find many confirmations of his conclusions. :

The curve for ‘‘Developmental Energy’’ on the chart a

indicates a progressive seasonal decrease of this capacity a

in the fertilized eggs; a decrease from spring to autumn.

ow the evidence is unquestionable for the lowest part

of the curve—the autumn. In general, least develop-

ment proceeds from the last eggs of the season. These

are the largest eggs of the year. There is also less de-

velopment in the second eggs of the clutch. These are a

the larger of the clutch. It is thus seen that the larger

the yolks the less ‘‘developmental energy’’ possessed by

them

The “Length of Life’’ of the several offspring of such

a sex-series tells again of an advantage possessed by the

earlier hatched birds, and of a more limited life-term

affixed to the later hatches. It is further probable that

within the group of clutches giving rise to females only,

a longer average life-term falls to those who hatched

from the first egg of the clutch, than to those arising

from the second. Here, then, as in correlation no. (2)

the smaller eggs of clutch and season are the eggs pro-

indicate that the basis of sex is a fluid, reversible rie that the basis

of adult sexual difference is a quantitative rather than a qualitative thing.’”

(e) Bulletin of the — owe emy of Medicine, aL 15, No. 5 (Oc-

tober, 1914) (pp. 205-295), The Determination of Sex and Its Experi-

mental Control.’’? ‘‘The sum of these results, together with the initial fact

of sex control itself, practically prove that the basis of sex is a fluid, re-

versible process, that the basis of adult sexual difference is a g '

rather than a atdan thing (p. 277), ” ete., ete,

No. 595] SEX CONTROL IN PIGEONS 401

. ductive of ‘‘strength.’’ The larger eggs both of clutch

and season more often display ‘‘weakness.’’ And in

passing we might note that by the procedures involved

in these sex-series it is possible to graduate the fatal

dosage, and in great measure to predict which of particu-

lar germs must come to an end first.

The fourth kind of fact pertaining to the eggs of this

series, proceeds from the results of more than 800 chem-

ical analyses of individual eggs. The results of earlier

studies of this nature were described in 19118 and 1912

more fully than time limits will here permit; but the na-

ture of these results can be noted with the help of Chart

3. It will be observed that not only does the size of the

egg increase with its later position in the series, i. e.,

with lateness of season, but the percentage of energy-

yielding or stored materials increases as much as, or pos-

sibly more than, is indicated by the size—or net weight—

of the yolk.

The importance, for our present purpose, of the re-

sults of these analyses is that they conclusively show (1)

that the male-producing egg of the spring is an egg that

stores less material than does the female-producing egg

of the autumn. (2) That the male-producing egg of the

clutch stores less material than does its female-producing

mate. (3) That the eggs of old females store more ma-

terials, and—as has been noted—yield a higher percen-

tage of females, than do birds not old. Therefore, it be-

comes evident that the egg of female-producing tendency

is one whose storage metabolism is high, as compared

with eggs of male-producing tendency. The analyses

show that during the season successive clutches present

higher and higher storage, i. e., the earlier clutches store

less—are more male-like; the later ones all store more—

are more female-like; and as we have seen, the eggs of

the low storage period give rise to males, those of the

high storage period produce females. Here we obtain a

close view of that upon which sex difference rests. Un-

8 Papers read before the American Society of Zoologists. — For —

see Science, N. S., Vol. 35, i 462-463, March 22, 1912.

402 THE AMERICAN NATURALIST [Von. L

mistakably, less storage and high storage pertain respec-

tively to the male- and female-producing germs. Un-

mistakably, our procedure—connected with generic

cross, season and overwork—delivers males from the

smaller storages in the earlier eggs. Unmistakably, these

procedures raise the storage in all of the later eggs, and

unfailingly we then find that these eggs yield only, or al-

most exclusively, females. And if we eliminate the factor

of wide—or generic—cross and mate the female with

one of her own species, then we see that the production

of males and females coincides from the first with two

sizes of eggs in the clutch—males from the smaller first,

female from the larger second. Only after overwork

and season have raised the storage value of the eggs, is

this situation, in such a mating, seriously disturbed. And

the disturbance—associated with an increase in the stor-

age metabolism of all the eggs,—delivers, as before, an

excess of female offspring.

The progressive increase in storage capacity of the eggs

during the season—under overwork—is to be interpreted

as a decrease in the oxidizing capacity of these same eggs.

Living cells in general dispose of ingested food material

by storing it, or by burning it. The products of the ox-

idation are removable and do not serve to increase the

bulk of the cell. Likewise the low-storage capacity of the

male-producing eggs as compared with the high-storage

capacity of female-producing eggs is therefore an index

of higher oxidizing capacity of the male-producing eggs

as compared with the female-producing eggs.

The fifth correlation relates to the percentage of water

in the eggs of spring and autumn, and in the two eggs of

the clutch. These figures for one series of analyses are

given on the chart (3) last examined. They show a

higher water content for the eggs of the spring (male-

producers) as compared with the eggs of autumn (female-

producers); indeed, each pair of eggs from the first of

the season onward has a slightly higher moisture value © o

than the pair that follows it. The analyses further show

a higher percentage of water in the first egg of the clutch —

` No. 595] SEX CONTROL IN PIGEONS 403

than in the second in all cases. If the results of my 800

analyses all ran as smoothly as do the 8 of this series there

would be no doubt of a perfect correlation of high mois-

ture values with small eggs, i. e., with male-producing eggs

—both small eggs of season, and small eggs of individual

clutches. But the results are not thus uniform and

smooth. There are some series which seem seriously to

depart from the order noted above. These can not be dis-

cussed here. We can, however, record our own belief that

the situation represented in the chart is, in the main, in-

dicated by the moisture determinations.

Now the evidence that higher water values are asso-

ciated with male-producing eggs, lower water values with

female-producing eggs is of high importance in connec- -

tion with our own generalization as to the basis of ger-

minal sex-difference; and is further of much interest as

being the means of demonstrating that in the—as I be-

lieve—several valid cases of sex-control now known, one

thing in common has really been effected, this though the

work has been carried out on a considerable variety of

animals, and though the procedures ‘have themselves been |

most various. The thing that seems to have been effected

in all cases has been the raising or lowering of the gen-

eral metabolism of the treated germs. If this conclusion

be definitely established biology may congratulate itself

that the further and complete analysis of this hereditary

character lies near at hand; is open to definite and easy

attack by methods already of demonstrated trustworth-

iness in this and other fields. And surely if such result

is possible it is timely, now when the ‘‘box within box’’

revival has the sex character, like all others, dissociated

from all processes that can be studied or measured, and

associated with a particle so minute as hopelessly to defy

all direct and functional investigation.

That higher water values in the tissues is associated

in development with increased metabolism is a fact well

established. We need cite here in reference to ‘‘tissue

growth and repair” only the well-known fact of the higher

water-content of embryonic tissues, and Minot’s calcula-

404 THE AMERICAN NATURALIST [Von L

tion that in a particular mammal 99 per cent. of growth

power is lost before birth. In respect to ‘‘heat produc-

tion’’ or the ‘‘basal metabolism’’ of embryo and adult the

data for comparison are not extensive, but it too lends sup-

port to the view that this basal metabolism is higher in

the young than in the adult. It may be added that Ben-

edict and Emmes® have recently shown by very exact

measurements that the basal metabolism of men is higher

by about 6 per cent. than that of women.

If a higher metabolism exists in male-producing germs,

and this is associated with higher water-content, as we

concluded in 1911, it is easy to see why a number of proce-

dures have since been shown to effect a control of the

production of sex. In 1912 Miss King desiccated toads’

eggs and obtained 87 per cent. of females. This was

the converse of the earlier experiments of Hertwig, and

of Kuschekewitch, who ‘‘over-ripened”’ frogs’ eggs—a

process during which they were found to take up water—

and obtained, in the experiments of the latter author, as

many as 100 per cent. of males. I think we can now see

it was a shifting of the metabolism, through the agency

of the water values, that produced the shifting of sex in

the eggs of the frog and the toad.

More recently still, Whitney has effected a change in the

sex of the offspring of the rotifer—Hydatina—a change

from female- to male-production by means which he con-

siders as serving to increase metabolism in the treated

forms. Confirmation of Whitney’s conclusion that it is a

heightened metabolism that brings about male-production

is now to be had in the result obtained by Dr. A. F. Shull!’

who finds that an increased oxygen supply leads toward

an increased production of males in Hydatina. It now

seems clear that a heightened metabolism in the Rotifers

is the agency of increased male-production.

? Benedict, F. G., and Emmes, L. E., ‘ʻA Comparison of the Basal

Metabolism of Norma Men and Wenn, ” Jour. of Biol. Chem., Vol. 20,

No. 3, 1915

10 Akae abstract of a paper to be presented at these meetings, Decem-

ber 29, 1915.

No. 595] SEX CONTROL IN PIGEONS 405

The greater production of males in cattle—indicated

by Thury, Russell, and several others—from eggs that

have remained unfertilized for a period of hours, is al-

most certainly correlated with an increased water-content

which these eggs obtain before fertilization. We do not

know by direct observation that the ova of the cow takes

up water from the fluids that it meets in the reproductive

passages. We do know that this is true for the eggs of

every amphibian, reptile and bird that has been investi-

gated. Von der Stricht has, however, described phenom-

ena in the yolk granules of the extra-ovarian egg on one

mammal—the bat—which phenomena I am quite assured

from my own earlier studies on the yolk spheres, definitely

indicate that in this one mammal in which the data permit

a judgment, the egg does take up water from the fluid that

it meets in the Fallopian tube. There is good reason to

believe that the changed sex-ratios of cattle can be asso-

ciated with changes in the egg-metabolism effected

through, or connected with, differential water values.

The important recent work of Baltzer convincingly

shows the plastic, fluid, controllable and reversible na-

ture of sex in Bonellia. And, it would be difficult to be-

lieve that the larva that attaches itself to the ‘‘riissel’’

of an adult, then quickly and fully differentiates, and be-

comes a male, is not displaying a higher metabolism than

is the larva that rests for long in the mud and sand, and

after prolonged growth becomes a female. Baltzer’s

results deserve a much more extensive statement than

can be given here.

Many points, too, in Geoffrey Smith’s illuminating

studies on sex in the spider crabs would seem to be in har-

mony with the view that the castrated males progressively

lose their initial advantages of a higher metabolism, and

that they then become more female-like as they approach

the lower metabolic levels which are normal to the fe-

males. Though Smith, so far as I am aware, has not

thus interpreted his results. o

The point to these citations is that sex control, in the

several various forms in which it has been accomplished,

406 THE AMERICAN NATURALIST [ Vou. L

has been accomplished fundamentally by the same means

in all—a changed metabolism, in which a higher water-

content of germ and higher metabolism for male-produc-

tion, and lower water-content and decreased metabolism

for female production, have been definitely shown to be

associated in a number of instances. Whitman learned

in pigeon hybridization an additional—an entirely dif-

ferent—means of accomplishing the same end of height-

ening the metabolism of the germ. And, this additional

means definitely tends toward male-production. The

wider the cross (within the limits of the ‘‘developmental

compatibilities’’ of the germs) the greater the vigor and

strength added by the mere act of crossing—and at the

same time the more assuredly will such crosses produce

males. Even the closely related varieties used in most

Mendelian crosses have not failed to indicate the greater

vigor of the heterozygote.

A sixth series of studies has been made on size of the

parents and offspring concerned in these sex-controlled

series. Seasonal and age fluctuations in the parents, and

in both sexes of both parent species; size of offspring as

related to their sex, to season, and to the egg of clutch,

have been studied during three and one half years. We

have found no subject that presents so many complica-

_tions as does the matter of the size of offspring in this

series. Only a single aspect of the matter will be treated

here. The seasonal fluctuation in size of the parents used

in the ‘‘overworked’”’ or sex-controlled series is, how-

ever, a simple matter. Our results show—as indicated

by the lower curve on the chart (1)—that such parents

weigh most in winter and spring; least in the autumn,

reaching a minimum in August and September. In other

words, during the period when the female parent lays

her largest eggs, she herself, and her consort, are smallest

in size. I have had no charts prepared showing the sea-

sonal curves for individual birds, but data for such curves

in great number are available.

_Now, the single word I wish to say on the relation of

size in the offspring to the order of the eggs of the clutch,

No. 595] SEX CONTROL IN PIGEONS 407

and as affected by the procedure of overwork, may be

more quickly said with the aid of the charts.

One chart (only Chart 2 is reproduced here) shows the

weight average of each individual hatched during the

year, from two simultaneous matings of alba X risoria.

Series I is from an older pair, previously overworked;

series II is from a younger pair, little—or not at wie

previously overworked. It will be noted that series I is

throwing large eggs, a predominance of females, and that

the size of the offspring—even of the males—is prevail-

ingly that of the females rather than the males of the

parent species.1: Series II is throwing smaller eggs, a

nearly equal proportion of the sexes, except at the end

of the season, and the size of the offspring is decidedly

larger than in series I; and, in fact, approximates to the

size of the males of the parent species. In both of these

Series it will be observed that size of offspring?? is also

correlated with the order of eggs in the clutch.

For series I, we have complete data for the year pre-

ceding and the year following the term covered by the

chart. The weights for the former were: Av. for ¢’s

172 gr.; 2’s 166 grams. For the succeeding year—early

1915— these weights are &’s 157 gr.; 2?’s 156 gr. Clearly,

during the three-year period a change in size of offspring

is progressively occurring; and the change runs from a

size comparable to that of the males of the parent species,

to a final size that is somewhat below that of the females

of the parent species. The egg-size was known in this

same series to have progressively and simultaneously

changed from greater male-producing tendency to a de-

cided female-producing tendency.

The seventh line of study intended to analyze the sea-

sonal and clutch deliveries of the sex-controlled series is

concerned with arrangements by which the sex-behavior

of the birds from such series is tested. In these pro-

11 The males, in both of these species, average 10-15 grams heavier than

the Teenie: m risoria birds are slightly larger on the average (5-10

grams) than

12 The rE given for individual birds — I the average of the

monthly, or bi- a wager for the year. r

408 THE AMERICAN NATURALIST [ Von. L

cedures female is mated with female, and male with male.

Such pairs—from a very few selected pairs of parents—

are kept mated for a period of six months. The three

and one half years that this study has been pursued has

enabled us—using 30 to 50 birds—to test one and the

same bird with seven others. Most of the birds used—

for lack of success with the incessantly fighting males— ~

have been females, and most of the seven successive tests

with each bird have been made with its sisters of the same

series. The members of the pairs are kept apart except

when under observation; when put together—as they are

twice daily—the records are taken of those females of

the pair which behave as males in copulation with their

mates. Three facts are definitely established by the data

obtained: (1) The females of the orientalis X alba cross

(they are dark in color) are more male-like in their sex

behavior than the females of the reciprocal cross (these

are white in color). (2) Females hatched from eggs laid

earlier in the season are more masculine in their sex be-

havior than are their own full sisters hatched later in

the season. And, several grades of females can be thus

seriated according to season of hatching. (3) The female

hatched from the first egg of the clutch is more masculine

than her sister hatched from the second of the clutch

in a great majority of the cases. And in nearly all these

latter matings the more masculine bird is so decidedly so

that she takes the part of the male a full 100 per cent. of

the time in copulating with her very feminine clutch-mate

sister.

A fuller account of this situation was given, with the —

assistance of charts too large to exhibit or describe here,

before the Society of Zoologists in 1913.12 The nature

of this behavior has been adequately recorded by means

of moving-picture films. Such records were also made

showing the reversal of the known sex-behavior of such

pairs by means of appropriate injections of ovarian and

testicular extracts. Those films were demonstrated in

this hall—or in one near-by—in connection with an ad-

18 Abstract in Science, March 20, 1914, - :

No. 595] SEX CONTROL IN PIGEONS 409

dress before the local chapter of Sigma Xisome 20 months

ago. )

The injection of the extracts of gonads, performed

now on the third series of birds, has resulted—quite

against our wish—in the death of a number of birds. In

the main the deaths from ovarian injections were of the

more masculine birds; while the deaths from testicular

injections have been among the more, or most, feminine

birds. The numbers concerned at present are not large,

and a further definite study of the matter will be made

before final conclusions are drawn. But the limited data

now at hand indicate that the eighth correlation listed

on Chart 1 is as it is exhibited there.

A ninth, and very accurate and convincing kind of in-

formation concerning the germs involved in these sex-

series has been obtained by means of the bomb calorim-

eter. The heat of combustion of some 200 egg-yolks

has been determined. One such series of determinations

for 1914, in which all available eggs were burned, is

shown on Chart 4. It will there be seen that the first

clutch of the season bore a higher caloric value than the

second, but is otherwise the smallest of the year. Be-

ginning with the second clutch laid in June, the succeed-

ing clutches to December 1 bear higher and higher heat —

values. In all clutches too, except the very first, the

second eggs show a higher storage of heat units than

do the first of the clutch. Here we find the conclusions

reached from studies on the weights of yolk, and on yolk

analyses, fully confirmed by a study of the burning value

of the materials stored. And confirmed by a method in

which the error involved in the determination is wholly

negligible. The most accurate method, for the study of

the storage values of male- and femaleproducing ova,

gives too the results most consistent with the breeding

ata. : i ‘

The tenth and last of these correlations deals with

embryological or morphological data. Itwas found that

some females dead at relatively advanced ages showed

persistent right ovaries. The right ovary in pigeons

410 THE AMERICAN NATURALIST [Vou. L

normally begins degeneration at or before hatching and

is wholly absent from the week-old squab. It soon be-

came evident that the persistent ovaries were found prac-

tically exclusively in birds hatched from eggs of over-

worked series. Further study has shown in addition that

they arise almost wholly from the eggs of autumn, and

predominantly then from the second egg of the cluteh—

that is from eggs otherwise known to have the greatest

or strongest female-producing tendency. These ovaries

have sometimes weighed nearly a third as much as the

adult left ovary with which they were associated, and

have been found in such birds dead at all periods from a

few days to fifteen months. We here attempt no ade-

quate description of this situation, but one can not have

observed the frequency of the persistence of this ovary

in the birds hatched from the eggs otherwise known to

be the most feminine from these overworked series with-

out conviction that the same pressure which carries the

eggs of spring from male- producing to female-produc-

ing levels, also carries the earlier female-producing level,

to another yet more feminine.

In conclusion, the studies that have thus far been made

on sex, and on the experimental control of sex, in pigeons

go very far, we believe, toward an adequate demonstra-

tion that germs prospectively of one sex have been forced

to produce an adult of the opposite sex—that germs nor-

mally female-producing have, under experiment, been |

made to develop into males; and that germs which were

prospectively male-producing have been made to form

female adults. That neither selective fertilization, dif- —

ferential maturation nor a selective elimination of ova

in the ovary can account for the observed results. Fur-

ther, and perhaps of more importance, these studies

throw much new light on the nature of the difference be-

tween the germs of the two sexes. This difference seems

to rest on modifiable metabolic levels of the germs; males —

arise from germs at the higher levels, females from the

lower; and such basic sex differenack are quantitative, oa

rather than qualitative in kind.

THE CALCULATION OF LINKAGE INTENSITIES!

Proressor R. A. EMERSON

CoRNELL UNIVERSITY

Two methods of estimating the intensity of linkage are

in use. One consists of crossing individuals heterozy-

_ gous for two or more linked genes with homozygous re-

cessives. This is the more direct method, because the

gametic ratio—barring differential viability—is exhibited

directly by the zygotic frequencies. The other method

employs ordinary F, ratios derived from selfing F, or

breeding together like F, individuals. Here the gametic

ratio can only be inferred from the numerical relation of

the zygotic classes. The results may be disturbed not

only by differential viability, as in the first method, but

also by selective fertilization, if that occurs, and may

often be materially influenced by chance in random mat-

ing where the numbers are small. In fact, this method

is so undesirable that it should not ordinarily be used

where the other method is practicable. It is true, how-

ever, that the mechanical difficulties of crossing certain

plants are so great and the number of seeds produced

per flower so small that often the ordinary F, results are

alone available. It is important, therefore, to have a

means of calculating gametic ratios from F, zygotic

numbers.

Since no direct formule for calculating gametiec ratios

from observed F, data have heretofore been available,

the problem has been attacked in an indirect way. A

series of F, zygotic ratios has first been calculated from a

corresponding series of gametic ratios. Next the ob-

served F, results have been compared with the calculated

series, the closest fitting calculated ratio determined, and

the corresponding gametic ratio taken as that responsible

for the observed F, results. - aoe

1Paper No. 54, Department of Plant Breeding, Cornell University,

Ithaca, N. Y. ea a

411

412 THE AMERICAN NATURALIST [ Vor. L

The method of determining the closeness of fit between

calculated and observed numbers used by Bateson, Pun-

nett and their co-workers was mere inspection. (See

Bateson and Punnett, 1911.) The unreliability of this

method was pointed out by Collins (1912) who made

use of Yule’s coefficient of association for the same

purpose. The well-known formula for this coefficient is

(ad — be)/(ad + be), where a, b, c, d are the frequencies `

of the phenotypic forms AB, Ab, aB, ab, respectively.

From a table giving the coefficients of association for a

series of gametic ratios, the best fitting gametic ratio

is chosen by inspection or interpolation. This method is

satisfactory except for the higher gametic ratios where

slight differences in the coefficients of association corre-

spond to wide differences in the gametic ratios. Since

the same intensity of linkage gives somewhat higher

coefficients of association for coupling than for repulsion,

particularly for the lower linkage values where the asso-

ciation coefficient method is most reliable, two tables must

be used.

Formule, by which gametic ratios can be approximated

directly from F, data without the use of coefficients of

association and without respect to whether coupling or

repulsion is involved, would seem to merit trial. Such

formule are presented later in this paper. Moreover, it

is often desirable to reverse the calculation, that is, to

determine zygotic frequencies from assumed gametic

ratios. A single formula suggested for this purpose —

gives accurate results for both coupling and repulsion.

This formula will be presented first because the others

are developed from it.

Bateson and Punnett (1911) snggested two empirical

formule for calculating zygotic frequencies from assumed

gametic ratios, one for coupling and the other for repul-

sion. Neither one, of course, is applicable to both types

of linkage, though both formule are true for independent

inheritance. If A and a are allelomorphic genes and Bo

and b are a similar allelomorphie pair—the capital letters a

No. 595] LINKAGE INTENSITIES 413

denoting dominance—and if 2n equal the sum of the

gametic series,? then the gametie series and the pheno-

typic zygotic series, AB, Ab, aB, ab, for coupling and for

repulsion are:

Gametic Series

$ Ab: eB

oupling

Ropuliön 622 rss. vee ee 1: n—1:n— 1: 1

eee Series

Ab T

Coupling .. 3n2— Ck 1): ie + 1: 2n— 1: n2 — (on 1)

Repulsion .. 2n24-1 :n2—1:n2—1:

That is, the formule of Bateson and Punnett are ex-

pressed in terms of the sum of the gametic series. But

the same thing can also be expressed in terms of the

several members of the gametic series. Thus, if r:s is

any gametic ratio, the usual form of gametic series is

r:s:s:r and the frequencies of the ten possible genotypic

classes and of the corresponding four phenotypic classes

are:

Genotypes Phenotypes

AB.AB—= r2

AB.Ab — 2rs

AB.aB = 2rs \ AB = 3r? + 4rs + 282

AB.ab — 2r2 :

Ab.aB = 2s?

Ab.Ab = 8? s :

Ab.ab D, E A

aB.aB =

aB.ab Fore fOB = ore + ;

ab.ab =r2 bab =r?

The general formula for calculating a phenotypic

zygotic series from a given gametic ratio is, therefore, |

3r? + 2(5? + Qrs) ss? + 2rs:s? + 2rs:1? (I)

The sum of the zygotic series is 4r?-+ 8rs + 4s? or

(2r + 2s)?, which, when expressed as

(rts+str(rts+sts+r), —

2 Bateson and Punnett considered n to be some power of 2, but this

limitation need not apply here.

414 THE AMERICAN NATURALIST [ Vou. L

indicates how the formula is derived. Reference to the

diagram will make this clear. Since r and s are any

positive quantities, formula I is applicable to coupling

r S S r

AB Ab aB ab

Titi ti i OE te eS id Se Ge a oe

LELI LITIN Serie Bae se:

- LELEII H ipe e memen m

r EY FHF rs HIRT rs EEE r2 HH

AB IHABAB HIH AB Ab ABaB HHIH ABab 4H}

CEEE TE PR EE A

| (8 @ WAS: i | 2S GS ses

E Be 8 ee Le FETE LI

Í L

| Ci

| HI

s eng rs HW 2 HF S E rs —

Ab IHF AbAB HHI AbAb QAT AbaBiGi— Abab -4

+- i

i L

i eS oe ees

ji TILLI

g * HHHH | Lil

rr 28 oe F TE s? rs

aBlHt+aBA taB.Ab +H- aB.aB aB.ab

HH LILL LL

| |

r IHE 7? 4A rs = pe

ab [Hab AB HHI ab Ab = abaB abab

HH LLL

DIAGRAM SHOWING IN TERMS OF r AND 8 THE NUMERICAL RELATIONS OF THE

F2 ZYGOTIC CLASSES THAT RESULT FROM COMBINATIONS OF THE GAMETIC CLASSES

B, Ab, aB, ab OCCURRING IN THE RATIO SERIES r:8: The dominant genes

and B are indicated by horizontal and vertical ice panier; while their

silattaleette a and b are indicated by the absence of such lines. (See formula I.)

(r>s), repulsion (r<s) and to independent inherit-

ance (r=s). It, of course, gives the same result as the

empirical formule of Bateson and Punnett, but is more

convenient in that one formula takes the place of the

two. It is easy to use since the fourth term of the

zygotic series is the square of r, the second and third

terms each the square of s plus twice the product of r and

No. 595] LINKAGE INTENSITIES 415

s, and the first term the sum of the second and third plus

three times the fourth.

An approximation of gametic ratios can be obtained

from observed zygotic ratios by simple formule derived

from formula I. If the actual values of s? + 2rs could

be assumed to be identical in all cases, it would follow

from formula I that 4r? — AB + ab — (Ab + aB) and r

—yV(AB + ab— Ab—aB)/4. Similarly, 4(s? +2rs)

—AB+Ab+aB—3r? and s=V(AB+4b+aB-+1") /4—r

=V (AB + Ab +aB +ab)/4— r. If E is the sum of the

extreme terms and M the sum of the middle terms of the

observed zygotic series, the formulæ for approximating

gametic ratios are, then,? he

r= 5VE—M (II)

s=dVE+M-—r

If it is desired to compare the observed F, frequencies

with a caleulated series of frequencies, the procedure,

obviously, is to calculate the gametic ratio by formule Il

—or by means of the coefficient of association—and then

to calculate the zygotic series by formula I—or by one of

the two formule of Bateson and Punnett. This procedure

is not always necessary, however, for a theoretical zygotic

series can usually be readily computed directly from the

observed frequencies. If AB, Ab, aB, ab is the series to

be calculated from the observed frequencies, it follows

from formule I and II that :

Ab = aB=M/2 i

"wE Ma (IIT)

AB=M+3ab ae

Since a zygotic series calculated in this way necessarily

meets the conditions imposed by formula I, the gametic

ratio can be approximated from it more readily than from

the observed frequencies. Since by formule I and It

ab=r? and s=5VE+M—r, ~

r= Vab

amcanai

p> (IV)

"TO

L—

‘*Since r and s are necessarily positi

416 THE AMERICAN NATURALIST [ Vou. L

Formule IV are not to be used in connection with ob-

_ served F, frequencies except when the latter approximate

closely the form demanded by formula I, that is, when the

first term of the observed freq ies equals approxi-

mately the sum of the second and third terms plus three

times the fourth term.

In cases of repulsion, where the fourth term of the

zygotic series is always relatively small and, therefore,

where the first term should be only slightly greater than

the sum of the second and third terms, it may happen that

the sum of the first and fourth terms, E, is actually less

than the sum of the second and third terms, M. In such .

cases, formule II (and consequently formule III and IV

also) can not be employed, for, if E is less than M the

quantity under the radical (E — M) is negative and has

no real root. In such cases, the gametic ratio must be

calculated by means of the coefficient of association.

The method here suggested for calculating gametic

ratios from observed frequencies never gives quite the

same results as that obtained by the association-coefficient

method except when the observed series approaches

closely the form demanded by formula I. Naturally,

then, the more widely the observed frequencies depart

from this form the greater the difference between the

results given by the two methods. Since the coefficient of

association gives reliable results if the tables to be used

with it are based upon sufficiently small differences in the

gametic ratios employed in its preparation, it follows that

the methods proposed in this paper give only approxi-

mate results. It is also true, therefore, that the nearer

the observed frequencies approach the form of formula I,

- the closer the approximation obtained by formule II (or

II and IV). :

The two methods have been applied to numerous cases

taken from published accounts of linkage studies and the _

goodness of fit tested by the method suggested by Harris”

(1912). The differences, o — c, between the observed fre-

quencies, 0, and the calculated frequencies, c, of the sev-

No. 595] LINKAGE INTENSITIES 417

eral classes are determined and §[(o0 — ¢)?/c] = æ? calcu- `

lated, S indicating summation.

With n, the number of classes, here equaling four, and

x, the probability, P, that departures from the calculated

series as great as those observed might occur through

the errors of random sampling, is obtained by reference

to Elderton’s (1901) table (see also Pearson, 1914).

Wherever appreciably different gametie ratios have been

obtained by the two methods, P has been found to be

greater for the association-coefficient method than for the

method based on formule II. The former method has,

therefore, given the closer fit. Since, in most of the cases

to which the test has been applied, 2? is less than one and

since such values are not listed in Elderton’s table, z? has

been used directly for the comparison of the two methods.

Where n is constant, the larger x? the less the probability.

While this test for goodness of fit has shown the asso-

ciation-coefficient method to be the better of the two, the

fact that in most cases x? was less than one for both meth-

ods indicates that the approximate method suggested here

ordinarily gives results such that the departures of ob-

served from calculated frequencies might well be due to

errors of random sampling. The method has been found

convenient and usually sufficiently accurate where only an

approximate determination of the gametic ratio is de-

sired. Where the observed frequencies depart widely

from the form given by formula I, this method should not

be used. It should be noted, however, that in such cases

no calculated series fits the observed results well. This

limitation to the use of the new method does not lessen

materially the convenience of using it where it is appli-

cable. By a mere inspection of the observed frequencies,

it can usually be told whether they conform fairly closely

to formula I, that is, whether the first term is approxi-

mately equal to the sum of the second and third plus three

times the fourth. a 2

A few examples will illustrate the use of the approxi-

mate method of calculating gametic ratios from observed

418 THE AMERICAN NATURALIST [ Vou. L

data and afford a means of comparing it with the associ-

ation-coefficient method.

Harris (1912) has quoted an example of coupling in

sweet peas from the studies of Bateson, Saunders, and

Punnett* and calculated P where the gametic ratios are

taken as 7:1 and 15:1, the only ratios considered in the

original paper. The phenotypic classes are based on shape

of pollen and color of flowers and the observed frequen-

cies are purple long 493, purple round 25, red long 25,

red round 138, total 681. As determined by Harris, on the

basis of a 7:1 gametic ratio, P=.0053 or t = 12.7699.

On the 15:1 basis, P = .3086 or 77 = 3.6375. The chances

against the 7:1 ratio are, therefore, 199 to 1 and against

the 15:1 ratio about 2 to 1. For this same material, Col-

lins (1912), using the association-coefficient method—

Coef. Assoc. == .982 + .004—naturally suggested a 12:1

gametic ratio—Coef. Assoc. also = .982—and pointed out

the fact that the deviation from the 7:1 ratio is 9 times

and from the 15:1 ratio about twice the probable error.

By formule III, the calculated series becomes 485.75

+25.0+25.0-+-145.25—681. By formule IV,r= 12.052 and

s==.996 or a gametic ratio of 12.1:1. The 12:1 ratio ob-

tained by the association-coefficient method gives a zygotic

series of 485.5 + 25.2 + 25.2 + 145.1—681. Both meth-

ods, then, give gametic ratios approximately the same and

practically identical zygotic series, namely, 485 + 25

+25+145. On the basis of this series, 2.4387 and

P is so large that it is useless to determine it. In short,

both methods give gametic ratios that fit the observed

data extremely well.

The next example of coupling presents a very different

condition. It has been quoted by Bridges (1914) from

Punnett’s (1913) summary of reduplication series in

sweet peas. The phenotypic classes are based upon ster-

ility of anthers and form of flowers and the observed

frequencies are fertile normal 165, fertile cretin 58,

sterile normal 58, sterile cretin 78, total 359. It can be

seen at a glance that these frequencies are far from

4 Rept. Evol. Com., 4: 11.

No. 595] LINKAGE INTENSITIES. 419

what formula I demands—58 + 58 + 3(78) =350, over

twice 165—and that therefore the approximate method

can not be depended upon in calculating the gametic

ratio. It is interesting to note, however, just how un-

reliable it is in comparison with the association-coeffi-

cient method. By formule III and IV, the calculated

zygotic series becomes 211 + 58 + 58 + 32=359, r—5.6,

s=3.8, and the gametic ratio is approximately 1.5:1.

Bridges referred the case to a 2:1 ratio (Coef. Assoc.

= .558), though the coefficient of association is .588 which

is equivalent to a gametic ratio of 2.1:1 (Coef. Assoc.

=.586). Punnett compared the observed frequencies

with a series derived from an assumed 3:1 ratio. The

zygotic series calculated from these ratios are, for the

2:1 ratio, 219 + 50 + 50 + 40—359; for the 2.1:1 ratio,

220 + 49 + 49 + 41 — 359; and for the 3:1 ratio, 230 + 39

+ 39+ 51—359. If now the criterion of goodness of fit

be applied to the four calculated series the values of x?

are, for the 1.5:1 ratio 76.1, for the 2:1 ratio 52.0, for the

2.1:1 ratio 51.4, and for the 3:1 ratio 51.3. Values of x?

above 30 are not listed in Elderton’s table, but where

£? = 30 and n==4, P=.000,001, which means that there

is only one chance in a hundred thousand of deviations so

great as the observed ones being due to the errors of

random sampling. Where neither of the two methods of

calculating the zygotic series gives a better fit than in

this case, it is immaterial which fit is the worse. ae

As an example of repulsion, the same characters, in

sweet peas may be used. The observed frequencies

(Bateson and Punnett, 1911) are 336+ 150 + 143+ 11

=640. Bateson and Punnett assumed that the gametic

ratio concerned was 1:3. The coefficient of association

is —.706, which is equivalent to a gametic ratio of 1: 2.74.

By formule II-I or IIL-IV, a ratio of 1:2.45 is indicated.

The values of x? are for the 1:2.45 ratio .649, for the

1:2.74 ratio .302, and for the 1:3 ratio .536. Here again

the association-coefficient method gives the better fit, but

the probability is great that the deviations of the ob-

Served from the calculated frequencies, even in ease of

420 THE AMERICAN NATURALIST [ Von. L

the approximate method, might be due to errors of ran-

dom sampling.

As an illustration of the fact that the approximate

method can not be used in some cases of repulsion, even

when the observed frequencies fit fairly well the series

calculated by the association-coefficient method, an ex-

ample of linkage between dark axils and fertile anthers

in sweet peas quoted from Punnett by Bridges (1914)

may be taken. The observed frequencies are 1335 + 643

+ 714+ 22694. The value of r can not be determined

by formule II nor by III and IV, because 1335 + 2 — (643

+714) is a negative quantity (— 20) and has no real

root. The coefficient of association is —.988, which is

equivalent to a gametic ratio of 1:17, though Bridges

assumed a ratio of 1:20. On the basis of this 1:20 ratio,

æ? = 5.68 and P—.1309. On the basis of the 1:17 ratio,

v?=4.04 and P—.2615, or odds of about 3 to 1 against

the occurrence of deviations as great as those observed.

It may be said, then, that the formule suggested here

afford a convenient method of approximating gametic

ratios from zygotic series, when the observed frequencies

are in fair accord with a series based on formula I—or

the formule of Bateson and Punnett. When the ob-

served frequencies are far from this type no method

gives a close fit between observed and calculated results.

LITERATURE CITED

Bateson, W., and Punnett, R. C. On Gametic Series Involving Reduplication

of erii Terms. Jour. Genetics, 1: 293-302. 1911.

Bridges, Calvin B. The Chromosome Hypothesis of A Applied to

Cases in Sweet Peas and Primula. AMER. Nart., 48: 534, 1914.

Collins, G. N. Gametie Coupling as a Cause of sch tne ate NAT.,

46:. 569-590. 1912.

Elderton, W. Palin. Tables for Testing the Goodness of Fit of Theory to

Okservation. Biometrika, 1: 155-163.

Harris, J. Arthur. A Simple Test of the Goodness of Fit of Mendelian

tios. AMER. NAT., 46: 741-745. 1912

Pearson, Karl. Tables for Statisticians sid Biometricians. Cambridge

University Press, Table XII, pages 26-28. 1914.

Punnett, R. C. Reduplication Series in Sweet Peas. Jour. Genetics, 3: TT-

103. 1913. if

THE MECHANISM OF CROSSING-OVER. IV

HERMANN J. MULLER

CoLUMBIA UNIVERSITY

Tue ‘‘map’’ of the first chromosome, based on these

experiments, is shown below:

YwA b cL

0 2355% 3 39- 3

Fig. 10. Map of chromosome I.

The figures represent the distances of the factors from

yellow, the first one in the line, and are calculated merely

by adding together the intermediate distances. This map

gives almost exactly the same proportionate distances be-

tween the different loci as does that obtained by com-

bining the results of linkage experiments performed by

other workers, in which usually the inheritance of only

two or three factors was followed at one time. Each set

of ratios, therefore, confirms the accuracy of the other.

The absolute distances in the present map are, however,

somewhat shorter, being % the length of those in the com-

posite map. This was caused mainly by the comparatively

large number of non-cross-overs produced by a few fe-

males; in the rest, the crossing-over frequencies were

about normal. It may, therefore, be concluded that chro-

mosomes which differ in regard to eleven pairs of factors

behave in the same way, so far as crossing-over is con-

cerned, as those which are alike except for two factors.

This is contrary to a suggestion made by Punnett. More-

over, the fact that chromosomes differing in so many fac-

tors behave normally is here especially noteworthy, be-

cause 11 of the 12 recessive factors were in the same

chromosome.

The results of the experiments with the second chromo-

Some may now be tabulated. 462 offspring of females

421 .

422 THE AMERICAN NATURALIST [Vou. L

heterozygous for the ten mutant factors used in this group

have been recorded. The table only gives the result with

respect to nine characters, however, as arc wing was not

followed in all of the experiments. (The data given later

as to its position have, accordingly, not been calculated

from quite as large a count of flies as have the data for the

other factors.)

CLASSIFICATION OF Factor COMBINATIONS TT BY FEMALES

Str bi pu Vgars

daj

HAVING THE COMPOSITION:

Cy A

| Streak 4 Not Streak | Total

Non-cross-overs

| Sbpvas 68 | djcb 82 | 150

Between Single Cross-overs

i BO Oe ee bd d oba: AT b 26

Ge Sn he ee ee E Fs Wo 24 db pos 19 43

ON ei a Sbe d 1? d 1?

P80 Dee a a a Sb. 6, -$ d jp s 6 9

Pi BH visa Sbp cba 14 djvs 20 34

Vg and cy Sb pm cba 10 dj esh 1

Oy ONO Spe ss cy E A B Upe Da 51 d3 c8 - 60 101

mae Desa rer eS, Sbpvsh, O dj c 0

Double Cross-overs

Str and da; pu and vg......... Sdj vs $ bpeba 1 oe

Sır and da; Vg atid Gy. 6.22.5. Sdj 1 b pocba 0 1

Str and da; cy and sp......... Sdj 5 b poba 8 13

Ge BHO. Das) ad Geo eo. Sj ms 1 Yo D 1

da and bis pa and vees.. oL. Sj vs 5 db p cba 5 10

dand -bi; vg and cy.......... Sj g 5 db pucbh 1 6

da and bi; cy and Bp... ....... Sj ces 5 poba 8 ER

j and pe; pu anid ve.. o 5... S bos 1 d jpa 1 2

jand po; vi and or.: uaaa 0 d jpvocba 1 1

j and pa; Oy and Sp... nei S b cs 7 d jpvba 1 8

Pu and vg; vg and cr......... S bp s 0 dj veb 2 »

Pu and Vg; Gy and Bp. -ses oii S ‘tp ce — 3 dj tbe 3 eee

e and Ov; cy and sp.......... S bpves 2 di be i ji

Be and de; dean bi: oranda | (Rite, 4

Str and da; j and pu; cv and sp . ube atone a b- ves 1

da and bisj and pu;puandvg..| Sj peba 1 et

d i db 1

da and bi; pu and vz; ev and s O EE db p cs 1

j and pu; pu and vg; cy and Sp.

No. 595] THE MECHANISM OF CROSSING-OVER 423

Total Single, Double, and Triple Crossing-over

Observed er Cent, 0

Between Number Crossing-over

i MNO Ga PAE a ee 45 9

Oe O8e i oooi eee Ti. 16,7

Be MOA Recah nce Gt cea 1? 0.29

$ GN Pu ei ce au dee eee: 25 5.4

Pie (AU Ves ie T e 59 12.8

Ve GU Oy a Ee aes 34 ga

ie ANG Bp) TOPE ae Oe 150 32.5

Mas ONG Pa nci cam cand eae S 0 0.0

In the case of this chromosome, too, the law of linear

linkage is graphically illustrated by the characteristic

‘‘sectional’’ mode of interchange between the groups.

The non-cross-overs here constitute only 32.5 per cent. of

the population, whereas the single cross-overs make up

51.1 per cent., the double cross-overs 15.2 per cent., and

the triple cross-overs 1.3 per cent. In making a map of

this chromosome, the chances of error are greater than in

the preceding case, since not so many flies have been ob-

tained. Nevertheless, the values correspond very closely

with estimates of the results obtained in other work, al-

though figures exactly representing the sum total of other

work are not just now available for comparison.

Ste da bij ps : vs Cr or __ Saba

" e us OM O e a =

Fic. 11. Map of chromosome II.

Let us now construct a curve showing the frequency

with which, in the experiment with the first chromosome,

points various distances apart showed coincidence of

crossing-over. Suppose that in this curve the horizontal

line represents the distance apart of the two coincident

cross-overs, and the vertical line the per cent. of cases in

which double crossings-over at such distances occur. For

example, if it were known that double crossing-over for a

distance anywhere between 15 and 16 units occurred in .2

per cent. of all cases the height of the curve above the

figures 15 and 16 would be made .2 vertical units. Now, —

each case of double crossing-over. that actually happens a

424 THE AMERICAN NATURALIST [Vou. L

among the 712 flies obtained for group I must represent

Ap or .14 per cent., of all the cases. If, then, a crossing-

over is found to occur somewhere between cı and v, and

one occurs coincidentally between s and r, the two points

of crossing-over may have been as far apart as cı and r

(36), or as close together as v and s (8), or at any inter-

mediate distance. Therefore we have no right to make

this case stand, in the curve, for a coincidence that hap-

pened at a particular distance (say 10-11) and to raise

the ordinates for this particular distance by .14 units.

Each distance between 8 and 36 is consequently given

partial credit in our curve for the occurrence of this coinci-

dence, and so each of the 28 ordinates between 8 and 36

is raised to an average height of ee =.005 approximately.

All the other cases are treated in a similar way, and thus

the curve shown by the heavy line in Fig. 12 is obtained.

An n Me

wy P N

# %

/ `,

Tà N,

’

/ N,

1 i

2 A

<OZMCOMIT

0- 10 20 30

Fig. 12, Curve showing the observed frequency of double crossing-over

in chromosome I, for points various distances apart. The dotted line shows

the frequency expected on pure chance. ;

No. 595] THE MECHANISM OF CROSSING-OVER 425

But although these ordinates are, on the average, raised

by this amount, each one is not raised equally, for there

is less chance that double cross-overs should have the

most extreme possible values than medium values. The

total addition of .14 units to the curve should hence be dis-

tributed among the different possible ordinates according

to the relative probabilities that the two points of crossing-

over should have been the distance apart represented by

these respective ordinates. These various probabilities

for the different ordinates, in the case of any specific

double cross-over, may be represented in the form of a

curve, and the main curve of double cross-over frequency

shown in figure 12 is thus really a composite in which these ~

individual curves for each double cross-over have been

added together. We may now consider the way in which

the individual curves of probability are calculated.

Let us take the case of the double cross-over that

occurred between c, and v and coincidently between s and

r. We have already calculated that the distance between

the two points of crossing-over must be somewhere be-

tween 8 units and 36 units (see second paragraph above).

The curve for this individual double cross-over will there-

fore start at 8 on the abscissa and continue to 36. What

height shall it have along the ordinates between these

points? Let the region c, — v be divided into 8 equal parts

—abedefgh—of two units each, as shown below.

13 29 37 ' 49

a — 16 ‘units — vy —-8 un.—s°—12-un.-, r

ee oe BS S a

PTA ht see

“bede igh

It will be seen that a double cross-over of 8 to 10 units

length (i. e., having 8 to 10 units between its two points of

crossing-over) which passes between the factors ¢, and v,

must go between them in the region h, if its other point of

crossing-over is to be between s and r. However, any

double cross-over of 10 to 12 units length which passes”

through either g or h will also pass between s and r, and

so there is twice as much chance for double cross-overs of 3

426 THE AMERICAN NATURALIST [Von L

this length to occur as for those 8 to 10 units long. Sim-

ilarly, those 12-14 units long may be three times as numer-

ous, for they may pass through f, g, or h, and so with each

increment of length, up to 20, there will be an equal addi-

tional amount of chance for a double cross-over of that

length (passing through the required sections, c, — v and

s—r) to occur. Thus our curve of probability rises in

regular steps from 8 to 20; if we could have divided the

distance c, — v into an infinite number of parts, instead of

into 8, these steps would each be infinitely small, and so

we should have a straight line rising from 8 to 20.

Beyond this point the rise in probability ceases; a

double cross-over between 22 and 24 units long has no .

more chance of happening than one of 20-22 units. Refer-

ence to the figure will show that a double cross-over of

20-22 units passing through any of the regions from c

through h will separate s from r and thus fulfill the re-

quirements, but a double cross-over 22-24 units long,

while it has the additional alternative of passing through

b, can not- pass through h without its second point of

crossing-over falling to the right of section s—r. Sim-

ilarly, one 24-26 long may not pass through g or h, though

it may pass through any region from a ta f; double cross-

overs of all these lengths therefore have the same chance

of occurring, and our curve along the corresponding ordi-

nates would hence be a horizontal line.

Double cross-overs longer than this would have less and

less chance of occurring; one 26-28 long could only pass

through regions a — e, one 28-30 only through a — d, and

so the curve falls again in a straight line to the zero level

at 36.

The same rules can be shown to apply to all cases: the

curve starts at a place on the abscissa representing the

distance apart of the innermost factors involved (in the

above case this distance was v — s, —8); it rises in a

straight line for a distance equal to the length of the

smaller section involved (above, this was the distance

S —r, = 12, so that the line rose to point 8 + 12, 20);

No. 595] THE MECHANISM OF CROSSING-OVER 427

it then proceeds horizontally until a distance from the

starting point of the curve equal to the length of the longer

section has been passed (above, this was the section

cĉ, — V, = 16; thus the line proceeded on a level to point

8 + 16 = 24°) ; then it falls in a straight line to a point on

the abscissa representing the distance between the outer-

most factors involved (above, the distance is c, — r, = 36).

The height to which the curve rose is determined by the

fact that its area (the sum of all the ordinates) must have

a value representing the per cent. of total cases in which

such a double cross-over occurred (above, each double

cross-over must have a curve with an area —.14, since

each fly was .14 per cent. of the total count).

It will be noted that for each individual curve the prob-

ability is calculated on a basis of pure chance, no account

being taken of possible interference, which, if present,

would tend to make the longer distances more likely than

the shorter, and so to raise the right end of the curve at

the expense of the left. In other words, each individual

curve represents the frequency with which double cross-

overs of different lengths would happen within the partic-

ular regions dealt with (in our case above, regions cĉ, — V

and s — r), if there were no interference and they had a

purely chance distribution, within these regions. . The

composite curve thus errs rather by showing too little

effect of interference than too much. All interference

which it does show—that is, all deviation between it and a

curve representing an entirely random distribution of

double cross-overs—must then be due solely to the way

in which the double cross-overs were found to be distrib-

uted among the various regions, as no assumption of

interference was made in calculating out the curve for

each double cross-over.

The curve representing the proportion of double cross-

overs of different lengths which would haye been found on

an entirely random distribution (no interference) 1s

€ The di een this figure (24) and that (26) found by the

method enipe raie BY ai : the region ¢,—v had been di-

vided infinitely instead of only into eight parts. |

428 THE AMERICAN NATURALIST [Yorn L

shown by the dotted line. To make comparison with the

other curve legitimate, it had to be constructed by the

same method,—namely, by making a composite of indi-

vidual curves, each of which represented the probabilities

for a certain type of double cross-over—only, instead of

using the observed numbers of double cross-overs of the

different types, in constructing it, it was necessary to use

the numbers of double cross-overs of the different types

that would have been observed if there had been no inter-

ference. (This curve hence represents the results of a

chance distribution both among and within the various

regions.) In the case of each type of double cross-over,

the way to find the per cent. of individuals showing it that

would be produced if there were no interference, is to

multiply the total per cent. of crossing-over in the first

region by the per cent. in the second region, as explained

in section 4a. (Thus, the per cent. of double cross-overs

passing between A and B and between C and D equals per

cent. of cross-overs between A and B times per cent. of

cross-overs between C and D.) This per cent., then multi-

plied by the total number of individuals counted, gives the

number of such double cross-overs theoretically to be ex-

pected in the absence of interference. When such calcu-

lations for each different possible kind of double cross-

over have been made, and the individual curve for each

then made, the latter may be combined to form a com-

posite curve like the curve shown by the dotted line.

The end desired is of course to compare the dotted and

the heavy-lined curves and see what proportion of the dou-

ble cross-overs various distances apart, that were expected

on pure chance, actually occurred. Therefore a new curve

(Fig. 13) may be made, representing this relative coinci-

dence, i. e., the per cent. which each frequency on the ob-

served curve formed of each frequency on the expected

curve (see sect. IVa). This curve consequently shows the

rise or fall of the index with which we are already famil-

iar, and which we have called simply ‘‘coincidence.’’

Owing to the fact that not very large figures have so far

No. 595] THE MECHANISM OF CROSSING-OVER 429

been obtained, we must be cautious about accepting the

exact values shown in the curve of coincidence; this

applies not so much to the main portion of the curve as to

the right-hand end (shown in dotted lines), for in the case

of very long double cross-overs, very few kinds are even

theoretically possible, compared to the number of different

y= z

Ki ee es

woe,

eo j nE E a

E na

0 20 30 40 50

DISTANCE

$ Fic. 13. Curve of coincidence for chromosome I, 4. e., the ratio of ob-

served double crossing-over for points in chromosome I various distances apart

to double crossing-over expected on a chance basis.

positions in which short double cross-overs of a given

length may be found. Accordingly, the marked fall, fol-

lowed by great rise at the very end of the first curve has no

true significance.

Certain points may be seen to stand out plainly, how-

ever. It is clearly evident that interference is great for

short distances—i. e., that relative coincidence is low; as

distance increases the coincidence rises, at first, quickly,

but beyond a certain point the rise ceases. :

There is no indication of a usual length of loop of less

than half the length of the chromosome, as cytological ob-

servations on strepsinema stages would suggest, and as

would therefore be expected on the view that crossing-over

occurs at that stage. The fall seen near the right hand

end is entirely unreliable, as has been explained. | But,

even if taken at its face value, the drop at this point can

have no significance for the question at issue, for a fall

due to the loop would have to be as long as the whole pre-

430 THE AMERICAN NATURALIST [Von. L

vious rise. In addition, the curve should, on this explana-

tion, rise high above the 100 per cent. level at its modal

point, whereas it is evident that, so far as the significant

figures go, it does not rise much above 100 per cent. at any

point. It would be premature, however, to generalize fur-

ther on these results.

The curve for group II will not be presented until

greater numbers of flies have been recorded. It may be

stated, however, that this curve too shows the phenomenon

of interference, although, since the factors are not so close

- together, the crossing-over for rather small distances can-

not so well be followed.

The great variability possible in the distance between

two points of crossing-over is shown not only in the above

curves, but may be graphically illustrated from a single

case. This fly was the triple cross-over in the first chromo-

some, which has already been mentioned. Its mother was

one of the tested females of the count, whose composition

ywAbvmerf

proved £6 have been —— , and it itself was a male

with the factors yrB;. Crossing-over, therefore, must have

taken place between y and w, s and r, andr and f. The

minimum possible distance between the first two points of

crossing-over is 42, the maximum distance between the

second two is 14. The latter is the smallest distance ever

observed between two points of crossing-over. It may

here be mentioned that it will be of great interest, when

more extensive figures are obtained, to see whether in the

second chromosome the same coincidence holds between

crossings-over on opposite sides of the middle point as

between crossings-over an equal distance apart, but on the

same side. The bend of the chromosomes in the middle, or —

some other structural difference here due to the attach-

ment of the spindle fiber at this point, might cause the re-

sults to be different in the above two cases.

Incidentally, the results demonstrate another Pes Be

lying in a somewhat different field of genetics. By fol-

lowing the method of keeping stocks constantly in heter- :

No. 595] THE MECHANISM OF CROSSING-OVER 431

ozygous condition, twenty-two factors have been contin-

ually outcrossed, in each successive generation, to their

allelomorphs. Yet after about seventy-five generations of

outcrossing, these characters do not show the slightest

contamination. The experiment therefore forms an ex-

tensive test and verification of the ‘‘purity of Mendelian

segregation.’’ Castle has, however, raised the point that

in determining whether characters change, we should not

be content with casual inspection. One of the characters

in the above experiment—dachs legs—lends itself readily

to quantitative work, since one of its main features is a

shortening of the tarsus and metatarsus. Measurements

of the legs of about a dozen of these dachs flies, derived

from the stock which had been subjected to continual out-

crossing, were therefore made, as well as measurements of

the legs of some dachs flies derived from a stock which had

been kept pure; the values for normal flies were de-

termined also. At the same time the thorax length of the

flies was observed, in order that any difference in leg

length due merely to variation in the size of the ‘whole

animal might be allowed for. The results for each individ-

ual are shown in the following table. Measurements are

given in eyepiece micrometer divisions, each of which

represented .026 mm.

In order to discover whether the character had become

more variable as a result of outcrossing, the standard de-

viation of the ratios of foot to thorax, in the two stocks of

dachs, was calculated from the above data. In the out-

crossed stock the standard deviation was found to be .036,

and in the original stock .035; that is, so far as these Te-

sults can show, the variability of dachs after outcrossing

has remained just the same. However this may be, the

fact remains that the character, after being subjected to

such long-continued outcrossing, had not approached steed

` whit nearer to the type of its allelomorph. The slight dif-

ference in the other direction observed between it and the

original mutant stock is of no significance, since just about

as great differences in thorax length occurred between the

432 THE AMERICAN NATURALIST [Vou. I

two stocks, but in opposite directions in the two sexes.

The judgment based upon measurements accordingly con-

firms the judgments based upon inspection.

FEMALES

Dachs from Outcrossed Dachs from Uncrossed Wild Flies

Stock Stock

3 £33

FE Seg 2 3 ; 3

too a pA [=] © ©

gF i ae ;

mS bet

aM RR, PAE EA 19 Beige pete eo 20 pd Pere een ae 31.5

BO us Sree eels Deh AE e a Sis eee 21 ro EURO te ee A ig 31

are hie 20 LEE lees BARE oF 19 Se TEE ae ata oo 31

cl a eeprom et Ry ON 5 cee ee 20 E E, ora 31

1s este gtiprae seers 20 OU es eo 20.5 Pk, Gulag 8 ae eae 32

eee eee 20.5 a ee are 21 AS aa 33

Averages Pi Me ae hs ee 34

Sie sen 19.8 Fe OE Ae 20.25

Ratio of foot to EBA a 31.9

thorax length: .567 .596 752

MALES

Dachs, Outcrossed Dachs, Uncrossed Wild

Cig arn a oN es 17.5

Be er a os 19 pS ERTS mae eee Be 19.5 OG oso te 24

ray) CLR ee Sree 20 ea ane ere eens 20.5 OB ekg fee 26

= LS Sepa t ease 21 BU eects ae 19 yo! ieee arta 26

8 BAe Sy valley as otic 17.5 DA weeks 22 1) jeepney i 28

RE code ae gta 19 by) ca al a 283.5 ate AE ween A 29

a oes 20 GSG ee 23 Roe

Averages: :

E a Sear eee j WR sc ween 20.7 Sh eae eee 27.5

Ratio of foot t

thorax: .650 677 .887

SuMMARY

1. Recent results complete the parallelism between fac-

tor groups and chromosomes in Drosophila. his

strengthens the evidence that separation of linked factors

is due to an interchange between chromosomes. :

2. The chief gaps in the information regarding the total

frequency of interchange in the different groups have been

filled, and it is found that the usual total frequencies of

separation correspond to the lengths of the chromosomes.

This constitutes specific evidence that crossing-over is the

method of interchange between the chromosomes, and tha

No. 595] THE MECHANISM OF CROSSING-OVER 433

the frequency of crossing-over between factors is deter-

mined by their distance apart in the chromosome. It sup-

plements the other evidence for these conclusions that had

previously been found by Sturtevant in the linear manner

of linkage of the factors.

3. It seems uncertain whether crossing-over occurs in

the strepsinema stage, as concluded by Janssens, or earlier

in synapsis. The cytological evidence at present at hand

would seem insufficient to settle this point. Possible tests

for various alternative mechanisms of crossing-over are

proposed.

4. In order to study the nature of crossing-over by

means of ‘‘interference,’’ stocks were made up that dif-

fered in regard to many factors. Females heterozygous

for 22 pairs of factors were thus obtained, and a special

method was devised for testing their output. Other

special methods for obtaining multiple stocks, and for

eliminating discrepancies due to differential viability,

have also been presented.

5. The results have been arranged in the form of a

curve showing the amount of interference for various dis-

tances. The results thus far obtained confirm those ob-

tained by less exact methods, and also give evidence that

interference decreases gradually with distance from a

point of crossing-over; this, taken together with —_

evidence from non-disjunction, lends some probability to

the view that crossing-over occurs at an early stage m

synapsis. i

6. A case of crossing-over in an embryonic cell of a

male is reported.

7. Incidentally, the experiments have afforded an ex-

tensive test of Castle’s assumption of contamination of

factors by their allelomorphs. Outerossing in each gener-

ation for 75 generations has failed to change any of the.

factors.

The author is deeply indebted to Professor Morgan, and

wishes also to convey his appreciation of the active co-

operation so often rendered him by E. R. Altenburg and

434 THE AMERICAN NATURALIST [Vorl

A. H. Sturtevant, who, moreover, on several occasions

helped to tide the stocks over critical periods during which

it was not possible for the author to carry on the work.

Thanks are also due to C. B. Bridges, for supplying sev-

eral multiple stocks as well as for the use of a number

of mutants which he had already located but an account

of which he has not yet published.

BIBLIOGRAPHY

1. Bridges, C. B. 1913. Non- ee of the Sex Chromosomes of

Drosophila. aes Exper. Zool., XV.

` 2. Bridges, C. B. 1914. Direct piel through Non-disjunction, ete.

Science.

3. Hoge, M. A. 1914. Another Gene in the Fourth Chromosome of

Drosophila. AM. Nat

4. Janssens, F. A. 1909. la Theorie de La Chiasmatypie. La Cellule,

XXVII

5. Metz, ©. W. 1914. Chromosome Studies in the Diptera. I. Jour.

Exper. Zool., XVII.

6. Morgan, T. H. 1910. Sex-limited Inheritance in Drosophila. Science,

XXXII.

7. Morgan, T. H. 1910. The Method of vag aa of Two Sex-limited

i Characters in = Same Animal. Proc . Soc. Exp. Biol. Med., VIII.

Inheritance. Sisa LV,

9. Morgan, T. H. 1912. Coniplats Linkage in the Second Chromosome

of the Male. Science, XXXVI.

10. Morgan, T. H., and Lynch, C. J. 1912. The Linkage of Two Factors

in Drosophila, ete. Biol. Bull., XXIII.

11. Morgan, T. H., Sturtevant, A. H., Muller, H. J., and Bridges, C. B.

The Mechanix of Mondelisn Heredity. 1915. Henry Holt & Co.

12 Muller, H. J. 1914. A Gene for the Fourth Chromosome of Droso-

phila. Jour. Senet Zool., XVII.

‘our. Exper. Zool., V.

13. Stevens, N. M. 1909. A Study of the Germ Cells of Certain Diptera.

14, Sturtevant, A. H. 1913. The Linear Arrangement of Six Sex-linked

Factors, as Shown by Their Mode of Association. Jour. Exper.

Zool., XIV.

15. buivtecnt, A. H. 1913. A Third Group of Linked Factors in Droso-

phila. Am. Nar

16. Sturtevant, A. H. 1915. The peters of = gine as Studied

tiroogh Linkage. Zeit. f. i bst

17. Sturtevant, A. H. 1914. The palea Hypothesis as Applied to

rosophila. AM. Nat., XLVIII.

18. Trow, A. H. 1913. Forms of Reduplication. Primary and secondary.

Jour. Gen., II.

19. Wenrich, 'D. H. 1915. eyed and the Individuality of the Chromo-

mes. Science, N. $., t

SHORTER ARTICLES AND DISCUSSION

DISTRIBUTION OF THE CACTI WITH ESPECIAL REF-

ERENCE TO THE ROLE PLAYED BY THE ROOT

RESPONSE TO SOIL TEMPERATURE

-AND SOIL MOISTURE

As is very well known, it is the common habit, when referring

to the relation of a ‘‘plant’’ to its environment, to mean the sub-

aerial portion only, leaving quite to one side the subterranean

paris. That there is little logic in this will be readily acknowl-

edged, although the possible causes are not far to seek. In the

first place, for patent reasons, roots do not greatly excite our ad-

miration or curiosity, and thus have received little attention in

the field. Further, relatively little experimental work has been

done on the roots of plants other than on seedlings and growing

in solutions. And besides these conditions which refer immedi-

ately to the plant, there is a nearly related one which has to do

with its environment, especially with the root environment. The

soils and the soil condition of whatever sort are probably more

difficult to study, and the results more difficult to express in a

manner capable of ready application than the subaerial environ-

ment of the plant. However, it has not been its difficulty alone

that has been the deterrent in the study of the environment of

roots since certain features, for instance the soil temperature, can

be easily learned by appropriate apparatus. Could we have a

comprehensive series of data touching this feature alone, to men-

tion no other, we should be in possession of a very useful engine

for use in comparative studies on causes underlying the distribu-

tion of plants, and, further, through it the study of the root-

systems of plants, and of their biological value, would be greatly

stimulated.

While it is here recognized that the presence ( of a planit i in oe

environment is an expression of the th 7

the entire environment, it is herann for the purpose in hand,

to ignore the responses of the shoots, and to focus our attention

for the time on the root relation alone. It can be noted, however,

as is very well known, that the activities of the latter may w

reflected in those of the former. Such a ee eee inter- |

436 THE AMERICAN NATURALIST [ Vou. L

esting possibilities, was observed at the Coastal Laboratory, at

Carmel, California, and may be briefiy referred to in this place.

Among the species growing in the experimental plots at the labo-

ratory are Opuntia versicolor and Fouquieria splendens from the

vicinity of the Desert Laboratory, Tucson, Arizona. Owing to

the usual low temperature of the air, and soil, these species gen-

erally make little or no shoot growth at Carmel. When, however,

the roots of the plants are kept in soil whose temperature is

25-30° C., the shoots remaining in the cool air, not only do the

roots grow rapidly, but new shoots and fresh leaves are promptly

formed. Without pursuing this phase of the matter further it can

be seen that analogous results might occur in nature should the

soil conditions, for instance its color or the relation of the soil

surface to the incident heat rays, be such as to bring about a

relatively warm soil environment. Under such conditions it is

clear that,only a study of the soil temperatures, and the responses

of the roots to soil temperatures, would provide the key to the

solution of the shoot behavior and to all of its accompanying

results.

It is generally recognized that the soil acts as a reservoir for

heat, and that the daily course of soil temperature is unlike that

of the air immediately above it. Thus, the roots are subjected to

temperature conditions which are quite different from those

affecting the shoot of the same organisms. The shoot is warmer

by day and colder by night than the root and it is improbable

whether the roots of most woody plants are often subject to

‘“‘optimum’’ temperature conditions, as must frequently be the

case of the shoots. An exception to this statement, however, is

to be found in the cacti where the most favorable soil tempera-

tures are of great importance among those environmental features

that may be called definitive. The roots of most cacti of the

Tucson region, and possibly elsewhere, lie near the surface of the

ground. For the most part they are less than 30 em. deep.

Inasmuch as the rate of root growth of the cacti, as will be shown —

er the

optimum,’’ the importance to these plants of a shallow position

of the roots will be apparent. It is only in the upper soil horizon —

that such favoring temperatures are to be found. It is of inter- —

1 Cannon, W. A., ‘‘On the Relation of Root Growth and Development to

the Temperature kad Aeration of the Soil,’’ American Journal of Botany,

Vol. 2, p. 211, 1915.

No.595] SHORTER ARTICLES AND DISCUSSION 437

est to note, on the other hand, that deeply placed root-systems,

such as of Prosopis velutina, may have a relatively rapid growth

rate at relatively low temperatures.? In such a case it is quite

possible that the rôle played by root response to temperature in

species distribution may be less important, or, at any rate, differ-

ent from that played by the roots of the cacti, for example, to the

distribution of members of that family.

We will now glance at the most striking conditions of soil tem-

perature as they obtain at the Desert Laboratory, where much of

the work here referred to has been carried on, before taking up a

résumé of the response of the roots of the cacti to the temperature

of the soil and the relation this suggests to the general distribu-

tion of the family,

Three series of soil thermographic records, which are now being

supplemented by others, have been kept at the Desert Laboratory.

These relate to three depths, namely, 15 cm., 30 cm., and about

2.6m. Although the records cover a series of years, it will serve

the purpose in hand if we refer to those for the year 1910 only.

The mean maxima and the mean minima temperatures for the

three depths will provide saints data for interesting com-

parisons,

At the shallowest depth, 15 em., the mean maxima temperatures

for midwinter and midsummer were 8.1° and 34° C., respectively.

The mean minima, for the same seasons, were 3.9° and 30.8° C.

At a depth of 30 cm. the maximal range was from 12.2° C., in

January, to 33° C., in July, and the minima temperatures, for the

same months, 10° and 32.2° C., respectively. It was observed that

from June to September, inclusive, the curve of the mean maxima

for this depth did not fall below 32.2° C

At a depth of 2.6 m., the mean maxima temperatures ranged

from 18.6° C., in January, to 27° C., in July.

Upon comparing, in a general way, the mean maxima for the

different soil depths we see that the shallowest soil is the warmest

from April to August, inclusive; that in September and October

only the highest temperatures are found at a depth of 30 em.;

and that in late winter-early spring the lowest level is also the

warmest. :

The relation of the rate of root growth in Opuntia versicolor,

as representative of the cacti, to different soil temperatures indi-

cates interesting conditions and — and will > given in

the following paragraph :

2 Cannon, W. A., l c.

438 THE AMERICAN NATURALIST [Vou. L

Very many experimental cultures, of various kinds, made both

at the Desert Laboratory and the Coastal Laboratory, have shown

that the growth rate of the roots of Opuntia, within limits, varies

directly with the temperature. It is relatively slow at 20° C.,

and most rapid at 34° C. The hourly increase in length of the

roots at 20° C. is about 0.3 mm., and at 30° C. it is approximately

twice this. Above 34° C., the rate falls off rapidly and ceases

at about 42.5° C. Below 20° C., the growth rate is very slow, as,

for example, at a temperature of about 16° C. an increase in

length of a perfectly normal root was found to be only 1 mm. in

14 hours. The maximum rate, taking place at about 34° C., is

about 1 mm. an hour.

Referring back now to the soil temperatures, it will be seen

that the roots of this species are exposed to optimum conditions in

July and August only, although the soil temperatures for one

month before and one month following this period, at a depth of

30 cm., or less, is also high enough for an effective growth rate.

The soil temperatures, at this depth, in the other months, and at

the lowest level throughout the year, are not sufficiently high for

the best root activity. However this may be, we find, in short,

that suitable soil temperatures obtain at the depths occupied by

the roots of the cacti during four months of the year. But it does

not follow that root growth goes on throughout this period for

the reason that the foresummer is arid and the shallow soils are

impossibly dry, having less than 10 per cent. of moisture. Active

root growth of the cacti, in fact, commences with the coming of

the summer rainy season, about the middle of July. It is ended

by the cooling of the soil in early autumn. The length of the

active growing season of the roots of the cacti, therefore, does not

usually exceed six or eight weeks.

It is in the response of the roots to the temperature and moisture

conditions, as just sketched, that lies the erux of the suggestion

offered in this paper, namely, that conditions being otherwise

favorable, the cacti, which are shallowly rooted, occur in such

regions as have the superficial soils moist.at the same time they

are suitably warm, and they are wanting where such soil condi-

tions fail.

With the reaction of the roots of the cacti to temperature in

mind, it will be instructive to examine briefly the leading climatic

features, so far as they affect the case in point, of the regions in

which the cacti form a conspicuous portion of the vegetation.

According to Engler and Prantl, the cacti occur mainly in the

No.595] SHORTER ARTICLES AND DISCUSSION 439

dry parts of Mexico, in the portions of the United States which

border ôn Mexico, in eastern and central Brazil, and in portions

of the Andes countries. Taking two or three genera as examples,

we learn, for instance, that Cereus occurs in Mexico, and in the

Andes of Argentina and Brazil. Echinocactus extends from

the southwestern part of our country to Brazil and Chili Opuntia

is found in Mexico, Peru, Chili, in Central America and in the

southwestern portions, especially, of the United States. Although

certain species are outside of this range, as especially certain

opuntias, where the winters are exceedingly cold, all are subject

in summer, when active growth takes place, to conditions which

are in rather close accord. A glance at the summer climates of

these regions will, I think, establish this point.

In the central part of Mexico, at Tehuacan, the annual rainfall

is about 15 inches, most of which occurs in summer, and at

Pueblo, 70 miles distant, and at a higher altitude, where the

annual precipitation is more than twice that at Tehuacan, 72 per

cent. of the rain comes in the warm season. The Tehuacan region

_ has been characterized as being the richest of any known in cacti.

At Chihuahua, where the rainfall is 10.86 inches, the amount

falling in the summer season is also over 70 per cent.

In the southwestern part of the United States, where the cacti

constitute a conspicuous portion of the flora, a relatively large

summer rainfall is also reported. At Tucson, for example, the

precipitation amounts to 11.74 inches annually, of which 54.7 per

cent. is received in July, August, and the first part of September.

Turning now to South America, and without especial regard as

to the presence of cacti at the particular stations quoted, eh find

that over a relatively large area, a large percentage of rainfall is

in the warm part of the year. For example, at Matto Grosso,

Brazil, the greatest rainfall is in December. From June to Au-

gust and generally for a month before and after this period, the |

climate is usually dry.* J

Along the east coast rain occurs from February to April, June

to September being dry. In the Cordilleras of Bolivia and Peru,

the rainy period is in December-March, and the climate is dry

from April to October. At La Paz, although rain may fall tig

month of the year, December to February is regarded as being

the season of rain.

2 MacDougal, D. T., ‘‘ Botanical Features of the North Amarican osetia,”

Carnegie Inst. Wash. Pub. 99, 1908.

+ Hann, ‘‘Handbuch der Klimatologie,’’ Bd. IT, 1910.

440 THE AMERICAN NATURALIST ` [ Vou. L

We have supplemental evidence that the cacti grow most suc-

cessfully in such warm temperate moderately arid regions as have

precipitation in the warm season from the work of the Australian

commission for the study of certain species which have escaped

from cultivation in several countries, especially Australia, and

have become a pest. In Queensland and New South Wales

species of Opuntia constitute a serious weed. At Westward and

Rockhampton, Queensland, where the cacti are particularly a

nuisance, over 50 per cent. of the annual rainfall occurs in

December—March, inclusive. Soil temperature data from Bris-

bane, depth one foot, show that the mean temperature from

October to April is between 22.7° and 27.9° C., and that during

the colder portion of the year the mean temperature at that depth

is below 20° C.°®

The commission studied the cactus problem in several different

portions of the world, among which were Cape Colony, central

and southern India, southeastern and southern South America

and the Mediterranean region. It will be instructive to sketch the

leading climatic features of definite localities where cacti were

found to have escaped cultivation.»

In southern Africa, species of Opuntia occur in a naturalized

condition in the Great Karoo and in the Transvaal. In parts of

the former region, as at Graaf Reinet, the species are abundant.

At Graaf Reinet, according to Knox,’ where the total precipita-

tion is 15.29 inches, 63 per cent. occurs in November—March. In

the Transvaal, where the escaped cacti are less numerous, the

rainfall is 26.94 inches, of which 81 per cent. occurs in November-

March.

In northern Africa the cacti escape from the oases very little,

and the same is to a degree true of other portions of the Mediter-

ranean region. In Algeria and Tunis, according to Knox, the

rains are almost exclusively restricted to the winter season.

In India species are naturalized over a large territory, as, for

example, in the Madras Province and in the Panjab. In Madras

the prickly-pear has become a formidable evil throughout several

districts. At Madras* 79 per cent. of the total precipitation takes

place in August-September. In the state of Mysore, also,

5 Report of the Prickly-pear Traveling Commission, Brisbane, 1914.

€ ‘‘ Results of Rainfall Observations made in Queensland,” H. A. Hand,

1914,

7‘‘ The Climate of the Continent of Africa,’’ 1911.

8 Hann, ‘‘Handbuch der Klimatologie,’’ 1. c.

No.595] SHORTER ARTICLES AND DISCUSSION 441

opuntia is commoon. At Mysore, according to Hann, 81 per cent.

of the rainfall is from May to October. At Lahore the prickly-

pear is not so abundant as further south, but it oceurs escaped,

nevertheless. Here the July-August rains comprise 55 per cent.

of the total annual precipitation.

In South America the Commission examined naturalized opun-

tias in portions of Brazil and Argentina chiefly. An important

prickly-pear region is northwestern Argentina, where native as

well as introduced species of cacti occur in abundance. At Salta

there is as good as no rain in the cold season, between May and

September. At Tucuman, 69 per cent. of the rainfall takes place

between December and March, inclusive (Hann), and at Cata-

marca, between November and March, inclusive, 81 per cent. of

the total annual precipitation occurs.

Without pursuing this phase of the matter further, it would

appear, in short, that in regions where cacti are abundant, either

native or introduced, rains occur during the warm season. It is

not intended to discuss in this place the actual amount of rain-

fall which falling in the warm season makes the presence of a

cactus flora possible. It is well known, however, that the amount

of precipitation in regions where cacti occur is extremely unlike,

and that it may vary from season to season in any one region.

This last, in fact, is one of the leading characteristics of an arid,

or semi-arid region. So far as regards the precipitation differ- `

ences in separate regions frequented by cacti, it is interesting w

note that at Rockhampton, Queensland, it is 40.09 inches,” while

at Phoenix, Arizona, it is 7.06 inches,” and that in the former

region 20 inches occurs in the warm season, while the amount or

summer precipitation at Phoenix is between 0.9 and 2.1 inches,

as means of the extremes."

In the Mohave the annual rainfall is 4.97 inches, about two

inches less than the mean precipitation for Phoenix. In the Mo-

have, however, 86 per cent. of the rainfall is in winter, which

greatly emphasizes the differences in summer aridity of these

regions, and points to a probable reason why cacti are almost

°“‘Results of Rainfall Observations made in Queensland,” H. A. Hunt,

l. c

10“ Botanical Features of North American Deserts,” D. T. MacDougal,

P. 95, 1908.

11 “‘ Climatology of the United States,” A. J. Henry, U. S. Dep. Ag. PN-

Q., 1906. i

12 MacDougal, l. c.

442 THE AMERICAN NATURALIST [Von. L

wholly wanting in the flora of the latter region. From these

climatic facts it appears that while soil moisture is a condition

sine qua non of the presenge of the cacti, the range of the actual

amount; of soil moisture must be very great indeed, so, in short,

it results that the temperature is the factor in direct control, thus

a very important limiting factor.

Should we sum up, therefore, the factors thus far mentioned

as being important among those which determine the distribu-

tion of the cacti, we find, in the first place, that the shallowly

placed root-system subjects the roots to the greatest possible ex-

tremes in soil temperatures, including those that are high, and, at

the same time, makes it possible for the plants to advantage from

the minimum effective rainfall. Further, an effective growth rate

of the roots takes place only at relatively high soil temperatures.

And, finally, a certain but highly variable amount of moisture

must be present in the soil. Since the cruz of the matter, how-

ever, appears to be the fact that the root-system of the cacti are

essentially superficial, there is the additional factor, or factors,

which bring about this circumstance. These are at present un-

proved, but the results of experimental studies, not published,

indicate that among them must be included the response to the

oxygen supply of the soil.

W. A. CANNON

DESERT LABORATORY.

THE INHERITANCE OF CONGENITAL CATARACT

In the February number of the AMERICAN NaTuRALIST there

is an article from the Bussey Institution by Jones and Masont

in which an attempt is made to show that congenital cataract be-

haves in heredity as a simple Mendelian recessive. The authors

from a study of family histories published by Harman in the

‘*Treasury of Human Inheritance’’ come to conclusions at vari- —

ance with those of Bateson and Davenport, which authors they

are perhaps unjustly disposed to criticize. The paper is well

written and embodies a considerable mass of data, so that the

reader not familiar with this particular problem might easily be

led to think that the older investigators had really made a mis-

take in interpretation. The evidence, however, does not seem to

1 Jones, D. F., and Mason, S. L., ‘Inheritance of Congenital Cataract,”

THE AMERICAN Nareks, Vol. D No. 590, pp. 119-126, February, 1916.

No.595] ` SHORTER ARTICLES AND DISCUSSION 443

warrant such a conclusion, as the present paper will attempt to

demonstrate.

It is stated on page 120 of the article in question that the data

used in the paper are taken from the tables accompanying Har-

man’s publication. Since we are concerned wholly with a ques-

tion of interpretation we may confine ourselves to these tables.’

The families recorded in the tables are classified by Jones and

Mason as follows (p. 120):

After discarding all the doubtful cases, and picking a sibship with its

parents from the table as a family, there is left a total of one hundred

and twenty-five families which are classified into three different cate-

gories, as follows: (A) Both parents normal with at least one abnormal

child; (B) one parent normal, the other affected with some form of

hast cataract, with at least one abnormal child; (C) both parents

abnormal, giving only abnormal children.

In each of these groups (A, B, and C) it is thought that evidence

is found in support of their contention that congenital cataract

is a recessive character. We may now consider this evidence in

the order in which it is presented.

In group A, 31 families are cited in which both sarii are

normal with one or more affected children. This is the strong-

est, or really the only, evidence that is offered in favor of the

recessive character view. Let us examine it more closely. On

going over Harman’s tables, we find that of these 31 families

there are 16 in which the affected individuals produced no off-

spring or nothing but normals. We do not wish to lay great

emphasis on this point, but in such cases one should bear in mind

the common clinical belief and the experimental proof (for rab-

bits and guinea-pigs)? that a certain number of congenital cat-

aracts are produced by intrauterine poisoning eous necessarily

any reference to heredity.

Another possible explanation for some of the examples in group

A is that they dT cases or e de novo. Jones and

Mason say:

2 Harman, N. Bishop, ‘‘ Congenital. Cataract,?’ $ i the PA D of Hu-

AUR Inheritance, ” Part IV, Section XIII a. Eugenics Laboratory Me-

moirs, XI, pp. 126-169. Pl. XXVII-XXXII. panus 1910.

mentellen Dreg ;

apa e pran Woch; 58 Jahrg., No. 32, pp- 1716-1717.

Aug. 8, 1911. (Reviewed in most of the eye j Re e

444 THE AMERICAN NATURALIST [Von. L

Surely it is not possible to explain so many cases as origin de novo

or as due to faulty classification of the parents.

With reference to the origin de novo of characters it may be re-

called that one does not‘have to search the literature long to find

instances of the same mutation occurring repeatedly in different

stocks and at different times, or of certain stocks that seem to be

especially prone to mutation.* Congenital cataracts occur in

many races of man and in other mammals. So far as the writer

is aware we are not at present in a position to state, either on the

basis of observed data or from à priori consideration just how

frequently mutations may occur in the human germplasm.

Again, since Jones and Mason elsewhere in the same paper

(p. 124) use the argument that ‘‘heterozygous individuals some-

times show the recessive character,’? we might, if necessary, use

the same argument to prove the dominance of cataract. On the

assumption that congenital cataract is dominant instead of re-

cessive it might be maintained that in those cases where both

parents of affected individuals seem to'be normal, one of them is,

after all, heterozygous, and affected children are therefore to be

expected.

Finally it should be recalled that i in their statistical study of

these 31 families Jones and Mason do not get the results that

their hypothesis demands. After having made the proper mathe-

matical corrections there still remains a discrepancy which they

do not adequately explain, the agreement between theoretical and

observed results being only .418 (p. 122). In order to test what

one should expect from the examination of such data when the

character is recessive, I have taken a paper by Usher* on retinitis

pigmentosa and summarized the charts in the same way that

Jones and Mason summarize those of Harman. Now retinitis

pigmentosa probably is a recessive character as is commonly be-

lieved. In the charts of Usher are recorded 44 families in which

4To cite a single case, we may mention the results of Barfurth in breed:

novo, but once having appeared is transmitted as a Mendelian dominant.

Barfurth, Dietrich, ‘‘ Experimentelle Untersuchung über die Vererbung der

Hyperdactylie bei Hühnern. V. Mittelung: Weitere Ergebnisse und Ver-

such ihrer Deutung nach den Mendelschen ae Arch. f. Entwichlungs- fe

mech. d. Organism., Bd. 40, pp. 279-309,

5 Usher, C.. H., ‘‘On the Ta of pene’ tinitis Pigmentosa with Notes

= feels The © Royal London Ophthalmic Hospital Reports, Vol. 19, PP»

236 Re

No. 595] SHORTER ARTICLES AND DISCUSSION 445

neither parent of the affected individual shows the defect (3.8,

both are presumably heterozygous). These 44 pairs of parents

are recorded as having 320 children, of whom 77 are affected—

24 + per cent. as compared with an expectation of 25 per cent.

If the data on cataract were to yield results as close as this we

would be more disposed to credit the view that the character is

recessive. 2

So far as the 31 families of category A are concerned it must

be admitted that absolute proof of the fallacy of the recessive

character view can not be furnished, but it will be apparent that

there is considerable evidence which not only fails to support

this view, but actually points decidedly against it. This fact,

taken in connection with the positive refutation which the data

in categories B and C supply, makes a very strong case against

the view that congenital cataract is a recessive character.

In the second category (B) where one parent is affected the

other normal, Jones and Mason remark that ‘‘the number of

affected children would be expected to be approximately the same

whether the character was inherited as a dominant or a recessive”?

(p. 121). But it must be borne in mind that the offspring of a

recessive show the 1:1 ratio only when the mate is heterozygous,

and in their second table Jones and Mason assume that the par-

ents of the children in group B represent the cross ‘‘Nn X nn.”

The question is not raised as to the probability of the oecurrence

of such matings nor does there seem to have been an attempt

made to trace the offspring from the normal and affected mem-

bers of the F, generation. In other words, the data of really

critical significance do not seem to have been considered. As it

stands, then, Table II seems to present no evidence either for or

against the above hypothesis, a point which the anen them-

selves recognize as the quotation indicates.

Since the authors have not tabulated the data which would

seem to be of most significance, we may return,to Harman’s orig-

inal charts assuming for purposes of the discussion that con-

genital cataract really is a recessive. On this assumption there

are two important conditions which we sbona expect to find ful-

filled.

1. If congenital cataract were a recessive, a cataractous person

married to a normal should in most cases produce only normal

children. This will be apparent when it is reoalled: m goi

446 THE AMERICAN NATURALIST [Von. L

genital cataract is so rare (perhaps 1 in 4,000 or 5,000) ® that the

number of heterozygous individuals in the general population

must be relatively low—theoretically not more than 1:30.7 In

other words, if congenital cataract were recessive the chances that

an affected individual in marrying would get a heterozygous

partner and thereby be able to produce affected children would

be only one in thirty and the chances that the same thing would

happen in several generations in direct descent as occurs repeat-

edly in the charts (the case in over 40 different family trees)

become extremely remote. We should not then expect families

with one ecataractous parent to contain affected children more

often than in the above proportion.®

2. If congenital cataract were recessive the normal children of

a cataractous parent should themselves produce affected children

in half as many cases as do their cataractous sibs and the total

number of affected children produced should be one half as great

in the first case as in the second. This expectation follows from

the assumption that the original (grandparental) mating was

nX nn. As a result of such a mating the F, generation can

be composed only of Nn and nn individuals. Neither of these

should produce affected children except when married to an Nn

(or an nn), and the chances of such a marriage are as great (or

as remote) in the one case as in the other. In other words, an

equal number of heterozygous and pure recessive individuals of

the F, generation should get heterozygous mates. In these few

families the expectation for F, would then be of a 1:3 ratio for

the Nn X Nn matings and a 1:1 ratio for the Nn X nn, which

would obviously give one half as many affected children in the

first case as in the second.

Harman’s charts afford sufficient data to settle these points:

conclusively. In regard to the first point there are 96 cases in

which the cataractous child of a cataractous parent has himself

` 6 This statement is an estimate based on data gathered from a number of

Sources, e. g., Jour. Amer. Med. Ass’n, quotations from the census reports,

ete. It is probably not too low, but if the incidence were as much as 1: 100

Breeding,’’ in Genetics, Vol. 1, No. 1, pp. 53-89. Jennings is not respon-

sible for the use of his formulae in this connection, but they obviously apply.

* Corroborative evidence on this point is furnished by the histories of

retinitis pigmentosa in the paper by Usher already referred to.

No.595] ` SHORTER ARTICLES AND DISCUSSION 447

reached maturity and produced one or more normal offspring,

thus proving on the assumption that cataract is recessive, that

his consort was either NN or Nn. Instead of finding, as we

should expect on the assumption that cataract is a recessive, that

only 3 or 4 (1:30) of these individuals would find an Nn mate

and therefore be capable of giving some affected children, what

` we really do find is that of these 96 families, 83 have produced

cataractous children—86 per cent. It is significant, incidentally,

that of the remaining 13, ten are families of three children or

less. A more striking refutation of the assumption could hardly ,

be found.

In regard to the second point we find that there are 47 normal

individuals in the same F, generation from the supposedly Nn

X nn matings who have also come to maturity and produced chil-

dren. In 42 of these families only normal children have resulted.

In 10 per cent. of such families, however, one or more cataractous

children have been produced. But the relation between 10 per

cent. on the one hand and 86 per cent. on the other is very far

from being the relation of one to two.°

These considerations based on the results from such matings

as are included in category B, and the F, descendants of such

matings, furnish convincing evidence that congenital cataract is

not a Mendelian recessive.

Finally as to the last point, which concerns category C, we can

agree with the authors when they say:

The critical test as to whether or not congenital cataract can be con-

sidered as a simple recessive character lies in the matings of abnormal

by abnormal. Families of this kind should have only abnormal chil-

dren. Only three such matings are available.

I find two of these, but have searched the charts in vain for the

third which must be an error or inadvertently drawn from out-

side sources. This really is immaterial since it would require

many such cases to prove the hypothesis, where a single bona fide

case in which two affected individuals produce normal offspring |

is sufficient to overthrow it. One of the three cases cited is such. .

Not to mention the possibility of faulty classification, possible mutation

ete., they might, for instance, belong to the class mentioned by Jones and

Mason in which the parent while really heterozygous appears in the recessive

“as THE AMERICAN NATURALIST [Von. L

In chart 342, IJI-28 and III-37 are shown as a pair of parents

both of whom are affected in both eyes. The descriptions quoted

from Nettleship by Harman (op. cit., p. 148) show that the diag-

nosis is evidently based on ophthalmoscopic examinations. This

is a clear case of abnormal by abnormal, and if we were to regard

it as ‘‘doubtful’’ we could find equal justification for so regard-

ing any other chart in the whole series. The offspring of this

marriage are seven children, of whom two have cataract, three

thought to have been free from it died in infancy, and two are

a definitely known to be normal. This is the one critical case that

is needed and, taken at its face value, it completely refutes the

argument for the recessive nature of congenital cataract.

In conelusion, the writer does not wish to insist on arguments

from a few particular cases, nor does he wish to make purely

academic distinctions in the treatment of data. In particular,

he does not wish to be understood as maintaining that congenital

cataract behaves strictly as a single dominant unit character—

a view to which he does not subscribe. The point upon which he

does insist, however, is that the view, presented in the paper

under discussion, namely that congenital cataract is due to a

single recessive character, not only fails to find support in the

data which was presented, but is in reality actually apv

by that data.

C. H. DANFORTH

DEPARTMENT OF ANATOMY,

WASHINGTON Untversiry MEDICAL SCHOOL

VOL. L, NO. 596° AUGUST, 1916

THE

AMERICAN

NATURALIAY 2

AMORTHLY SORELLE o o

Devoted to the Advancement of i eee ee i 7 o

Special Reference to the Factors of Evolution Gre E a

The American Naturalist

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

sent to the Editor of THE AM MERICAN NATURALIST, Garrison-on-Hudson , New York.

Short articles containing summaries of research work bear aring on the

seamen efi of oraraa evolution are especially welcome, and wili be given preference

in pu

One e ndred reprints of spre rage are supplied to authors free of charge.

Further “reprints will be supplied atc

iptions and ienei should be sent to the lagri The

subsoription price is four dollars a year. Foreign postage is fifty cents and

anadian postage twenty-five cents additional. The charge for single copies is

Pty cents. The advertising rates are Four Dollars for a page.

THE SCIENCE PRESS

Lancaster, Pa. Garrison, N. Y.

_ NEW YORK: Sub-Station 84

Ent ter, April 2, 1908, at the Post Offi ster, Pa., under the Act of

Congress of March 3, 1879.

NATURAL HISTORY OF MAN The Diato RERS of en pe Vicinity

Ad 1 Bie i th rii diat a ‘fi Í ‘Philadelphia

e o z

18,000 Specimens illustrating the daily life of sad the scription of th “a radius of one hundred miles,

Savage Races, ay; ecole now extinct, s, Orna- | cluding the pre- and postglacial deposits E Drar PA DA

ments, Deities, i Tools, etc. ices 1e clay of the ancient Delaw ver has just been issued

Guaran Soa y ae Da oe ind Te Coll k form by the J. B. amei, it Company. 143 pages.

age. ype ions for monet ha 700 original illustrations, enlarged 800 times, with

Museums a Specialty. cens eae a eda 12. Bound in Art Vellum cloth.

jent postage pai

LDMAN Brixton Hill, London Address Charles S. Boyer.

9 AK 7 : England | 1649 Columbia Avenue Tniladelphia, Pa

PUBLISHED BY THE CAMBRIDGE UNIVERSITY PRESS

our Professor of Genetics i in the University of ee

‘EXTRACT FROM THE PREFACE ©

present litle woke Mia hoks 9 Witte. -o

Ih Lo a rea he ive in visit aop pica

: butterfly them.

(eee ays ae ee E

a MIMICRY IN BUTTERFLIES |

-Bound in Buckram. Royals a BR i Eat ere of which |

= By Recixauy aes Punyert, F a Fellow of Gonville and Caius cae fg

“Thistle book E been written in the w that it may. soe to o severi l ce

THE

AMERICAN NATURALIST

VoL. L. August, 1916 No. 596

THE FORM OF EVOLUTIONARY THEORY THAT

MODERN GENETICAL RESEARCH SEEMS

TO FAVOR

DR. CHAS. B. DAVENPORT

CARNEGIE STATION FOR EXPERIMENTAL EVOLUTION, COLD SPRING

HARBOR

Nature produces those things which being continuously moved by a

certain principle contained in themselves arrive at a certain end.

—Aristotle.

CONTENTS

I. General Statement of Theory of Evolution ......--+++++++++++s

II. Support for the Theory from Collateral Piplden oes wie ee

Embryology . o eon ce eee ee aie = St

9. Paloontolowy: axis fives Wen w len oe renee tree VERISED

3. Experimental Breeding ....--++-+-+e++ssctrrrertresttees

4: Radiation Studios r.-s 50 eee ee cee ee ke

III. Certain Consequences of the Theory ....--++++++++erersrertett

. ton o seour oh ee A A ee EAS A

2. Rôle of Selection .. Ler ce parni rat e Ea e een

IV. Sommaty <<. seeen i oen pe a A a errerer

Y. Literature tod 1r: -brh resserrer Tirra AANA we cents tr?

I. GENERAL STATEMENT oF THEORY of EVOLUTION

Tue history of evolution, all will agree, has been from

the less specialized to the more specialized. The pre-

vailing view is that this greater specialization has been

achieved by adding qualities one by one to the less spe-

cialized until there has become built up through the ages

so complex structures as the higher types of organisms.

Organic evolution, from this view, has proceeded along

the same lines as the evolution of an old English manor

449

450 THE AMERICAN NATURALIST (Vor L

house through the accretions of successive generations;

or is like some medieval cathedral to which each genera-

tion has added some sculptures, some stones to the

steeple, some new stained-glass windows. Organisms,

the theory runs, began simple,—they varied, why or how

no matter; variation is given and is in all directions.

Positive variations toward greater complexity were

usually advantageous and individuals showing such

elbowed out of existence their less-favored cousins and

established a new and higher level from which evolution

might proceed. :

This view has certain resemblances to an old view of

ontogeny. The embryo is simple. Food is added and

food makes this part and that grow. Why only certain

_ parts grow in the presence of food and not all—the dif-

ferential nature of growth—that is given. Tt is the

nature of the organism that all parts should not grow

equally; but the essential thing is that food and water

and heat are the things that add parts and make the

embryo larger and more complex.

This view of development has, I think it will be ad-

mitted, now become generally abandoned. To-day we

recognize rather that the egg or embryo is not so simple,

but that, on the contrary, it has wrapped up in it all of

the potentialities that are eventually realized; potentiali-

ties that will be realized, however, only if conditions of life

(food, water, heat, etc.) are appropriate.

The alternative view of evolution, like the modern view

of embryonic development, lays more stress on the in-

ternal factors of evolution. It postulates that the primi-

tive organisms, like the eggs, are not so simple as they

look but have a molecular constitution of great com-

plexity; and, just as the egg has a mechanism by virtue —

of which (under favorable conditions) it develops, 80 the :

ancestral protoplasm had a mechanism by virtue of whic

(under favorable conditions) it evolved. We do not hee

know much about the specific molecular machinery that

determines the specific nature of differential growth, but — 7

- No. 596] EVOLUTIONARY THEORY 451

we have reason to think that it is located chiefly in the

nucleus. Similarly we are ignorant of the specific nature

of the machinery that determines phylogenetic varia-

tions, but we have reason to think that it is located in the

germ plasm and that the karyokinetic phenomena, espe-

cially the movements of chromosomes at and around the

time of fertilization, have a great deal to do with such

phylogenetic change. But as the egg with its given in-

ternal mechanism of development under adequate ex-

ternal conditions will develop into the specific adult form,

so the primitive germ plasm with its internal mechanism

of evolution under adequate external conditions has devel-

oped into all those forms or kinds of germ plasm that

are responsible for the great variety of present and past

organisms. Just as the egg-nucleus contains probably

fewer kinds of molecules, but each more complex, than a

nerve cell, e. g., of the adult, so the ancestral form of

protoplasm probably contained fewer kind of molecules

each more complex than the derived forms. The derived

forms, conversely, have more kinds each of simpler

constitution.

In this view (if we may let our imagination picture

the consequences) the foundation of the organic world

was laid when a tremendously complex, vital molecule

capable of splitting up into a vast number of kinds of

other vital molecules was evolved! The capacity for

thus splitting off molecules determines the possibility of

production of the organic species with their vast number

of characteristics. oe

It is to be kept in mind, however, that the number of

genes is probably less than the number of elementary

species. For numerous elementary species differ by

only one or two genes and in many cases the species

differ only in new combinations of the same set of genes.

This theory of evolution is not new; it is that briefly

expressed by Bateson in his Australian address; it is

very clearly expressed by Hagedoorn in Roux’s Vor-

träge” and has received the support of Lotsy (1913). It

452 THE AMERICAN NATURALIST [Von L

is, however, essentially Nageli’s theory of evolution from

within by virtue of a perfecting or progressive tendency,

in support of which Nägeli himself draws the parallel

between embryology and evolution. It has certain points

of resemblance to Eimer’s orthogenesis, but differs from

it tremendously in that Eimer thought evolution was

directed by the external world and that there were summed

in the germ plasm the impressions of that world received

in successive generations. Huxley was inclined to accept

the theory of internal factors in evolution. He says:

I apprehend that the foundation of the theory of natural selection is

the fact that living bodies tend incessantly to vary. This variation is

neither indefinite, nor fortuitous, nor does it take place in all directions,

in the strict sense of these words. ... A whale does not tend to vary

in the direction of producing feathers, nor a bird in the direction of

developing whalebone.

Mivart and many others have long contended that evolu-

tion is due to internal factors. Their views are in accord |

with Aristotle’s. |

Finally, I may close this section by quoting a simile

from Bergson, as translated by Mitchell.

The evolution movement would be a simple one, and we should soon

have been able to determine its direction, if life had described a single

course, like that of a solid ball shot from a cannon. But it proceeds

rather like a shell, which suddenly bursts into fragments, which frag-

ments, being israel shells, burst in their turn into fragments des-

tined to burst again, and so on for a time incommensurably long. We

perceive only what is nearest to us, namely, the scattered movements of

the pulverized explosions. From them we have to go back, stage by

stage, to the original movement. 3

When a shell bursts the particular way it breaks is explained both by

the explosive force of the powder it contains and by the presistance of

the metal. So of the way life breaks into individuals and species. It

depends, we think, on two series of causes: the resistance life meets

from inert matter, and the explosive foree—due to an unstable balance

of tendencies—which life bears within itself.

Il. Support FoR THE Tee FROM COLLATERAL FIELDS —

The view that the course of evolution (like the develop-

ment of the individual) is chiefly determined by internal

No. 596] EVOLUTIONARY THEORY 453

changes receives support from various collateral fields

of investigation.

1. First, from embryology, by analogy. The develop-

ment of the embryo is directed from within. The process

of development is one of specialization; a great number

of tissues is produced but these tissues have lost the

capacity, which the embryo has, of producing all kinds

of tissues. Development is essentially an irreversible

process just as evolution is, and for the same reason—

that a fragment can not produce the whole. The adult

individual is more complex than the fertilized egg, yet

the egg has greater potentialities than any tissue-cell has.

Regeneration depends on the presence of embryonic, i. e.,

non-tissue, cells lying latent amidst the tissue cells. The

greater complexity of the adult as a whole over the egg

does not hold for a given tissue cell—that is less complex

than the egg. The complexity of the adult is due to the

fact that there are in the body many kinds of tissue cells,

each simple, while the egg is just one kind of cell—but

very complex in constitution. Similarly, we may infer

that while the vast number of kinds of germ plasms in

the higher organisms, differentiated in respect to their

chromomeres, contrasts with the condition in the pro-

tista, it is prebable that each chromomere of the protista

is composed of much more complex organic molecules or

molecule complexes.

2. Another mass of evidence for this theory is sup-

plied by paleontology. By this science are offered ex-

tensive series showing: (a) the usual beginning of a new

character as a simple, often inconspicuous trait, (b) the

increase of variety in the course of evolution of a phylum

ending in great outburst of extreme and bizarre forms

immediately preceding the extinction of a phylogenetic

line; (c) the irreversibility of the process of evolution;

and (d) parallelism in evolution of allied lines.

(a) The Simple Beginning of a Trait.—The early his-

tory of a number of spinose groups of species shows

(Beecher, 1898) that each group began its history ım

454 THE AMERICAN NATURALIST [Vou. L

small, smooth or unornamented species. The septe of

ammonites begin simple and later evolve their extraordi-

nary foldings. The horns of titanotheres ‘‘have exces-

sively rudimentary beginnings phylogenetically, which

can hardly be detected on the surface of the skull’’ (Os-

born, 1912, p. 253). Also, these ‘‘rudiments arise inde-

pendently on the same part of the skull in different phyla

at different periods of geological time.’’

(b) The History of Progressing Phylogenetic Develop-

ment.—While paleontologists have no knowledge of

germinal conditions they can study the course of evolu-

tion of a particular character through a lineage compris-

ing thousands of generations. Paleontologists are

agreed that characters tend to become more and more

complex. So Beecher (1898, p. 354) writes:

The smooth, rounded embryo or larval form [terms used in the phylo- .

genetic sense] progressively acquires more and more pronounced and

highly differential characters through youth and maturity. In (paleon-

tological) old age, it blossoms out with a galaxy of spines, and with

further decadence produces extravagant vagaries of spines. So in the

titanotheres the horn rudiments evolve continuously, and they gradually

change in form, . . . they finally become the dominant characters of the

skulls, showing marked variations of form in the two sexes.

' The saber-toothed and other tigers gained canines that

they could not use. The mollusc, Hippurites, gains a

shell a foot thick. In the Labyrinthodonts the infolding

of the teeth has been carried to an extraordinary degree,

ete. These are illustrations merely of what is said to be —

a general rule. F. B. Loomis! writes of it under the

head ‘‘Momentum in Evolution.’’

has been often remarked upon. It leads to exaggerated

developments in one direction; lateral and backward

variations are relatively uncommon.

Thus, the horse has shown divergent lines of evolution

but none returning to the four-toed ancestral type. The

ammonites went from simpler forms of septum to more

and more complex without reversal, except at the very

end of their phylum. D. Rosa says:

1 AMER. Nat., 39 (1905). :

No. 596] EVOLUTIONARY THEORY 455

An organ which in the course of its phylogenesis once disappears has

disappeared for ever... . Not an exception is known to the rule.

Even an organ once rudimentary, like flying in some

ground birds, never returns to full activity. A striking

example of dropping out is that of cilia in Arthropods;

which, ubiquitous in other groups of animals, are in this

group gone throughout; they fail to develop even in

spermatozoa. Toward the end of the process of evolu-

tion, a character tends to break up into a great number

of new characters. These usually affect a certain organ,

like the suture of the ammonites, and precede extinction

of the phylum. We may say that the number of genes

has become, through fractionation, so great that the re-

sulting complexity or the resulting extremes of develop-

ment of a trait are prejudicial to the development of the

organism. or

(d) Parallelism of Evolution in Allied Lines.—This

parallelism was recognized by Darwin who writes:

The principle formerly alluded to under the term of analogical vari-

ation has probably in these cases often come into play; that is, the

members of the same class, although only distantly allied, have inherited

so much in common in their constitution, that they are apt to vary

under similar exciting causes in a similar manner; and this would ob-

viously aid in the acquirement through natural selection of parts or

organs, strikingly like each other, independently of their direct in-

heritance from a common progenitor.

Of such parallel variations Osborn (1915, p. 216) speaks

as follows: :

Similar rectigradations may arise in all the descendants of similar

ancestors at different periods of time; they always give rise to parallel-

ism or convergence between the members of related phyla. |

Tilustrations of such are afforded in many paleontological

monographs.

3. Another line of evidence for the theory of the

primacy of internal factors of evolution is found in ez-

perimental breeding. This evidence appears in the facts

(a) that many mutations begin small and ean be rapidly

evolved into highly developed characters, (b) that simi-

lar variations appear in related organisms, (c) that

456 THE AMERICAN NATURALIST [Von L

mutation is limited to certain lines and (d) that experi-

mental evolution seems chiefly due to dropping out of

enes.

(a) That characters often arise as rudiments and only

in the course of generations realize their full potentiality

is a well-known experience of breeders. ‘Thus De Vries?

states that the double Anemone coronaria was produced

by the owner of a nursery who, observing in his beds a

flower with a single broadened stamen, saved its seeds

Separately and in the succeeding generations procured

beautifully filled flowers. By appropriate matings

Castle succeeded in 5 generations in getting much better

expressed polydactylism in guinea pigs than he had at

the outset. Dr. F. E. Lutz (1911) found a slight ab-

normality of venation in the fruit fly, Drosophila am-

pelophila. By in-breeding abnormally veined flies and

selecting as breeders the extremely abnormal flies he

eventually secured in the later generations some highly

abnormal individuals. From a hen that showed only a

slight extension of the web between certain toes I suc-

ceeded in breeding a race of profoundly syndactyl de-

scendants.

(b) That mutations (‘‘saltations’’) run in parallel

lines in related species is well brought out in a table

given by Osborn (1912, p. 191), which I reproduce here

with certain modifications.

That each germ plasm can vary only within certain

limits and that related germ plasms show only a limited

number of variations and the same in the different

species indicate that variations of the specifie rank are

not determined by anything outside the organism, but by

the very nature of the organism. Thus, the rabbit shows

in its coat color the agouti coloration, and so does the

guinea pig. From this in the guinea pig have arisen

yellow, chocolate, black, albino and other colors; and a

similar series has been obtained for rabbits. The guinea

pig has produced an angora coat and so has the rabbit,

2 ‘Species and Varieties,’ p. 491.

No. 596] EVOLUTIONARY THEORY 457

COMPARATIVE TABLE OF SALTATIONS

CBRE Nn be Ge Oe ae Se) | 10 | 11

Siga

; 3 Aj B32 i | 8

E Hi o|@ : 4 H |g Se 3

i “1. Proopie erbedi abbrevia-

Pa EEE E SE x x

2. Sudden “develope of horns on

OGG. ovis outs seams iy x eee Ne x

3. Absence a ase on horned races. apse A Gk D Grae

4, Jaw ap Be eee rs he tet eA Ea Oras Wey ae 2

5. Taillessness, absence of caudals. >TO SOUND A Sp x x

6. gee “leggedness, or limb abbrevia- :

T EAEE EE E E XEX ia Kii

+ Consolidation of paired hoofs, syn- ‘

sib W8ie ei 6a alco ea A eee 5 Ge ON A A

8. pe ry AES E E E WER T KIRIATA ILAR PALAI KI TN s

9. Epidermal npn ec ee hawt KISRA Pas Peele lie as x

10. ttled skin Pings 2). oa x| xX x .

ir, yanin hairiness, p length of

E AA ov E AN X EATX x1 x1 xX) xX

12. Haicleasnese, entire absence of

Halt ETN eo es MOT eel CO Oe re be T ee

13. Excessively fine or silky hair.... . MP RSE RHE RA SE E OCP ee

14. Reversed hairs: ...: 2.2... -. 3c P EO ese ech wo) wee Pet

15- Curled-hatr 200.) Sa ee WTX | ERN Bn cs Pe fs a

the cat, the dog, the goat, the sheep (Lincoln) and others.

So in different species of Drosophila, I am informed by

Dr. ©. W. Metz, the same mutations oecur—of course,

without any relation to environment.

In a strain that has produced a mutation once we are

apt to find the same mutation a second time. Thus De

Vries states that the peloric toadflax does not set seed in —

nature, yet it occurs repeatedly in a given locality and

even in distant localities. We conclude there is some-

thing in the structure of the germ plasm of the toadflax

that permits a wholly useless, and indeed not naturally -

perpetuated, mutation to occur easily. Similarly, doub-

ling of the floral parts occurs again and again in wholly

unrelated species; often combined with complete sterility. —

Again, poultry sometimes show a great extension of the

web between the toes. This occurs chiefly between digits

III and IV, exactly where the syndactyl web occurs in

man and where it occurs in wading birds that have only

a single web between the toes—of which 8 goera e ee Oe

458 THE AMERICAN NATURALIST [Vou. L

be named. So far as I know, a marked extension of the

web between digits I and II has never been observed;

between II and III it is relatively rare; again, between

IV and V it has, so far as I know, not been noted in

mammals.

case of syndactylism in poultry well illustrates the gen-

eral principle of the limitation of mutation to particular

narrow lines; and this is commonly the case. Thus, Dr.

J. A. Harris has examined himself and with the assist-

ance of others over 1,000,000 bean seedlings, and while ~

extraordinary variations have been found, yet in the

later hundreds of thousands no new ones have appeared;

nevertheless, the possibilities in leaf form, variegation,

etc., if we may judge from books on plant teratology, are

by no means exhausted. And when we contemplate the

variety of form assumed by first leaves in different

species, is not the constancy in the form in beans striking

evidence of the narrowness of variation and its restric-

tion to certain lines? :

(d) Evolution by Loss of Genes.— Finally, it has long

been recognized that an extraordinarily large proportion

of the mutations we meet with are recessive to the wild

type. Castle has noted this of rabbits and guinea pigs?

and Morgan has noted it in the case of Drosophila. In

my work with poultry I was not impressed with it, as

taillessness, polydactylism, syndactylism, white color of

the Leghorn, rose comb, etc., are dominant mutations.

Still, other genetics work (Baur, Shull and many others)

is strengthening the conclusion that current evolutionary

changes under observation are chiefly due to the dropping

out of genes, and this supports the theory that evolution

is proceeding largely by the loss of genes. However, the

fact of dominant mutations can not be derived. Take,

for instance, foot abnormality. Here a disturbing factor

has appeared in the organism that was not there before.

There is no a priori reason for doubting that the break-

3‘ Heredity,” p. 86. 3

No. 596] EVOLUTIONARY THEORY 459

ing up of one of the genes that makes for the normal foot

might have a residuum that positively interferes with the

operation of other foot-forming factors.

The genetical studies also accord with other evidence

as to the general irreversibility of evolution. We see

that when a population contains only a recessive trait it

is impossible by breeding inside that population to get

back the dominant allelomorph. From a pure blue-eyed

population we do not get back the primitive brown-eyed

condition.

It might be thought that if evolution proceeds chiefly

by loss of factors it would tend to simplification rather

than increase of the number of factors. It seems prob-

able, however, that the loss is not merely of a whole gene,

but of some part of it; a fractionation, as it were, by

which the gene becomes altered or split up into two or

more. For example, it is probable that in man the loss

of a gene (or a part of one) releases special nervous

States, as, for instance, that in which musical combina-

tions run through the brain, or numerical relations are

rapidly worked through, ete. Thus the general result of

experimental work in genetics gives support to the view

of evolution by loss of genes and by their fractionation.

4. Evidence from Evolutionary Changes in the Inor-

ganic World. Radiation Studies.—The view that evolu-

tion is primarily by internal changes receives unexpected

support from the recent discoveries concerning the evolu-

tion of the elements. It is now well known that the ele-

- ment, uranium, under certain conditions of temperature,

ete., undergoes a spontaneous change into ionium, ionium

into radium, radium into polonium, an essentially lead-

like REUE O ‘Similarly, thorium passes through

mesothorium and radiothorium. Indeed, in a particular

series, side branches may be given off; thus polonium is

not derived directly from radium, or from radium C, but

from one form of radium C called radium C,. deed,

‘‘radiums’’ of different sorts, called radium 4, radium

B, radium C, are recognized and the atoms of these differ

460 THE AMERICAN NATURALIST [Vou. L

from one another by, probably, one electron lost from

each successive stage of the series. The series is thus

shown graphically by Sieveking (1913):

_, RaC,

Ra > Em > Rad > RaB > RaC, > RaD —> Raz > Raf

ILI. CERTAIN CONSEQUENCES OF THE THEORY

1. The acceptance of this theory requires a special

explanation to account for adaptation. Eimer’s theory

of orthogenesis posited the direct action of environment

on the germ plasm, a view which wider knowledge of

facts does not support. It follows naturally from the

hypothesis that new traits bear, at first, no relation to

environment any more than the polonium that is derived

from the uranium does. Darwin recognized that varia-

tions were not necessarily adaptive in their origin; also

that it was not necessary that they should be adaptive in

order to survive. Darwin says:*

We clearly see that the nature of the conditions is of subordinate im-

portance in comparison with the nature of the organism in determining

each particular form of variation.

How then is adaptation brought about? Strictly, we

may say adaptation is not the thing that is brought about,

but rather absence of non-adaptiveness. Such adjust-

ment as we find is, doubtless, only such a residuum of

variants as has not proved incompatible with conditions

of existence. Two kinds of variations may survive:

(a) Those not incompatible with the conditions of the

present environment and (b) those which, while incom-

patible with present environment, are not incompatible

with some other environment into which the species may

migrate.

2. Relation to the Réle of Selection in Evolution.—

There is going on to-day a great discussion as to the

importance of selection in evolution. How does the

matter look from the standpoint of this theory.

First, all are agreed that nothing has importance for

***Origin of Species,’’ p. 9.

No. 596] EVOLUTIONARY THEORY 461

evolution of the race except what modifies the germ cells,

because they are all that goes from one generation to the

next. However, we learn little or nothing about the

potential traits of the germ cells by looking at these cells.

Our knowledge of their hereditary composition depends

upon the traits shown by the individuals that develop

out of them. We may infer the genotype by observing

the phenotype. But -the phenotypical condition of a

person is a more or less imperfect index to the geno-

typical condition of that person. The soma is an imper-

fect index to the germ plasm. The difference between

the two schools, one asserting, the other denying the value

of ‘‘selection,’? is based primarily upon the reliance

placed on the sufficiency of this index. Castle says, in

effect, the somatic condition of my rats in respect to the

coat-pattern is so good an index of their germinal condi-

tion that whenever I select rats for a quality of the

pattern I am selecting them very closely for the cor-

responding quality of their germ plasm. Pearl says, in

effect, in poultry the egg-laying capacity of the hen is

hereditary; yet it is so poor an index of her germinal

idiosyncrasies in this respect that an individual with the

somatic characters of high laying is no more apt.to have

the genes for high laying than an individual with the

somatic character of low laying. Pearl says to make

progress I must select for breeders those which have `

proved their germinal quality by belonging to a race of

high layers. If the various sisters and daughters of the

hen are high layers, that is more important than the egg-

laying performance of the one individual, merely.

It seems to me that the whole question is a pragmatical

one. If the somatic condition of a trait is a good index

of germinal conditions, in any case progress can be made

by selecting the soma in that case; if the soma is an in-

adequate index, then little or no progress will be made by

selecting the soma. We know this by breeding experi-

ence. If we select parents because of absence of pigment

in skin or eye, we select also a germ plasm devoid of the |

462 THE AMERICAN NATURALIST [Vou. L

capacity for forming such pigment. If we select parents

because they have medium-brown hair we get offspring

with blond or golden or red hair even, and it would take

a long time to get a pure brown-haired race by this

method of selecting. If the trait that we are trying to

improve is very sensitive to environment, then the somatic

conditions will be a very bad index of the germinal; but

if not sensitive, the somatic may be a good index of the

germinal. Thus, since in the fruit fly the number of

bristles varies closely with food conditions, selection of

breeders merely on the somatic state will not lead to

much, if any, genetic progress.

The question of the potency of selection in nature

comes back to the matter of value of the soma as an index

of the germ plasm. If blue eye color affords insufficient

protection from tropical conditions elimination of the

blue-eyed individual kills off blue-eyed plasm and with

each death the race is rapidly purified. But there is a

limit to the process because brown eyes are preserved

equally whether the germ plasm does or does not carry

the blue-eyed condition. As there are twice as many

simplex as duplex brown eyes, the complete elimination

of the blues will not be brought about quickly.

When, therefore, we hear a breeder say: ‘‘I made this

character by selection’’ what he really means is: Somatic

‘ variations in the desired direction were afforded; there

was a large correlation between somatic and germinal

conditions, so that I was able, merely by choosing as

breeders individuals showing the desired trait somati-

cally, to get a race with the determiners of the character

pure, or practically pure, in the germ plasm.

One other difference of opinion there seems to be be-

tween selectionists and the others. Castle evidently

doubts if factors for characters are always discrete and

do not change. He is inclined to hold that there may be

genes which vary pari passu with a variation of a ‘‘unit

character’? and in somewhat the same degree. This —

view seems to be quite in accord with an expectation that —

No. 596] EVOLUTIONARY THEORY 463

is based on experience, that traits shall be found the

germinal bases of which are undergoing current evolu-

tionary changes through loss of genes or through frac-

tionation. And it is in accord with expectation that

mutations shall reveal themselves just at the extreme of

a series as Castle’s plus mutation revealed itself. Others

are inclined to think that the varied color pattern of

Castle’s rats is really determined by several factors (re-

strictors or extensors of the gene for ‘‘hoodedness’’)

which were all in his race of rats at the beginning of his

experiment. Upon one point all geneticists are, how-

ever, agreed—that we must interpret all of our results in

terms of genes alone.

One other bearing of the orthogenetic theory deserves

‘to be pointed out. If the germ plasm is capable of

undergoing a spontaneous mutation which is the main

source of evolutionary change, this fact would seem, at

first blush, to indicate the futility of trying to control

genetic change experimentally, except by the selection of

germ plasms. Naturally, under these circumstances our

effort would be limited to what nature affords. How-

ever, we do not yet know enough to put these limits on

‘Cexperimental evolution.” There is some evidence,

although not as critical as might be wished, that the

germ plasm is not beyond the reach of modifying agents.

At least we must continue experimental efforts in that

direction.

IV. SUMMARY

A theory of evolution that assumes internal changes

chiefly independent of external conditions, E é, spon-

taneously arising, and which proceeds chiefly by a split-

ting up of and loss of genes from a primitively complex

molecular condition of the germ plasm seems best to

meet the present state of our knowledge.

Such a theory receives support from various fields.

1. From ontogeny, where the differentiated end stage

is derived from a relatively undifferentiated, but prob-

ably molecularly complex egg.

464 THE AMERICAN NATURALIST [Vou. L

2. From paleontology, where the history of the phylum

seems governed by internal laws.

3. From experimental breeding where progress is

afforded only as internal changes permit.

4. From analogy, with evolution in the inorganic world,

so far as may be inferred from the studies on the ‘‘rare

earths.’’

Such a theory makes clear that success in ‘‘selection’”’

depends on rate and amplitude of internal change and

ability to judge of germinal from somatic conditions.

It renders less hopeful (but not hopeless) the prospect

of being able to control completely by experimental

methods evolutionary change.

ARNEGIE INSTITUTION OF WASHINGTON,

STATION FoR EXPERIMENTAL EVOLUTION,

COLD SPRING HARBOR, LONG IsLAND, N. Y.

V. LITERATURE CITED

ioh, C. E. 1898. »The Origin and Significance of Te Amer.

Jou , VI, 1-20; 125-136; 249-268; 329-359. July—

wean E. "1913. Oneative ivalution. Transl. by A. Leeched 407

pp. si be

Castle, W. E. 1914, Piebald Rats and Selection. Publ. 195. Carnegie

Inst. of Washington, 56 pp. la

Castle, W. E. 1915. Some eee = in Mass Selection. AMER. NAT.,

49: 713-726. Dec.

Castle, W. E. 1916. Can Selection cause Genetic Change? AMER. NAT.,

50: 248-256. April.

deVries, H. 1906. Species and Varieties: Their Origin by Mutation. Ed.

y D. T. MacDougal. 2d ed. 847 pp. Chicago, Open Court Publ.

Hagedoorn, A. L. Autokatalytical Substances the Determinants for the

Inheritable Characters, a Biochemical Theory of Inheritance and Evo-

lution. Vortr. u. Aufs. über Entwicklungsmechanik der Organismen,

von Roux, Heft 12.

Loomis, F. B. 1905. Momentum in Evolution. AMER. Nat., 39: 839-844.

Lotsy, J. P. 1913. Fortschritte unserer Anschauungen über Deszendenz

seit Darwin und der jetzige Standpunkt der Frage. Progressus Rei

Botanicæ, 4: 361-388.

Lutz, F. E. 1911. RAEE with Drosophila ampelophila Concerning

Evolution. Publ, 143. Carnegie Institution of Washington.

Mivart, St. G. 1871. The Genesis of Species. London and N. Y., Mac-—

millan. 333 pp.

No. 596] EVOLUTIONARY THEORY 465

Nägeli, C. 1865. Entstehung und Begriff der Naturhistorischen Art.

2te A PENA 55 pp.

Nägeli, C. v. 1898. A Morpema E Theory of Organic Evo-

lution. tran by s A. Waugh. Chicago, yi Court Publishing Co.

eY H; 912. The Continuous Origin of Certain Unit Characters

5 Uaa và the ae Amer. NAT., 46: 185-206; 249-278.

io. —May.

Osborn, H. F. 1915. Origin of Single Characters as sou in Fossil

3 and Living Animals and Plants. AMER. NAT., 49: 193-239. April.

Pearl, R. 1916. m si in the Domestic Fowl and the Selection Prob-

lem. Amer. Nat., 50: 89-105. Feb.

Sieveking, H. 1913. os der jetzigen Stand der DER iiber Radio-

aktivität. Naturwissenschaften, I: 129-133. Feb.

COMPARATIVE RAPIDITY OF EVOLUTION IN

VARIOUS PLANT TYPES

PROFESSOR EDMUND W.’ SINNOTT

CONNECTICUT AGRICULTURAL COLLEGE

Durtne the course of evolutionary development among

the higher plants certain groups have evidently altered

with such exceeding slowness as to retain their ancient

constitution practically intact for very long periods;

whereas others, by their more rapid accumulation of

heritable variations, have during the same time under-

gone far-reaching changes and become developed into

new and radically dissimilar types. These differences in

the rate of evolution are apparently due in part to dif-

ferences in mutability, in the extent to which hybridiza-

tion has occurred, in the degree of diversity presented by

the environment or in the keenness of the struggle which

is waged for survival. The purpose of the present paper

is to call attention to still another factor which seems to

be of much importance in determining the rapidity of

evolutionary change among plants, namely the length of

the generation or period from seed to seed.

The time necessary for the attainment of reproductive

maturity among plants is definitely correlated with the

growth-type to which a species conforms. Trees are very

slow to mature, some arriving at an age of eighty years

or more before their first flowering time and very few in- |

deed, under natural conditions, fruiting before the tenth

season. The scanty available observations seem to indi-

cate that the average period from seed to seed in arbores- —

cent forms is in the vicinity of twenty years. Among-

shrubs the generations are decidedly shorter, varying

usually from three to ten years. The herb is the most

rapidly maturing of all, a single year sufficing to develop _

seed in an annual and but two (rarely more) ina biennial

or perennial. Most herbaceous species will thus have

466

No. 596] RAPIDITY OF EVOLUTION 467

from fifty to one hundred generations a century, most

shrubby ones from ten to thirty and most trees only

four or five. The degree of variability and other factors

being equal, therefore, one would expect changes to accu-

mulate much more rapidly and evolutionary progress

consequently to be much faster among herbs than among

woody plants. Is there evidence that this is actually the

case?

In an attempt to obtain such evidence an analysis as to

growth-habit was first made of the endemic portion of the

floras of several regions. Upon the biological isolation .

of an area the new varieties, species and genera of plants

which gradually take their origin in its flora are neces-

sarily limited in their distribution to the region in ques-

tion, or are ‘‘endemic”’ in it; and these local types will

evidently be produced first and in greatest abundance by

those elements in the flora which are changing most

rapidly. Consequently that particular growth type which

is found to predominate among such endemic forms may

justly be regarded as the one whose members are under-

going the most rapid alteration. In an analysis of the

endemic element in any flora, however, caution must be

used to distinguish carefully two radically different types

of endemic plants: those under discussion, which were

local in origin and have never spread abroad; as con-

trasted with those which owe their present localization

rather to the fact that they are isolated survivors of

genera or species at one time much more widely distrib-

uted. The former category, which we may call the ‘‘in-

digenous’’ endemics, will evidently represent a new ele-

ment in the flora; the latter, or ‘‘relict’’ endemies, a very

old one. It is not always easy to separate sharply these

two types in a given flora, but species, and especially

genera, which stand apart and possess no near relatives

in the region are, in most cases, at least, evidently to be

looked upon as relicts; whereas in a group of species or

genera the members of which are numerous and closely

related one to another we doubtless behold a body of

plants undergoing evolutionary development on the spot.

468 THE AMERICAN NATURALIST [Vou. L

It is an analysis only of the latter type which is of sig-

nificance for the present problem. ,

The genera of dicotyledons endemic’ in temperate North

America (Canada, the United States and northern Mexico)

and in Europe (inclusive of. the entire Mediterranean

floral province) were studied in this connection. 400

genera were recorded as being endemic or essentially so in

temperate North America. In this imposing array of

types which are limited in their distribution to this region

approximately 130 stand in isolated positions in the flora,

being quite without near relatives, and are presumably

‘‘relicts.’’? To this category belong Carya, Planera, Mac-

lura, Garrya, Sassafras, Xanthorrhiza, Baptisia, Nemo-

panthus, Ceanothus, Dirca, Dionaea, Hudsonia, Rheaxia,

Ptelea, Decodon, Houstonia, Symphoricarpos and many

other familiar plants. That these now exclusively Amer-

ican genera were indeed at one time much more widely

distributed is indicated by the fact that many of them are

found as fossils in Europe and Asia to-day. They un-

doubtedly represent a very ancient element in the flora.

The most noteworthy feature of these ‘‘relicts’’ is that

they include practically all the genera of woody plants

endemic to this region, the typically American trees and

shrubs.

Such ancient and isolated relicts, however, compose but

a minority of the endemic genera. The remainder evi-

dently belong to the other category and owe their ende-

mism rather to the fact that they have been developed in

America and have never spread beyond its borders. It is

such endemic types which are of interest to our problem.

That they are actually of local origin is rendered probable

by their occurrence in groups of closely related genera

(many of which are rich in species) each group presum-

ably representing a separate center of evolutionary de-

velopment and the nucleus for a new subfamily. There

1 Genera in which 90 per cent. of the species or more are confined to the

region in question were considered as endemic there. Genera rather than -

species were studied since they represent a greater degree of evolutionary

change; and since endemism among species is so nearly universal on the

two continents as to make its investigation of little value.

No. 596] RAPIDITY OF EVOLUTION 469.

are over sixty such groups of allied genera in the endemic

flora of North America. Notable among these are the

alliances centering about Hriogonum in the Polygonacex;

about Streptanthus and about Lesquerella in the Crucif-

ere; about Eschscholtzia in the Papaveracee; about Heu-

chera in the Saxifragacee; about Cercocarpus in the

Rosacee; about Godetia in the Onagraceæ; about Cy-

mopterus in the Umbellifere; about Pterospora in the

Pirolacee; about Cryptanthe in the Boraginaceæ; about

Trichostema and about Agastache in the Labiate; about

Pentastemon and about Castilleja in the Scrophulariacee,

and about Brickellia, about Solidago, about Boltonia,

about Silphium, about Rudbeckia, about Hemizonia, about

Baeria, and about Microseris in the Composite. Most of

these groups of genera have their center of distribution

in the southwestern United States or in northern Mexico.

The great majority of endemic genera in Europe are also

evidently ‘‘indigenous’’ rather than ‘‘relict’’ in charac-

ter. Centering chiefly in the Mediterranean region there

are seventy or more groups of closely allied genera, no-

table among which are the alliances dominated by Dian-

thus, Brassica, Alyssum, Lotus, Scandix, Asperula, Scabi-

osa, Anchusa, Anthemis, Carduus, and Cichorium.

These ‘‘indigenous’’ endemic genera constitute a very

characteristic and important part of the present-day flora

of North America and Europe. Had they existed in any-

thing like their present numbers and importance at the

period when an easy exchange of plants was possible be-

tween the two northern continents, it is hard to believe

that they would not now be well represented in the floras

of both; particularly since herbs, of which we shall see

these genera to be almost exclusively composed, tend more

quickly than any other plant type to loose their endemic

character because of the power to migrate rapidly and

populate wide areas which is conferred by their ability to

produce seed from seed in a single generation. This is

well shown by the fact that in DeCandolle’s famous list of

117 species, each of which at present occupies at least one

half of the land area of the globe, there are none but her-

470 THE AMERICAN NATURALIST [Vou.L +

baceous forms. Since we have cause to believe that com-

munication between North America and Europe existed

until well into the Tertiary, the supposition is altogether

reasonable that the genera in question have undergone at

least the greater part of their development and dispersal

since that time; and that in contrast to the ‘‘relicts’’ and

the non-endemic genera they represent an element of very

recent development in the floras of the two regions.

It is noteworthy that these indigenous endemic genera

are composed almost exclusively of herbaceous species,

Cercocarpus and its allies among the Rosacee furnishing

practically the only exception to the rule in America. If

the conclusion is correct that such endemic types are the

most recently developed members of a flora, this domi-

nance ôf the herb among them constitutes excellent evi-

dence that it is the herbaceous element which has indeed

been undergoing the most rapid evolutionary change.

In striking contrast to the highly endemic and local

character of so many of the north temperate genera of

herbs is the wide geographical range almost universal

among the woody types. Nearly all the tree genera of

temperate North America exist to-day on some portion of

the Eurasian continent or give evidence by fossils that

they once did exist there. The whole study of endemism

in its relation to growth forms presents us with the piĉ-

ture of a very slowly changing woody vegetation, one

which since the separation of the two northern land

masses has given rise to few or no generic types, but

which has been accompanied by a rapidly developing her-

baceous flora so quick to originate new forms that upon

isolation it has produced not only a throng of local species,

but even a goodly number of genera.?

Evidence of value in the present problem may also be

derived from a study of the relationship of the members

2 The predominance of woody plants which is so pronounced in the en-

demic element of the floras of oceanic islands and of somewhat isolated con-

tinental areas in the south temperate zone, is evidently due to the fact that

most of the herbaceous portion of the vegetation here has so recently arrived

from its seat of origin in the great land areas of the north, that it has 1r !

yet had time to develop into endemie species and genera on a large seale.

No. 596] RAPIDITY OF EVOLUTION 471

of the various growth forms and their distribution in the

modern system of plant classification. During the his-

tory of the vegetable kingdom, the continual evolution of

species, genera and families has been opposed by their

continual extinction, for one cause or another. One would

expect that those types in which evolution was proceeding

most rapidly would occur in groups, usually rather large,

of closely related species and genera, and that members

of such groups which subsequently came to be isolated by

the extinction of their allies would soon become centers of

development and give rise again to new groups. Mono-

types would consequently be rare among them. The more

slowly evolving forms, on the other hand, would be able

to repair the ravages of extinction much less rapidly and

easily and would therefore tend to occupy more or less

isolated positions in the system, frequently as monotypic

genera or families.

To determine the distribution of herbs as contrasted

with woody plants in the present scheme of classification

an analysis was made of the dicotyledons in Engler and

Prantl’s ‘‘Naturliche Pflanzenfamilien,’’ supplemented

and brought up to date as far as possible by the seventh

edition of Engler-Gilg’s ‘‘ Syllabus der Pflanzenfamilien.’’

The figures obtained of course can not be regarded as

exact, but they are at least definite enough to bring out

certain general facts. ,

One hundred and eight thousand species of dicotyledons

were counted, grouped in 6,840 genera and 238 families ;

4,030 genera, comprising slightly over 50,000 species, were

found to be composed entirely of woody plants, and 2,630

genera, with slightly under 40,000 species, entirely of herbs;

180 genera, containing over 18,000 species, included both

woody and herbaceous members and were disregarded in

the count.? The average number of species in the woody

genera is therefore 12.5, in the herbaceous ones, w. In

their large genera (including 10 species or more) the two

3If 90 per cent. of the species of a genus were woody, that genus was

counted as ‘‘woody’’; if 90 per cent. were herbaceous, it was counted as

‘‘herbaceous.’? ‘Mixed’? genera include more than 10 per cent. of each

type. : : ;

472 THE AMERICAN NATURALIST [Vou. L

types are practically the same, there being an average of

45 species per genus in the former and 46 in the latter.

This of course pulls the general averages together. The

more scattering distribution of trees and shrubs, however,

is made evident by the fact that of small genera (10 spe-

cies or less) they possess 3,115, as compared with 1,890

for herbs.

The two types are somewhat more diverse in the num-

ber of genera per family. In an analysis to determine

this distribution, only land plants of normal growth-

habit were considered. The Balanophoracex, Rafflesi-

aceæ, Hydnoracee, Lennoacee and Cynomoriacee were

excluded as aberrant; and families characteristically

aquatic were also omitted, since from their uniformity of

environment or other reasons unknown they are notori-

ously poor in generic types. Two hundred and twenty-

four families of dicotyledons enumerated in the seventh

edition of Engler’s ‘‘Syllabus’’ remain to be considered;

130 of these are exclusively woody, 60 have both woody

and herbaceous members, and only 34 are exclusively her-

baceous,* making it evident at a glance that woody plants

have much the wider taxonomic distribution. In the 190

families which include trees and shrubs, there are just

4,000 genera, an average of 21 per family. In the 94 fam-

ilies which include herbs, there are 2,590 genera, an aver-

age of 27.5 per family. (The 180 ‘‘mixed’’ genera were ©

left out in this count.) The fact that the bulk of genera in

both are massed in a few large families again pulls the

averages together; but the more scattering distribution of

woody genera over a large number of small families is

decidedly emphasized when we note that there are no less

than 39 monogeneric families of trees and shrubs, but only

7 of herbs (exclusive of aquatics). Of families with 5

genera or less there are among woody plants 83, among

herbs only 24. In number of species per sree the differ-

4 Families like the Ranunculacee, Papaveracee, Crassulacer, Gerani

and Umbellifere, which possess exceedingly few woody oe have been

counted as strictly herbaceous; and others like the Sapindacee,

and Bignoniacee, in which there are very few herbs indeed, have been

counted as strictly woody.

No. 596] RAPIDITY OF EVOLUTION 473

ence between the two types is still more marked, for the

190 woody families include about 59,000 species, an aver-

age of 310; the 94 herbaceous families about 49,000 spe-

cies, an average of 510.

It is therefore evident that herbs tend to be massed in

a comparatively few large genera and families. Among

trees and shrubs, on the other hand, although the majority

of species are also naturally in large, successful groups,

there is a very much greater proportion of small genera

and especially of small families, widely scattered through-

out the whole taxonomic range, which have been isolated

by the wholesale extinction of related forms and are ap-

parently very slow to develop into larger aggregations.

Such evidence as this, like that derived from a study of

indigenous endemism, seems to indicate that the rate of

evolution among herbs is decidedly higher than among

woody forms.

There are doubtless numerous exceptions on both sides.

Crategus, Eucalyptus, Acacia and other genera of trees

and shrubs appear to be rapidly developing new species,

whatever the cause thereof may be; and many herbs, far

from producing new forms, show every indication of be-

ing stationary or even of becoming extinct. These are —

both evidently exceptions to the general rule, however.

Of course these cases introduce the consideration of

another factor, naturally thought of first in connection _

with rate of evolution among any organisms, namely,

their comparative ‘‘variability,’’ using the term in its

broadest sense. To contrast members of the two growth

types as a whole in this respect is necessarily very diffi-

cult, but an attempt was made to do so by comparing the

proportion of varieties and named forms among the

woody plants with that among the herbs in a number of

floras. The results are shown in the table (p. 474).° |

The proportion of varieties and forms is therefore

practically the same among woody plants as it is among

herbs, and if this is to be regarded as at all a criterion of

variability, there is little to choose between the two growth

5 Dicotyledons only are considered. sere eee Ss

474 THE AMERICAN NATURALIST [Von. L

Species Varieties and Forms

Woody | Herbaceous | Woody | Herbaceous

N: E. United States....| 532 (23%) | 1,748 (77%) | 144 (31%) | 322 (69%)

CSUN hs Bee oes 1,123 (63%) | 670 (37%) | 165 (60%) | 112 (40%)

Australia. .............| 3,970 (70%) | 1,741 (30%) | 752 (70%) | 325 (30%)

forms. The greater plasticity. probably belongs to the

herb, owing to the larger independence of environment

which its discontinuous existence allows. Brevity of life-

eycle, however, rather than higher variability, is probably

the cause of the more rapid rate of change exhibited by

members of this type.

The conclusion that herbs have been evolved much more

rapidly than trees or shrubs has certain important cor-

ollaries. There are at present approximately 2,600 gen-

era of dicotyledonous herbs and 4,000 genera of woody

plants. The much wider taxonomic range which we have

noted in the latter type makes it probable that in extinct

species and genera their numerical superiority is still

greater. If herbs are thus decidedly fewer than woody

plants in number of species and genera, but are neverthe-

less being produced at a much more rapid rate, it is highly

_ probable that the herbaceous element in the flora of the

world must have had a shorter evolutionary history than

the woody one. This is in harmony with a considerable

body of evidence derived from a study of the history,

structure and distribution of the Angiosperms, which in-

dicates that the most ancient members of the group were

woody and that herbaceous vegetation has made its ap-

pearance in comparatively recent geological time.® The

great steps in the evolution of the higher plants seem to

have taken place while trees and shrubs were the domi-

nant growth types, for all the important orders and the

great majority of families are still composed wholly or in

part of such plants. Herbs oceur to-day in less than half

the families of the dicotyledons, and although such a

dominant portion of the vegetation in many regions, they

seem to be of relatively recent appearance. |

6 Sinnott, E. W., and Bailey, I. W., ‘The Origin and Dispersal of Her-

baceous Angiosperms,’’ Annals of Botany, XXVIII, 1914, pp. 547-600.

No. 596] RAPIDITY OF EVOLUTION 475

A study of the comparative rate of evolution among the

Angiosperms also throws a little light on the antiquity of

this great group. Its fossil representatives have as yet

failed to make their appearance in strata lower than the

lowest Cretaceous, and it is commonly assumed that they

had their origin at about that period. We have already

noted the probability that they were represented at their

inception only by woody forms. There is evidence that

the herbaceous type, save perhaps as a negligible portion

of the flora, did not arise till the early Tertiary, since

which time its members have undergone practically their

entire evolutionary development. As to the relative

lengths of geological periods there is no very definite

evidence, but authorities agree that the Mesozoic was of

considerably longer duration than the Tertiary, a con-

servative estimate placing the former at about three

times the length of the latter. Let us grant for purposes

of argument that this is a reasonably close approach to

the truth; and let us also assume that herbs have evolved

at a rate averaging twice as great as that of woody plants,

surely a moderate supposition. All the 2,600 exclusively

herbaceous genera presumably developed as herbs; and

there are 4,200 genera which contain woody plants and

which probably originated as woody plants. Now if these

2,600 herbaceous genera have been developed since the

beginning of the Tertiary, woody plants, producing new

forms at half this rate, would during the same time have

given rise,let us say, to 1,300 genera. To produce the other

2,900 woody genera (to say nothing of the hundreds which

have become extinct) would require, assuming the same

rate of evolution, another period twice as long as the Ter-

tiary, thus thrusting back the origin of the Angiosperms

to a date distant from the present thrice the length of the

Tertiary. According to the estimate mentioned above,

this would be about the beginning of the Jurassic. Of

course this is all extremely hypothetical, but it serves to

emphasize the probability that in order to have developed

their great number and diversity of slowly changing

476 THE AMERICAN NATURALIST [Vou L

woody forms, the Angiosperms must have been in the

process of evolution for a period many times as long as

that since the origin of herbs, evidently beginning at a

date far earlier than that at which the first angiospermous

fossils occur.

This conclusion is still further strengthened by other

facts. The earliest fossil Angiosperms seem to have been

trees, presumably the most slowly alterable type of all.

Furthermore, whatever may have been the importance in

recent times of hybridization through cross-fertilization

by insects as a cause of accelerated evolution, this factor

was evidently inoperative at the origin of the Angio-

sperms, since, according to Handlirsch, flower-loving in-

sects did not make their appearance till the Tertiary.

The specialized character and apparently high phyloge-

netic position of many of the earliest fossil Angiosperms

also renders it highly probable that they were the product

of a long evolutionary history.

Evidence from all sources therefore seems to agree that

the origin of these higher seed plants took place at a time

very much earlier than their paleontoldgical record indi-

cates. That they were not preserved as fossils in hori-

zons lower than the Cretaceous is perhaps due to the fact

that the earliest Angiosperms, as Professor Bailey and

the writer have suggested,” appeared under essentially

temperate climatic conditions, which in the Mesozoic were

mainly confined to upland regions where fossilization

would take place much less commonly than in lowlands.

This apparent predilection of these primitive Angio-

sperms for an environment cooler than the tropical, to-

gether with what we have noted as to their high antiquity,

suggest that the refrigeration of climate which took place

in the Jurassic might have been a factor in their origin;

and even tempts one to look farther back toward the

epoch of markedly low temperatures subsequent to the

close of the Paleozoic, which caused so many radical

7 Sinnott, E. W., and Bailey, I. W., ‘‘Foliar Evidence as to the

and Early Climatic Environment of the Angiosperms,’’ Amer. Jour. Bot., ee. |

IT, 1915, pp. 1-22.

No. 596] RAPIDITY OF EVOLUTION 477

changes in the organic world, as perhaps the time when

the angiospermous stock began to be differentiated from

its gymnospermous ancestors.

We may point out, in conclusion, that a recognition of

differences in the rate of evolution between various

growth forms is evidently of importance in many prob-

lems concerned with the phylogeny, ecology or distribu-

tion of the higher plants. Upon the question, for ex-

ample, as to whether the boreal species in the floras of

southern South America and Australasia are of ancient or

recent arrival there, some light is thrown by the fact that

these species are almost all herbs. That so many mem-

bers of a growth-form which is usually subject to rapid

change still maintain specific identity with distant north-

ern types argues strongly for their comparatively recent

arrival in the south. A similar question is raised by

Willis’s studies of the flora of Ceylon.’ He regards the

endemic element here as one which is of local and recent

origin and as much younger than the non-endemic ele-

ment; but the fact that of the endemic species, by impli-

cation the ones which are changing the fastest, less than

one fourth are herbs, whereas nearly one half of the non-

endemic species belong to this growth-form, would sug-

gest the opposite conclusion; for had this large body of

herbs been in existence here for a very long period, as

Willis supposes, we should have expected it to develop a

relatively much greater body of endemic species. In such

questions as this it should be borne in mind that evolution

has not been a steady and uniform process among all

forms, but that the decided differences between the vari-

eus growth types in the rapidity with which they tend to

~ accumulate heritable variations introduces a factor which

it is always necessary to consider.

SUMMARY

1. The most recently evolved element in the floras of

temperate North America and of Europe, as determined

8 Willis, J. C., ‘“‘The Endemic Flora of Ceylon, ge reference to Geo-

graphical Distribution and Evolution in General,”’ . Trans., B, CCVI,

1915, p. 307.

478 THE AMERICAN NATURALIST [Von L

by a study of the indigenous endemic genera, is composed

almost entirely of plants which are herbaceous in habit.

2, Herbs tend to be grouped in fewer and larger genera

and families than woody plants.

3. Itis therefore concluded that herbaceous plants, pre-

sumably because of the brevity of their life cycle and the

rapid multiplication of generations consequent thereto,

are in most cases undergoing evolutionary development

much more rapidly than are trees and shrubs.

4. From this conclusion are drawn inferences as to the

origin of the herbaceous habit and the antiquity of the

Angiosperms. '

EGG PRODUCTION AND SELECTION

DR. H. D. GOODALE

MASSACHUSETTS AGRICULTURAL EXPERIMENT STATION,

AMHERST, Mass.

Axsout three years ago breeding for increased egg pro-

duction was begun at this station with Rhode Island Reds.

A report on some of the chief features of the work will be

published in the near future, but because of the recent dis-

cussion in this journal, by Pearl and Castle, of egg pro-

duction in relation to selection certain features of egg

production in this breed are of particular interest at this

time.

The winter record of a hen depends upon two main in-

ternal factors aside from possible environmental factors.

First, the date at which the first egg of a pullet is pro-

duced, which may be taken as an index of the attainment

of sexual maturity; and second, the rate of production

after the onset of egg-laying. The former in turn depends

upon the time when she was hatched and upon her rate of

growth. The latter factor, i. e., rate, is controlled in part

at least by an internal mechanism. In the Reds it is clear

that of the two factors the preponderant effect is exer-

cised by the factor of maturity. The observed differences

in rate of production are of less, though by no means neg-

ligible importance. In the Barred Rocks, as far as I have

been able to learn, conditions are reversed, in that the

rate of production appears to exercise a greater influence

on the kind of record a hen makes than the age at which

she produces her first egg, although the latter is also a

factor. Identical records as expressed in number of eggs

per unit of time may therefore result, but it is clear that they

are not directly comparable. The use of the age at first

ege as a criterion of sexual maturity involves some diffi-

culties, but it is the best objective criterion readily avail-

able and corresponds at least roughly with the general

- es

480 THE AMERICAN NATURALIST [ Vou. L

bodily conditions familiar to poultrymen as indications of

the attainment of maturity. These conditions may be ob-

served in the male as well as in the female. There is also

an apparent tendency on the average for large pullets to

make lower records than their smaller sisters, as shown

by the negative correlation between high winter egg pro-

duction and large size, due probably to a longer growth

period in the larger pullets.

The relation between age at first egg and rate involves

the following points: A pullet that matures late can not

give expression to high fecundity, although she may carry

the genes for it, but the record of a pullet that matures

fairly early will depend upon the rate at which she lays.

On the other hand, a zero producer will be also late ma-

turing, judged according to the criterion suggested, be-

cause she does not lay until spring, provided, of course,

that she was hatched at the proper season. It is obvious

of course that a high record bird must mature early in

order to make her record, and that in this sense early

maturity and high productiveness will go hand in hand,

but early maturity of itself does not insure a high record,

for early maturing birds may lay poorly and so make a

low record. On the other hand, late maturity is bound to

insure a relatively low record. In the flock of Reds with

which we have had to do there is comparatively little dif-

ference in the rate of production. In one flock the coeff-

cient of correlation for the time elapsing between the first

egg and the 1st of March and the number of eggs laid was

calculated and found to have a value of .8612 + .0132

which must mean that the flock was fairly homogeneous in

respect to rate of production. In other flocks there is

clearly a negative correlation between age at first egg and

winter egg production but the values have not yet been

determined. The difference between a pullet that begins

to lay December 1 and lays 60 eggs, another the first of

January and lays 40 eggs, another that begins the first of

February and lays 20 eggs, must be a question of start.

The last bird, too, must be essentially different from a

No. 596] EGG PRODUCTION AND SELECTION 481

bird that begins to lay in November and lays 2 eggs that

month, 10 in December, 0 in January and 8 in February,

although she lays the same number of eggs as the individ-

ual that began February 1. The former is clearly a

mediocre producer, the latter a high producer of late

maturity.

The flock averages for the past winter, arranged ac-

cording to the month in which the birds were hatched, are

instructive on this point. The mean egg production for

the (140) March-hatched birds was 39.8 eggs, for the

(172) April-hatched 29.81, and for the (158) May-hatched

18.1 eggs. If the pullets be grouped according to Pearl’s

classification of zero, 1 to 30 (mediocre), and over 30

(high), there were among March-hatched pullets 6.4 per

cent. zeros, 30.0 per cent. mediocre, and 63.6 per cent.

high; among the April pullets 11.6 zeros, 37.3 per cent.

mediocre and 51.1 per cent. high; while among the May-

hatched pullets there were 19.9 per cent. zeros, 58.8 per

cent. mediocre and only 21.3 per cent. high. The average

winter production for the mediocre and high producers

respectively, is for March-hatched birds 17.2 and 54.6; for

April 16.6 and 46.2; for May 16.1 and 40.7. In this con-

nection it should be recalled that there is a tendency for

the abstract numbers involved to yield similar averages,

which are an indication to some extent at least of two

genotypes rather than one. Clearly, the time of year dur-

ing which a Rhode Island Red pullet is hatched plays an

important part in determining the class in which a given

individual falls.

In general the use of a hen’s record expressed in num-

ber of eggs per unit of time as the sole criterion of her

capacity for egg production seems to us to be essentially

wrong. Attention should also be directed to the elements

that enter into the make-up of the record. The factor of

maturity is not wholly a fecundity factor, though it is

closely bound up with egg production and may modify a

pullet ’s fecundity record very considerably. Broodiness

is another factor that, ortn closely associated with

482 THE AMERICAN NATURALIST [Vou. L

egg production, is after all rather of the nature of a sep-

arate entity which may or may not be present, but which

when present operates to reduce the egg record of a hen

very materially. On the average, a hen’s record for the

broody months is only about 60 per cent. of the same hen’s

record for the non-broody months. There is always a

sharp break in monthly egg production with the onset of

the first broody period and as a hen once broody continues

as a rule to alternate brief periods of production with

broody periods, it follows that the number of eggs laid

depends in a large measure upon the number of broody

periods.

Our viewpoint in regard to modifying factors for egg

production may be observed in morphological as well as

physiological characters. Thus, the color, arrangement

and length of hair could not be studied in a hairless race,

although the genes for color, arrangement and length may

actually exist in the germ plasm of the hairless individ-

uals. Variability in part at least is due to the existence of

numerous germinal factors which can come to somatic

expression only when some other germinal factor is pres-

ent. Thus, there are several types of rose comb which

appear to be due to definite factors that affect only the

rose comb. The spike, for example, may be telescoped

into the body of the comb. The telescoped condition is

clearly inherited but it does not appear in singles of the

same family, although there is no obvious reason why the

blade should not be telescoped quite as readily as the —

spike. Before the rose comb can be said to be understood,

the various genetic modifying factors involved, their re-

lations to each other, and the mode of inheritance of each

must be made out. Any study of selection that fails to

take into account such internal modifying factors is in-

complete.

There is no question regarding the inheritance of egg

production in the sense that certain families make much

better records than some other families, though in the

Rhode Island Reds the families that make the better

No. 596] EGG PRODUCTION AND SELECTION 483

records also mature at an earlier average age than those

making poor records. Whether or not the mode of in-

heritance of fecundity in Rhode Island Reds follows

Pearl’s theory is still uncertain. Since there is great

variability in the age at which the first egg is produced

and since its frequency polygon indicates, at least up to

305 days, the essential homogeneity of the flock, we are

inclined to believe that the Reds are unfavorable material

for a solution of this question. More work, however, re-

mains to be done before this point can be cleared up.

On Pearl’s theory the continued selection of genetic

high producers and the continued use of their sons as

breeders should in the long run tend toward an improved

egg production, even though individual pedigrees are not

kept, as he himself has pointed out.

The rate at which the selection becomes effective de-

pends much on what classes are assumed to be available

at the start. In the following discussion it is assumed

that the matings are made on a scale sufficiently large to

permit the various matings to take place in the propor-

tions indicated. The scale required, however, is so large

that it would be scarcely practicable to test the matter

experimentally.

In the first place, it should be pointed out that of the

nine genetic classes of males possible on Pearl’s theory,

classes 5, 6 and 9 are not produced by any mating of class

1 and 2 females—the only classes laying 30 or more eggs

—with any class of males. Moreover, it can be shown

that if these three classes be assumed to be the only males

available and that if only class 1 females (l, Ly, L.1,) be

bred throughout the test and if the male offspring be bred

in each generation in the same proportion in which they

were thrown by the preceding generation, at about the

seventh generation there would be male classes 1, 3 and

7 only, existing in the proportion of 1:2:1. Classes 2, 4

and 8 would indeed be present, but the three classes to-

gether would amount to only a little over 1 per cent. of

the population. This small percentage decreases one half

484 THE AMERICAN NATURALIST [Vou. L

in each succeeding generation. The females produced by

this group of males mated to class 1 females only, would

occur in the proportion of ‘‘over 30” females 75 per cent.,

‘‘under’’ 25 per cent. The ratio 1:2:1 among the males

maintains itself after once reached, so long as class 1 fe-

males only are used for mates.

The same ratio is reached if it be assumed that all nine

classes of males exist in equal numbers at the start and

that they are bred in each generation in the proportions

thrown by the preceding generation and only to class 1

females.

The assumptions made have been chosen as those most

likely to lead to decreased egg production or to a main-

tenance of production on a level. I have not worked out

many of the possible assumptions, but as long as only fe-

males belonging to classes 1 or 2 are used in the matings,

males of higher classes, i. e., classes 1, 2, etc., are thrown

by those of the lower classes in the same manner as

- pointed out for classes 5, 6 and 9, so that eventually males

of the higher classes come to exist in definite ratios.

If class 2 females (L, L, L,1,) only are bred, somewhat

different results are secured. If males belonging to classes

1, 2, 3, 4, 7 and 8 are used in equal proportions at the

start (classes 5, 6, 9 being omitted for reasons stated

above) then in the seventh generation nearly 98 per cent.

of the males will belong to class 1, the other 2 per cent.

belong to classes 2 and 3, and these will diminish one half

in each succeeding generation. The final result, then,

should be a race of high producers.

Since it has been shown that when class 1 females are

used exclusively in the matings, that eventually a stable

flock of birds is produced in which male classes 1, 3 and 7

exist in a definite ratio, and since if class 2 females are

bred exclusively a flock consisting solely of high pro-

ducers will result in time, and since class 1 females mated —

to males of classes 1, 3 and 7 throw 25 per cent. class 2 __

females, 50 per cent. class 1 females and 25 per cent. —

mediocre producers, it would seem inevitable that eventu-

No. 596] EGG PRODUCTION AND SELECTION 485

ally a homozygous race of high producers should result

from the continued breeding of the sons of high producers

to high producers.

Since Pearl has repeatedly stated that his results apply

to annual production as well as winter production, it

might be supposed that selection during the Gowell

régime should have been effective. Pearl has interpreted

the results as due to Gowell’s failure to use the progeny

test. Two other possibilities, however, suggest them-

selves. Under the Gowell régime the males came from

200-egg hens, but all hens laying over 150 eggs were used

as breeders. Possibly the latter value is too low. The an-

nual production corresponding to the 30-egg division-

point for winter production has not been stated, as far as

I know, but since the mean annual value of a flock of high

producers is stated by Pearl to be about 166 eggs, 150

eggs “would seem sufficiently high to exclude mediocre

producers. The possibility, then, that this value is too

low may be disregarded.

The second possibility relates to the number of males

-used. In an actual experiment only a relatively few

males are used each year. The scale at which the Gowell

experiments were carried on was hardly sufficient to equal

more than one or two years’ work on the scale that would

be required to get the males in something like the propor-

tions expected, and hence little or no change in egg produc-

tion could be expected in the period during which the mass

selection experiments were carried on.

While, theoretically, mass selection should be effective

under the conditions stated, it would require large num-

bers in order to be uniformly successful. The progeny

test, on the other hand, produces results quickly and defi-

nitely,

SHORTER ARTICLES AND DISCUSSION

ON SELECTIVE PARTIAL STERILITY AS AN EXPLA-

NATION OF THE BEHAVIOR OF THE DOUBLE-

THROWING STOCK AND THE PETUNIA

My attention has only just been drawn to the paper by Howard

B. Frost: which appeared in the issue of the AMERICAN NAT-

URALIST for October, 1915, under the title of ‘‘The Inheritance of a

Doubleness in Matthiola and Petunia.” In this paper, which is :

a preliminary communication, the writer states that from a con-

sideration of the data contained in the accounts which I have

published of my experiments on the cross-breeding of pure single

and double-throwing strains of stocks (Matthiola), he has been

led to form a view differing from that which I have put forward

as to the interpretation to be placed upon these results. Ac-

cepting the essential points requiring explanation to be as I have

stated them, he discusses the scheme which I have suggested as —

underlying and accounting for the facts observed, and also the

explanation of these facts proposed by Goldschmidt.? Though

conceding that the factorial scheme which I have formulated?

would give the results observed, he rejects it on the ground that

it is unnecessarily complex, and claims that his own interpreta-

tion, which he then gives, is both simpler and supported by defi-

nite evidence. Though it is evident that a final decision on the

points raised must await further investigation, we can in, the

meantime examine in the light of our present knowledge the two

main grounds upon which Frost claims that his explanation is |

to be preferred, viz., (1) its greater simplicity and (2) the ex-

istence of definite evidence in its favor. ce

For the purposes of this comparison we need take into account — -

only the three following outstanding facts: a

1.. That whereas some single stocks yield only singles in each

successive generation other strains yield a mixture of singles and

1 As Mr. Frost mentions in a note to his paper that he has received no

answer to a letter addressed to me in May, 1914, I may take this oppor-

tunity to say that no letter or paper from him has ever reached me, adt

can only suppose that his letter in some way unfortunately miscarried.

2 Zeitschrift f. induktive Abstammungs- u. Vererbungs-lehre, Bd. 10, p- 74,

1913. a

3 J. of Genetics, Vol. 1, No. 4, 1911, and later J. Roy. Hort. Soc., Vol. XL,

Part III, 1915. E

486

No.596] SHORTER ARTICLES AND DISCUSSION 487

doubles in each generation. So far as we know the pure-breeding

type and the ever-sporting (double-throwing) type differ in no

other respect than this particular character. One can obtain

parallel strains of the same color, the same habit, the same form

of surface character, the one pure-breeding, the other double-

throwing.

2. That double-throwing singles give a small but constant ex-

cess of doubles, which we have no reason to doubt is due to a

constant excess production among the functional gametes of those

carrying the double character.* Were this excess merely acci-

dental we should expect to obtain deviations from equality in the

direction of deficiency as well as of excess, but such deviations

have not been found to occur. The excess of doubles sug-

gests a ratio of 9D:7S (564 per cent. of doubles) or possibly

8-5D:7-58 (534 per cent. of doubles)

3. That the results of eross-breeding are such as to show that

in the pure-breeding single all the functional egg cells and male

germs carry singleness, whereas in the double-throwing strain all

the male germs and rather more than half the egg cells carry

doubleness, somewhat less than half the egg cells and none of the

male germs carry singleness. For if the cross is made in the form

no-d single 2? X d-single ¢ all the F, singles give a mixture of

singles and doubles in F,, whereas in the reciprocal cross only a

proportion—generally rather more than half—of the F, cross-

breds behave in this way, the remainder give only singles in F,.

To explain the above facts I have suggested:

1. That singleness is due to the presence of two factors (X and

2. That in the present-day true-breeding single these two factors

4 That is to say this excess is not apparently due to any selective mor-

tality after fertilization or germination. In previous accounts I have not

introduced into my statements the qualifying expression — 7? but

I do not gather that any misapprehension has arisen in consequence, since I

have clearly shown that my scheme was based on the results ai from

fertilization—results, that is to say, arising from the nature of those germ

cells which did actually function. Until there was any evidence fortheoming

of the formation of non-functional germ cells it seemed more misleading to

introduce the word than to leave it out. As, however, Frost bases his view

entirely on the supposition that: some of the germ cells are constitutionally

incapable of functioning it becomes advisable for the sake of clearness to

express this limitation, previously inferred but not stated, without, however,

oo thereby the atipa of the _prodastion of non-functional

gametes,

488 THE AMERICAN NATURALIST [Vor L

occur linked together (XY), i. e., they exhibit the kind of

relation which is now generally assumed as the interpreta-

tion of a widespread class of phenomena.

3. That in the ever-sporting single this coupling or linkage is

only partial so that X and Y can occur dissociated.

4. That in this type of single these two factors are distributed

differently to the functional male and female germ cells.

The male germ cells are unable to carry either factor. The

female germ cells may contain both, or one or other, or

neither, the four different combinations occurring in the

ratio 7:1:1:7 or possibly 15:1:1:15.

These suppositions provide us with a working hypothesis which

covers the facts detailed above and enables us to form a concep-

tion of how it comes about that the two classes of singles behave

as they do.

If 1 and 2 are true, and they postulate nothing beyond what is

held to be the probable explanation in the case of other charac-

ters studied by other observers, we must suppose that 3 is also.

Since were this not so, the offspring of an ever-sporting single ae

having X and Y linked together in any of its functional germ c

cells, would some of them behave like F, cross-breds from the

cross no-d single 9 X d-single § and would yield only a small

proportion of doubles, whereas observation has shown that all

the offspring of ever-sporting singles give an excess of doubles

like the ever-sporting parents. Finally as to 4, the supposition

of a sex-limited distribution of factors finds also a parallel in

other cases: the special features in the present instance are that

a sex-limited distribution of factors should occur in an hermaph-

rodite organism, and that on the male side it should be complete.

So much for the main points .of my scheme which are ac-

knowledged by Frost and clearly set forth by him. I now turn

to his criticisms.

1. His first point is that I give no reason why singleness rather

than doubleness is eliminated on the male side, or why this elimi-

. nation is uniform. This is true. But do we know the reason

‘“‘why’’ in any case where factors are lost or where they show &-

sex-limited inheritance? Take, for example, such a case as the —

red-flowered stock in illustration of the first point. We know

that the original wild form was purple-flowered, and we suppose —

that at some point a mutant arose from which a certain factor B

capable of turning a red color blue, much as an alkali turns rd

No. 596] SHORTER ARTICLES AND DISCUSSION 489

litmus blue, which was present in the wild form, was eliminated.

But we do not know why the factor B happened to be eliminated

rather than either of the other two factors C and R which we

suppose to produce red-colored sap, and to be present in the wild

form as well as the factor B. We none the less regard the con-

ception of the existence of the factors B, C and R as satisfactorily

expressing the color relations in the stock so far as we know them.

Neither are we able to say why at some point in the course of

evolution a mutant arose, which, although inheriting the factors

(or factor) governing singleness, produced some functional germs

in which these factors were absent. With regard to his further

point, viz., that I offer no explanation of the fact that this elimi-

nation is uniform (by which I understand him to mean com-

plete) on the male side in the double-throwing single, seeing that

it is partial only on the female side, I may point out that neither

does he. Indeed it appears to me to be more difficult to explain

the facts regarding the egg cells on his view (see below) than to

reconcile the production of a uniform type of male germ with |

mine. So far indeed as the female germs are concerned he does

not attempt to formulate any definite scheme. As it is upon their

behavior that the observed excess of double offspring depends,

one is constrained to ask why the criticism which he passes on

Goldschmidt’s scheme, viz., that it ‘‘gives at most only an in-

definite implied explanation of the deviation of the double-single

ratio from equality in the double-throwing races,’’ does not

equally apply to his own hypothesis. Stated briefly and apart

from certain alternative suggestions put forward tentatively in

the absence of any positive evidence, Frost’s theory appears to

amount to this: that single-carrying as well as double-carrying

pollen is formed by the ever-sporting single, but that all the

single-carrying grains contain some other factor which in some

way prevents them from functioning. He finds it further neces-

sary to suppose that a similar lethal process occurs in some of the

ovules, but that for some (unexplained) reason some only of the

single-carrying egg cells are rendered functionless, the others are

always capable of fertilization.» With regard to his question

why it is singleness which is eliminated on the male side and

not doubleness, it seems unnecessary to point out that as the

double-throwing single has arisen from a pure-breeding single,

5I conclude that Frost would accept this as a general statement of his

position, but he makes no — statement regarding the ovules, ”

490 THE AMERICAN NATURALIST [ Vou. L

if the mutation is one of loss the character lost must be singleness.

But a more important line of argument is that dealing with the

question as to when the elimination of singleness takes place.

At what stage in the sequence of cell divisions occurring in the

direct line between the fertilization of the egg cell at the begin-

ning of the cycle and the formation of the male sperm by the

factors, and is the disappearance of the single character from the

male germs illusory? In other words, is it the case that both

single- and double-carrying pollen grains are formed, but that

those carrying singleness always fail to achieve their purpose—

fertilization? This brings us to a consideration of the question

whether, if elimination occurs, it takes place at some point in the

series of cell-divisions which culminate in the formation of the

pollen mother cells, or at the reduction division, or at one of the

succeeding divisions which give rise successively to the complete

sporetetrad, the vegetative and generative cells of the pollen

grain, and the twin sperms formed from the generative cell. The

first alternative Frost dismisses on the ground that we have no

decisive evidence in support of somatic segregation and ‘‘an over- on

whelming convergence of probabilities against it.” That we y

have as yet no actual evidence of such segregation in stocks

themselves is true but in the light of known facts in regard to i.

bud variation it seems difficult to escape from the view that so- 2 k

matic segregation not only might, but does sometimes occur, and

that this possibility is not excluded in the present case.

It appears to be rather generally assumed, though perhaps

without sufficient reason, that the segregation of all allelo-

morphs must necessarily occur simultaneously. Those who hold

that the evidence justifies the belief that segregation takes place

at the reduction division are naturally committed to the view of

simultaneous separation of all allelomorphs. But it must be aĉ-

knowledged for the reasons given above that there is a certain

difficulty in the way of accepting the view that the reduction — |

division of the germ cells constitutes the sole sorting mechanism —

for the allelomorphs. May it not be, even if the majority of the

allelomorphs segregate as a rule at some particular point, that

occasionally or even regularly, segregation of one or another pair a

of allelomorphs may occur prematurely or again may be post-

poned? All that we certainly know is that by the time the

gametes are formed the sorting has been completed. As regards

the possibility of elimination at a later stage than the formation

No.596] SHORTER ARTICLES AND DISCUSSION 491

of the pollen mother cells Frost states that, although he -has

looked for indications suggestive of degeneration among the

spor tetrads, he has seen no signs of it at any stage. I may say

that the appearance of the pollen in ever-sporting as in true-

breeding singles is that it is all consistently good, whereas, any

supposition attributing the behavior of the ever-sporting single

to degeneration in the grains must needs assume that about half

of them are consistently bad. In the absence of any histological

evidence in support of elimination of factors through degenera- `

tion of the pollen grains themselves Frost does not advance it.

Adopting the view (as I gather) that segregation must occur at

the reduction division, he does not either consider the possibility

of elimination at a later point, though the removal from the cell

genealogical tree of one daughter cell in each of two successive

divisions of the pollen grain leading to the formation of the twin

sperms affords precisely the kind of evidence of which he is in

search. There remains the alternative view which Frost puts

forward, viz., that both single-carrying and double-earrying pol-

len is produced, but the single-carrying grains never achieve

fertilization. In order to get over the difficulty that the male

germs of the true-breeding single are undoubtedly single-carry-

ing but are yet able to effect fertilization Frost is compelled to

make one of the following alternative assumptions—and here we

come to the main point in his scheme:

1. That the pure-breeding single being regarded as SS i in con-

stitution, the double-throwing single must not only be supposed

to have lost one S factor, but the remaining original S factor

must have become altered, or a new lethal factor completely

linked with it must have been produced in the double-throwing

mutant ‘‘by which the presence of S becomes incompatible with

pollen formation.’’ In other words, the two processes which

Frost postulates as leading to the appearance of the double- —

throwing single, require (1) the loss of one § factor, and (2)

either the production of a new factor producing singleness which

we may indicate by S, or the complete coupling of the original

remaining S factor with some new factor. That is to say, he

also finds it necessary on either view to suppose the existence —

of at least two factors, S and S, and in the second alternative

to assume further that the two are completely linked. It

seems difficult to see wherein this scheme shows greater sim- . . a :

plicity than the one I suggest, which is Joes ee: is ar pees

492 THE AMERICAN NATURALIST [Vou. L

to two factors which are completely linked in the true-breed-

ing single type and dissociated in the double-throwing single.

2. As an alternative hypothesis to the above Frost suggests

that the pure-breeding single from which the double-throwing

single originated may not have been the same kind of single as

those which one meets with to-day, but that it differed from them

in one or other of the two ways suggested under 1. But again is

this a simpler explanation?

Simplicity, however, though not to be disregarded, is not neces-

sarily the final test. Let us consider Frost’s second argument

that his view can be supported by definite evidence. Failing to

find any positive histological evidence that can be taken as indi-

cating the required process of factor elimination, Frost suggests

that the case of the ever-sporting single is to be considered as

that of a hybrid showing selective sterility. To quote his own

words:

Selective partial sterility seems to be rather a common pas

and it very probably occurs here.

Functional sterility is to be supposed in the case of all pollen

grains carrying the factor (or factors) essential to singleness,

i. e., according to Frost’s scheme, either S,—a modified form of

S, or S with a lethal factor linked to it. In other words, half

(we are to presume) of the pollen, though apparently good, is

supposed to be incapable of fertilizing the ovules. He further

adds that if we also assume ‘‘a slight tendency to selective elimi-

nation of S-carrying eggs’’—a somewhat vague supposition to

account for a very definite fact—or if these egg cells were less

often fertilized than those which are s in constitution, or if there

were selective elimination of the embryos (Ss) that would pro-

duce singles, we should have ‘‘a simple and direct’’ explana-

tion of the constant excess of doubles. It is in fact suggested

that the postulated inability of the single-carrying pollen to

fertilize the ovules may be due to want of vigor. As bearing

on this point he calls attention to the greater vigor of the

double as compared with the ever-sporting single, as shown

in the vegetative habit of the growing plant, and in the greater

viability of the double-carrying embryos (seed). This greater

vegetative vigor of the ss over the Ss zygote Frost contends

may possibly be the outcome of a similar difference between

the s and S gametes. But if this were so, if the lesser vigor

of the ever-sporting single is due to the presence of one S fac-

No.596] SHORTER ARTICLES AND DISCUSSION 493

tor, surely we should expect that the true-breeding single, which .

is SS in constitution, would be less vigorous still. It seems

somewhat gratuitous to suppose that the character singleness is

sometimes due to a factor S associated with greater vigor and

sometimes to a factor S, associated with diminished vigor; or, to

put it another way, to assume that if it is the lack of one dose of

the factor for singleness in the ever-sporting plant which makes

it less vigorous, that the lack of the double dose in the double

plant leads to the opposite result of greater vigor. Is it not

almost a certainty that the greater vigor of the double-flowering

plant is due to the fact that the energy of the individual is not

exhausted in the formation of the reproductive cells, but is ex-

pended in producing a more vigorous vegetative growth? And

hence that a check to vegetative growth, similar in cause and in

degree, is operative alike in the pure-breeding and the double-

throwing single? To obtain strict proof that this is so is diffi-

cult since it might always be argued that the particular pure-

breeding single strain used as a control was not precisely identical

in all other respects with the double-throwing strain with which

it was being compared. It can, however, be stated that in some

commercial material supplied as double-throwing, but which

proved to be a mixture of pure-breeding and ever-sporting

singles otherwise apparently identical, no indication was observed

of any difference in vigor between the two kinds of singles. The

second argument which Frost urges in support of his hypothesis

of differential sterility is the fact that the seeds (embryos) which

produce doubles have on the whole rather greater viability than

their sister seeds which give singles. But this second argument,

depending for point on the same entirely unsupported assump-

tion as that derived from vegetative habit, is open to the same

objection. I have never observed that the seed-producing true-

breeding singles showed any superiority as regards viability over

that yielding the ever-sporting singles, but rather, as with vege-

tative habit, that the distinction is to be drawn between any

kind of single-producing seed on the one hand, and double-pro-

ducing seed on the other. I have sown a large quantity of ap-

parently excellent seed of a pure-breeding single after the lapse

of some years and failed to get any germination, just as I have

failed sometimes to get any singles from old double-throwing

seed although a few doubles were obtained. Lastly, Frost brings

forward the somewhat out-of-date view that gs — of

494 THE AMERICAN NATURALIST [ Vou. L

doubles is increased by starvation treatment. Though this treat-

ment is still practised by German growers, it survives from a time

when it was not yet appreciated that the capacity of the individ- |

ual to become a single or a double depended upon its inherited

constitution and not upon the effects of environment.® It may be

pointed out that the French growers have been in the habit of

pursuing precisely the opposite method of treatment with the

same object in view. The inhibition of flowering which Frost

has observed to be more marked in the singles than in the doubles

in the case of one variety when subjected to a high temperature

does not seem to me to bear on the question at issue,’ which is

whether there is any direct evidence of the selective sterility of

ovules and pollen in the ever-sporting as compared with the true-

breeding single. For I gather that Frost is not prepared to

maintain that sterility occurs regularly also in the pure-

breeding single as well as in the ever-sporting single. It is

further to be noted that besides assuming the definite sterility

of all the S, pollen in an individual of S,s constitution pre-

sumably producing equal numbers of S, and s grains, Frost

has to have recourse to the vague assumption that there 1s

only a slight tendency to selective elimination of the S, egg

cells out of a total composed presumably of equal numbers of

S, and s. One is fain to ask on what grounds it is possible to

uphold the view that the same factor can destroy the func-

tional activity of every pollen grain carrying it, but is only

able to affect some of the egg cells in which it is borne?

this connection it may be mentioned that it is no uncommon cir-

cumstance even when self-fertilization is left to nature to obtain

pods both in pure-breeding and double-throwing singles where

every ovule has been fertilized, and this can always practically

be ensured where fertilization by hand is carried out in gooi

weather. Though he instances no examples I gather from his

previous reference to Belling’s work: that he has in view such —

eases as that of Stizolobium decringianum (the Florida Velvet

‘‘bean’’) and other species investigated by that observer, Nico-

tiana on which recent experiments in this connection have been

6 The real advantage of the German method of treatment is that the seed —

harvested is all well ripened.

7 On this cei of the inhibition of flowering, however, Frost promises

further informat

8 Zeitschr. f. ma hes: u. Vererbungslehre, Band XII, 1914.

No.596] SHORTER ARTICLES AND DISCUSSION 495

made by East,’ and @nothera studied by Geerts.’? But in these

cases we are dealing with species hybrids, with plants in which

partial sterility is a demonstrable fact and due to a demonstrable

cause. In the case of Nicotiana the parent plants were both

(one certainly, the other in all probability) self-sterile and thus

the cause of self-sterility in the cross-bred offspring is explained.

In Stizolobium Belling found, as Geerts had previously noticed in

(nothera, that in the hybrids a certain proportion of both ovules

and pollen grains were shriveled and malformed. Here again

the cause of the sterility is plain. If, as Frost suggests, the ever-

sporting single stock should be regarded as a hybrid, it differs

completely from the cases referred to above, for in this case there

is no obvious sign or cause of sterility either in the plant itself or

in the pure-breeding single from which we suppose it to have

arisen. It becomes then a question whether to attach weight to

the argument from analogy in a case where it goes against all the

evidence available.

rom the considerations here reviewed I am led to sum up the

position as follows:

Evidence is wholly lacking in the stock itself in support of

Frost’s hypothesis that the behavior of the ever-sporting —

(double-throwing) single is to be accounted for as the result

of the selective sterility of ovules and pollen. Not only so, but

it may be claimed that the facts on which he relies to support

his argument can equally well be adduced in favor of the opposite

point of view. The selective elimination of embryos or the more

frequent fertilization of egg cells of s as compared with those of S

(or 8,) constitution seem both untenable in view of the fact that

the usual excess of doubles is obtained in cases where every ovule

is fertilized and every resulting seed germinates. In the formal

scheme which I have put forward we have a working hypothesis

which enables us to correlate the present known facts. Does

Frost’s hypothesis give us more than this? Are we not, in the

end, still left debating whether his various speculations unsup-

ported by facts really carry us further, and whether they can

justly lay claim to the merit of greater simplicity?

There is, however, one further point of interest bearing on this

discussion to which I would call attention here. According to

the scheme which I have suggested the constitution and gameto-

Senesis of the two kinds of single can be represented thus:

° THE AMERICAN Narurauist, Vol. XLIX, No. 578, 1915.

19 Récueil des Trav. Bot. N eerl., Vol. 5, 1909.

496 THE AMERICAN NATURALIST [Vou. L

Ever-sporting Single Pure-breeding Single (as it is Found

XYxy in Ordinary Commercial Material)

KYAT

Gametes Gametes

Ovules Pollen Ovules Pollen

7 XY or15 Xy all xy all XY all XY

1 Xy

Sad

T xy or 15 xy

the presence of X and Y in the zygote always producing single-

ness, whether the factors are coupled or not.

Now on this supposition the mating d-single 2? X no-d single J

will give some plants of the constitution XY XY. These when

self-fertilized. will presumably give some pure-breeding F, plants

of XYXY composition. If these pure-breeding singles in which

X and Y are not coupled are crossed back with the pollen of an

ever-sporting single we shall again have a plant with the consti-

tution XYxy. In this way we may hope perhaps to synthesize

the ever-sporting form. This experiment is already in progress.

For this purpose the ever-sporting strain known as sulphur-

white is particularly well suited if used as the female parent in a

mating with a pure-breeding cream. As the ovules of the sul-

phur-white appear to be of the four kinds, XYW, XyW, xYw,

xyw, we know that only F, single whites will serve our purpose.

We also know that we can disregard among the F, families de-

rived from the self-fertilization of F, whites those which contain

doubles, and proceed to cross individuals in the pure-single fam-

ilies with pollen of an ever-sporting type. By the choice of a

sulphur-white as the female parent in the first cross we are sav

much trouble in identifying those F, plants which contain XY ee

unlinked. In this way we may hope to obtain further light on

the different condition of the factors for singleness in the pure-

breeding and ever-sporting single, respectively.

After dealing with the stock Frost suggests that his hy pothesis

of selective sterility no doubt also explains the case of Petunia,

and adds that in any case my view that singleness is here domi-

nant is untenable. He further mentions that both Goldschmidt

and Belling" hold the view that doubleness and not mena is

11 Belling’s statement is that ‘‘in the Petunia doubilons may be incom-

pletely dominant as in the greenhouse carnation.’? THE AMERICAN ar

RALIST, Vol. XLIX, No. 578, p. 126, Note 1, 1915.

No.596] SHORTER ARTICLES AND DISCUSSION 497

dominant in this case. In my account™ of the results obtained

with horticultural strains of Petunia violacca and P. nyctagini-

flora and a hybrid double form I suggested that we might account

for the following facts:

1. That the singles when self-fertilized or inter-bred gave only

singles.

2. That when crossed with the pollen of doubles they gave both

singles and doubles in F,, the singles being in excess though

not always in the same ratio.

by supposing either

(a) That some factor essential to singleness was absent from all

the ovules of the singles and from some of the pollen of

the singles, or conversely

(b) That some factor was absent from all the pollen of the

doubles, but only from some of the ovules of the singles,

the single-double character, and

(d) That singleness is dominant.

It will be seen that the above suppositions provide only a gen-

eral basis of explanation; they do not constitute a full solution

even of the facts observed. Frost, putting (c) on one side on the

ground that it concerns only the deviation of the ratio from 50

per cent., argues that both (a) and (b) are untenable. He re-

gards the facts as indicating that doubleness and not singleness is

dominant, and holds the view that if some of the pollen of the

doubles be assumed to carry doubleness and some singleness the

hypothesis of partial sterility will explain the rest. But though

the formulation of (c) was intended primarily to provide for the

occurrence of more than one ratio, it was essential also to the

suppositions (a) and (b). If (c) is negatived then (a) undoubt-

edly becomes impossible, but if (c) is true then (a) or (b) might

represent a basic part of the explanation with which some further-

complication was combined, as to the nature of which, however,

the data available afforded no clue. Owing to the complete ster-

ility of the double plant, it was impossible to make the reciprocal

cross. The singles employed might be, in fact almost certainly

Were, of mixed descent. It was realized that at best either scheme

offered only a partial solution. Unfortunately the efforts made

m the course of the experiments and since, to obtain seeds of the

wild species, have only been partially successful. The position

12 Journal of Genetics, Vol. 1, No. 1, 1910.

498 THE AMERICAN NATURALIST: [Von L

still is that we are unable to say for certain whether doubles in-

variably occur when the above-named species are crossed

with the pollen of a double. When this evidence is available we

may expect it to throw further light on the question as to which |

character is dominant. At present decisive proof on this point is

lacking. Comparison with the other types carries us no further.

Singleness has been found to be recessive in carnation, hollyhock

and Meconopsis cambrica; on the other hand,—it is dominant in

wallflower and probably in sweet william. Moreover, these

forms differ from Petunia in that they give a uniform F, when

single and double are mated together. The case of Petunia

therefore still remains one of balance of probabilities. In regard

to Frost’s further suggestion that the facts observed are due to

selective sterility I think that this hypothesis may very possibly

be correct and certainly has some evidence in its favor. P.

vidacea is recognized as a self-sterile species and many of the

singles which I used proved to be so. A large number of individ-

uals were tested for this character, but further investigation was

postponed in the hope of obtaining pure material for comparison.

If Frost’s hypothesis is confirmed for Petunia, and the work of

Belling and East*® points in this direction, it may offer a com-

plete explanation of the facts and render the supposition of a

differential distribution of factors to ovules and pollen in this

genus unnecessary.

Epita R. SAUNDERS

NEWNHAM COLLEGE

GAMETOGENESIS IN PLANTS

Tue evolutionary origin of the reproductive cells furnishes

one of the most fundamental problems connected with genetics,

for upon a clear understanding of the subject depends the satis-

factory solution of many subsidiary problems relating to animal

and plant breeding. The value of hybridization and inbreeding;

the meaning of the pure line hypothesis; the principle of cumU- a"

lability, etc., may here be mentioned. Therefore, whether or not :

one agrees with the conclusions presented, studies from widely

18 Saunders (unpublished). The carnation has also been investigated by —

Norton. (See paper read at the meeting of the Society of Horticulture in

Philadelphia in December, 1904. Also Gard. Chron., Jan., 1905.)

14 Loe. cit.

15 Loe. cit.

No.596] SHORTER ARTICLES AND DISCUSSION 499

divergent standpoints which bear upon the question are to be

welcomed. It is only through an analysis of the opinions thus

advanced that there will develop a perspective which will eventu-

ally permit the solution of the problem.

It is in this connection that the conclusions of Professor Coulter

as set forth in ‘‘The Evolution of Sex in Plants’’ are of interest,

representing as they do the views of one whose attainments in

biology have by no means been confined to the field of plant mor-

phology. Presented in a clear and interesting manner so far as the

facts are concerned, the volume furnishes a valuable résumé of

the subject from the botanical standpoint. It is evident, how-

ever, that a certain narrowness must exist in such a presentation,

for a problem of this nature demands that plant biology and ani-

mal biology supplement one another from the experimental as

well as from the morphological and cytological side. Gameto-

genesis had its beginning not, as Coulter suggests, among organ-

isms far above the most primitive plants, but among unicellular

flagellate forms whose representatives partake of the nature of

both plants and animals and from which have arisen the various

groups of plants in general. Sexuality, once having arisen, may

have been partially or even wholly suppressed in various plant

groups, but its subsequent reappearance by no means makes it

necessary to affirm its polyphyletic origin. Our present knowl-

edge of Mendelian behavior is of interest in connection with such

a view.

It will be well to examine some of the more definite conclusions

which Coulter has presented. Few of these are original, never-

theless they are of decided value since they are in most cases sup-

ported by unique observations bearing directly upon the point of

view. It is merely unfortunate that the bibliographic references

which would illustrate the development of the ideas are entirely

absent, in consequence of which a false impression may be con-

veyed to many readers.

Early in the volume it is stated that sex in oe higher animals

has become the only method of reproduction. Logically this

view is not to be maintained, as has already been pointed out by

LeDantee (’03) as well as by Chamberlain (’05) evidently in

ignorance of the conclusions reached by the previous writer.

More recently Janet (712) has considered the subject. If the

criterion by which the sporophyte is to be distinguished from the

— John Merle Coulter, head of the department of poe, Univ. af

leago. Univ, of Chicago Press, December, ss

500 THE AMERICAN NATURALIST [Vou L

gametophyte rests upon the 2x as compared with the x condition

of the chromosomes, we find that among animal organisms the

asexual phase has actually become the dominant method of re-

production and the sexual phase is represented only by the para-

sitic cells arising through the reduction division. In accordance

with this view one is prepared to accept the spore mother cells

of plants as homologues of the cells preceding the reduction di-

vision in animals. In a subsequent discussion Coulter states in

accordance with the view first advocated that the animal body

produces gametes and not spores. When reduction occurs at the

time of the first maturation division in animal organisms it is

quite clear that the cells thus produced may correspond to spores

which in the next division give rise to gametes. It may even be

asserted that they are megaspores or microspores dependent on

the sex represented. When chromosome reduction is moved for-

ward to the second maturation division, however, it is possible to

agree with Coulter, but seemingly more logical to admit that the

change is a secondary one and that the first maturation cells may

still represent the spore cells.

In accordance with the proposition that spores unite as gam-

etes to form a single cell, evidence should either be presented to

show that an identical chromosome composition exists between

the actual spores and the so-called spores functioning as gametes

or consideration should be given to subsequent reduction division.

Otherwise the conclusion scarcely merits the value of an opinion.

Furthermore, the argument that the fusion of a sperm and cell

among the angiosperms to form a nutritive endosperm justifies

the conclusion that pairing and fusing do not represent the es-

sential features of sexuality, can not be considered. This is only

one of numerous examples where changes in form or function of

parts occur without having any bearing on the actual origin of

the part. Even in this case a fusion is represented and may

have a value similar to that among gametes.

It is in connection with ‘‘A Theory of Sex’’ that it seems neces-

sary to decidedly differ from Professor Coulter. Here the two-

main theses are that sexuality has arisen (1) to carry an organ-

ism through an unfavorable environment, and (2) to make evo-

lution more rapid by presenting a greater diversity of forms.

The first deduction is based on the proposition that gametes

in many plants are produced at the close of the vegetative period. — ae

Such a conclusion—post hoc ergo propter hoc—does not rest upon

a sound basis. With the fulfilment of a function having the

No. 596] SHORTER ARTICLES AND DISCUSSION 501

importance of a gamete production, it is quite logical that the

cycle of development should close, but to state that the closing of

the cycle has brought about the production of the gametes, is

quite another thing. The acceptance of this would lead to the

inference that gametes arose in fresh or brackish water forms

where pronounced seasonal changes took place and not in larger

bodies of water like the ocean, the most probable place.

The second deduction is a restatement of a conclusion reached

by Weismann (’76) to the effect that amphimixis increases vari-

ability with the assumption that variations thus assumed to be

produced are inherited in a cumulative manner. The evidence,

however, available at present, supports a view directly contrary

to this, namely that the gametie condition makes evolution slower

by decreasing the diversity of available forms. Mendelian com-

binations may occur but the result is a de-

crease in variability when the parental populations are compared

with the F, or with a succeeding generation. The amphimuta-

tions are transitory and there is no evidence that they present

anything actually new in themselves.

Regardless of the opinions here at variance some of which can

only be established as sound generalizations through long experi-

mental investigation, the summary of gametogenesis by Pro-

fessor Coulter will be read with profit and pleasure by those in-

terested in problems of evolution as well as by those particularly

concerned with plant morphology and development.

L. B

WALTON

KENYON COLLEGE,

_ GAMBIER, O.

NOTES AND LITERATURE

LIFE HISTORIES IN THE RED ALGZ.

THE past decade has given us a number of excellent life his-

tory studies in the Rhodophycew which have been ‘of material

assistance in clarifying a very difficult field in plant morphology.

The more important of these contributions will form the subject

of this review.

Yamanouchi! in 1906 published an account of the life history

of Polysiphonia violacea based on cytological investigations of all

of its significant phases. He arrived at the following chief con-

clusions: (1) Carpospores have 40 chromosomes and tetraspores

20, i. e., half of this number. (2) The sexual plants in their

vegetative mitoses showed 20 chromosomes and therefore were

believed to arise from tetraspores. (3) The asexual or tetrasporie

plants showed 40 chromosomes throughout their vegetative his-

tory and consequently could be assumed to come from carpo-

spores. (4) The production on the asexual plants of tetraspores

is the result of a reduction division since the first mitosis in the

tetrasporangium is clearly heterotypic as shown by the character-

istic pairing of the 40 chromosomes and the separation of the

members of each pair; thus each tetraspore comes to have 20

chromosomes and is prepared to develop a sexual plant. (5) It

was therefore clear that the formation of tetraspores determines

the end of a sporophytie generation in the manner characteristic

of plants. (6) The gamete nuclei of the sperms and carpogonia

take the 20 chromosomes of the sexual plants and fertilization

gives the zygote nucleus with 40 chromosomes. (7) Descendants

from the zygote nucleus enter the carpospores which conse-

quently have 40 chromosomes and on germination would give

tetrasporic plants; certain complicated cell fusions during the

development of the cystocarp are purely cytoplasmic in char-

acter and evidently nutritional in function. (8) ‘‘There is thus

an alternation of a sexual plant (gametophyte) with a tetrasporic

plant (sporophyte) in the life history of Polysiphonia, the cysto- 3

carp being included as an early part of the sporophytie phase.”’

Yamanouchi’s contribution on Polysiphonia has in great meas-

1 Yamanouchi, S., ‘‘ The Life History of Polysiphonia skola i Bot. a a

XLII, 401-449, 1906. i

502

No. 596] NOTES AND LITERATURE 503

ure been the inspiration and largely furnished the basis for the

comparative studies that shortly followed. Lewis? in 1909 pre-

sented an investigation of Griffithsia Bornetiana which is in essen-

tial agreement with the conclusions of Yamanouchi that an alter-

nation of sexual and tetrasporic plants occurs. From the zygote

nucleus, containing 14 chromosomes, are derived the nuclei of the

earpospores. The tetrasporic plants have 14 chromosomes and

were assumed to come from carpospores. The first mitosis in the

tetrasporangium is a reduction division so that 7 chromosomes

enter the tetraspores and these were believed to produce sexual

plants. Lewis, however, held that the sporophyte generation was

represented by the sporogenous cells of the cystocarp and con-

sidered the tetrasporic plint to be a phase of an homologous

alternation of generations even though it was clear that the tetra-

sporic plant led up to the critical period of chromosome reduction

in the tetrasporangium. From a study of Griffithsia corallina

Kylin? questions the accuracy of the details of nuclear structure

and mitoses as given by Lewis and also his count of the chromo-

somes which in G. corallina appear to be as high as in Polysi-

phoma, Kylin describes and figures the reduction phenomena in

the tetrasporangium of Griffithsia in substantial agreement with

Yamanouchi and in an earlier paper gives a similar account for

Rhodomela virgata+

Extremely interesting are the results of experimental cultures

made by Lewis’ through which fruiting plants have been grown

from sporelings established in the laboratory on oyster shells that

were then favorably placed in the sea. From carpospores of

Polysiphonia violacea 6 tetrasporie plants were developed and 23

sterile. Tetraspores of Griffithsia Bornetiana produced a total of

60 sexual plants (32 male and 28 female) and 15 sterile. Tetra-

Spores of Dasya elegans gave 149 sexual plants (143 male and 6

female) and 139 sterile, the apparently large proportion of male

plants in the culture of this species probably being due to slower

maturing of the females. The sterile plants of these cultures were

2 Lewis, I. F., ‘‘ The Life History of Grifithsia Bornetiana,’’ Ann. of Bot.,

XXIII, 639-690, 1909.

: ? Kylin, a, “Die Entwicklungsgeschichte von Grifithsia corallina

(Lightf.) Ag., Zeitsch. f. Bot., VIII, 97-123, 1916.

*Kylin, H., ‘Studien über die Entwicklungsgeschichte von Rhodomela

virgata Kjellm.,’? Svensk. Bot. Tidskr., VIII, 1914.

“Lewis, I. F., ‘ Alternation of Generations in Certain Floridew,’’ Bot.

Gaz., LITI, 236-242, 1919. pees

504 THE AMERICAN NATURALIST [Vou. L

of course individuals which at the time of the examination had

not yet developed fruiting organs. More extended studies of

Lewis® have established the seasonal habits at Woods Hole, Mass-

achusetts, of the above species of Polysiphonia, Griffithsia and

Dasya. Tetraspores and carpospores were germinated on shells

that were fastened to piles and left over winter. In June tetra-

sporic plants of carposporie origin were abundant, which, re-

leasing their tetraspores in July, produced a crop of sexual plants

that matured their carpospores in August or early September.

The small plants from the carpospores winter over and produce

the tetrasporic plants of the first summer generation. Thus the

sexual generation is conspicuous in the late summer while tetra-

sporie plants surviving the winter are characteristic of the

spring. Belated growth of tetrasporic plants may result in their

fructification late in summer, so that a few small carposporie

plants also winter with the tetrasporic but they are rela-

tively scarce. The seasonal history is then in the main char-

acteristic ; tetrasporic plants appear in the spring and through

their spores produce a summer crop of sexual plants, from the

earpospores of which a generation of small tetrasporic plants in

favorable situations carries the species through the winter. This

experimental work thus supports at all points the conclusions of

Yamanouchi based on cytological studies.

The most important cytological work on life histories in the

red alge since the paper of Yamanouchi on Polysiphonia is that

of Svedelius presented in studies on Martensia, Delesseria, Nito-

phyllum and Scinaia. Part of this work, described in the next

paragraph, concerns the development of tetraspores in multinu-

cleate tetrasporangia (Martensia and Nitophyllum). All of it

supports to the fullest degree Yamanouchi’s theory of alternation

of generations in such species as have tetrasporie plants in their

life histories. The work on Scinaia presents a most interesting

hypothesis for the life histories of such red alge as do not develop

tetraspores.

The life history of Delesseria sanguinea is discussed by —

Svedelius’ in three papers embodying cytological work of a high

, I. F., “The Seasonal Life Cycle of Some Red Alge at Woods

Hole,’’ Science, XX XIX, 253, 1914.

7 Svedelius, N., ‘‘ Ueber den Generationswechsel bei Delesseria sanguinea, ?

Svensk. Bot. Tidskr., V, 260-324, 1911. ‘‘Ueber die Spermatienbildung bei a

Delesseria sanguinea,’’ ibid., VI, 239-265, 1912. ‘‘Ueber die Zystokarpien- —

bildung bei Delesseria sanguinea,’’ ibid., VIII, 1-32, 1914.

No. 596] NOTES AND LITERATURE 505

order. Tetrasporic plants have 40 chromosomes in their nuclei,

This number is reduced by the first nuclear division in the tetra-

sporangium. During synapsis the 40 chromosomes become

grouped to form 20 pairs, which become exceptionally clear in the

stage of diakinesis. The first or heterotypic mitosis separates the

members of the pairs, thereby halving the number and giving

after the second mitosis 20 chromosomes to each of the tetra-

spores. Vegetative mitoses in the male and female plants show

uniformly 20 chromosomes and it must be assumed that these

sexual plants develop from tetraspores. The chromosomes are

organized in the prophases of mitosis directly from a chromatic

network and without the interpolation of a spirem stage. The

spermatangia are cut off in pairs to the right and left of a mother

cell and in each spermatangium a uninucleate sperm is organized

(20 chromosomes) which on escaping leaves behind an empty

-eyst. The young carpogonium is uninucleate but an early divi-

sion differentiates the female gamete nucleus (20 chromosomes)

which remains in the carpogonium, and a trichogyne nucleus that

shortly breaks down. The carpogonium terminates a 4-celled

carpogonial branch borne by a cell (‘‘tragzelle’’) from which

develops also an auxiliary cell and certain sterile cells. The

sterile cells enlarge greatly after the fertilization of the carpo-

gonium and then break down, forming a mass of slime which

apparently serves to give space and protection for the develop-

ment of the gonimoblasts. In some way not clearly understood

the zygote nucleus of the fertilized carpogonium enters the auxil-

lary cell and from this cell as a starting point the gonimoblasts

arise as a dense growth of short filaments. The cells of the goni-

moblasts contain nuclei derived from that of the zygote and con- —

sequently have 40 chromosomes which are passed on to the carpo-

Spores developed in rows. From the carpospores must come the

tetraspori¢ plants with their 40 chromosomes. We have there-

fore in Delesseria an antithetic alternation of generations exactly

parallel with that of Polysiphonia, a diploid phase including the

gonimoblasts and the tetrasporie plants alternating with a oe

phase represented by the sexual plants.

Svedelius’

algæ with hi

in Martensia, one of the Delesseriacew. The yo bea de cell

* Svedelius, N., ‘Ueber den Bau und die Entwicklung der Florideengat-

Martensia,’’ Kungl. Svensk. Vet.-akad. Handl., XLIII, No. 1 T

opened a new vista in cytological studies on the seo o o i

s discovery of multinucleate tetraspore mother cells ee

506 THE AMERICAN NATURALIST [Vou. L

like its neighbors several nuclei which by division increase in

number as the cell enlarges until about 50 are present. Then a

general nuclear degeneration sets in coincident with an increase

in the amount of cytoplasm and only one nucleus in the center of

the plasma mass survives to give rise to the 4 nuclei that enter

the tetraspores. A similar situation was found by Svedelius® in

Nitophyllum punctatum, a related form of the same family, which

is described in greater cytological detail. Typical mitoses in the

young tetraspore mother cell give it a dozen or more nuclei all

with about 40 chromosomes, the diploid number for the species.

Many of the nuclei shortly begin to show signs of degeneration

by the disappearance of the chromatin so that only the nucleolus

remains to take the stain, and there is also a shrinkage of the nu-

clear membrane. The degeneration is not simultaneous, a few

nuclei increase in size and give indications of preparation for the

heterotypic mitosis as shown by clear stages of diakinesis. How-

ever only one nucleus carries the history of reduction further and

thus becomes the surviving nucleus of the tetraspore mother cell,

a nucleus very much larger than the degenerating structures that

lie about it in the cytoplasm. At diakinesis pairs of chromosomes

are clearly shown, about 20 in number. The members of the pairs

are separated by the first mitosis which is therefore heterotypic

and a reduction division as in Polysiphonia, Griffithsia and

Delesseria. The second mitosis, homotypic, gives the four nuclei

of the tetraspores, each with about 20 chromosomes. It is very

interesting to note that the red alge present illustrations of

nuclear degeneration at a period of reproduction when it may

be desirable to conserve the cytoplasm of a cell for a limited num-

ber of reproductive elements. This nuclear degeneration appears:

to be strictly analogous from a physiological point of view to

that exhibited in the oogonia of Vaucheria, Saprolegnia and

Albugo, in the sporangium of Derbesia, and in the oogonia of

certain forms of the Fucacee.

With the cytological and experimental evidence in complete

accord and so strongly in favor of the theory of an antithetic

alternation of generations in those red alge which have tetra-

sporic plants certain observations which at first thought appear

to offer exceptions to this theory naturally take on a high degree

of importance. The literature records a number of species which cee

9 Svedelius, N., ‘‘Ueber die Tetradenteilung in den vielkernigen Tetra-

sporangiumanlage bei Nitophyllum punctatum,’’ Ber. deut. bot. Gesell, Fe:

XXXII, 48-57, 1914. ——

No. 596] NOTES AND LITERATURE 507

have been reported to bear tetrasporangia on sexual plants. facts

which would be significant if it were established that such tetra-

sporangia were the seat of reduction divisions. Three of these

species have been studied cytologically and there is good evidence

that the reduction divisions are not present and that tetraspores

are either not fully matured or that the ‘‘tetrasporangium’’

develops a monospore. Yamanouchi and Lewis in their studies

found occasional sexual plants bearing what seemed to be tetra-

sporangia but in these cells the nucleus remained undivided in

Polysiphonia or produced several nuclei in Griffithsia while

cleavage furrows proceeded only a short distance into the cyto-

plasm. Svedelius’® has reported on a cystocarpie plant of Nito-

phyllum punctatum bearing tetrasporangium-like structures.

These cells were found around points in the thallus where pro-

carps had been formed and only reached their fullest development

when the procarps remained unfertilized. Their position on the

thallus therefore indicated a close correlation with the nutritional

physiology of the plant. These cells in their early history follow

exactly the course of normal tetrasporangia in this species; there

is a multiplication of nuclei and then a degeneration of all but

one which takes its position in the center of the cell. The sur-

viving nucleus does not divide but the entire protoplast slips

out of the thallus as a uninucleate monospore. Since the cysto-

carpie plant was haploid a reduction division in this tetraspo-

rangium-like cell would have been most irregular; it does not take

place and the monospore in its chromosome count has the same

value as a tetraspore. Agardhiella (Rhabdonia) tenera presents

another problem brought forward by observations of Osterhout.

Tetraspores of this plant sometimes germinate while still im-

bedded in the tissues of the parent with the further peculiarity

that the tetrad group behaves as a unit so that all four cells enter

into the formation of a sporeling. These epiphytic sporelings

commonly become sexual plants, as would be expected from the

germination of tetraspores. Osterhout, however, reports that

occasional tetrasporic plants are developed which would be irreg-

ular unless it were found that such plants came from tetra-

sporangia in which reduction divisions had been suppressed so

that such tetrasporangia, behaving like monospores, give rise to

diploid plants (tetrasporic) similar to the generations on which

1° Svedelius, N., ‘Ueber —~ an a ee von aie, oe :

A Beitrag zur Frage d :

er. deut. bot. Gesell., XXXII, ene 1914. =

508 THE AMERICAN NATURALIST [Vou. L

they were borne. It is probable that other red alge will exhibit

peculiarities of life history related to apospory, as seems prob-

able in this ease of Agardhiella.

There is enough evidence before us in these three studies by

Yamanouchi, Lewis and Svedelius to demand the utmost caution

in the consideration of other cases of ‘‘tetraspores’’ upon sexual

plants which have been brought forward by critics of the theory

of alternation of generations as applied to the red alge, cases

which have not as yet been subjected to the tests of cytological

research, One of the most interesting of these is Platoma Bairdii

(Farl.) Kuckuck! which in Helgoland apparently produces no

antheridia but develops cystocarps parthenogenetically. On

these cystocarpie plants are also found tetraspores indicating

that the plants are diploid in character which if true would ac-

count for their apogamous behavior. Platoma Bairdii therefore

appears to be one of the cases in which sexual organs are devel-

oped upon a diploid plant and not one in which tetraspores are

found on a haploid generation. A cytological study of this form

would be a matter of great interest and the only way in which the

facts may be determined. Also, there are the problems of the

paraspores or polyspores characteristic of a number of species in

the Ceramiacee. These are spores borne numerously in chains

(seirospores) or in dense glomerules and are found on tetrasporie

plants. Schiller”? regards them as homologous with tetraspores,

and as supporting the view of Oltmanns that the latter are repro-

ductive spores without significance for an alternation of genera-

tions, but until we know the cytology of their development an

opinion can have little value. Should the paraspores be formed

without reduction divisions, as seems probable, they will rank in

the same class with monospores and play no part in an alterna-

tion of generations. Should there be found evidence of a hetero-

typic mitosis in the development of these exceptional reproductive

cells they might perhaps constitute a modification of the tetrad

group characteristic of reduction divisions, but it does not seem

likely that this will prove to be the case.

There is left for consideration those red algæ which produce

no tetraspores at any phase of their life history. They include

a number of well-known types such as N emalion, Batrachosper-

11 Kuckuck, P., ‘‘Ueber Platoma Bairdii gpa bese 1? Wiss. Meeres- .

untersuch. Biol. daatant Helgoland., V, 187-2 :

12 Schiller, J., ‘‘Ueber Bau, Entwicklun ung, of see und Bedeutung der —

Parasporen der Ceramiaceen,’’ Oester. bot. Zeitsch., LXIII, 144, 203, 1913. —

No. 596] NOTES AND LITERATURE 509

mum, Scinaia, Helminthora, ete., and at present constitute the

field where research is most needed in connection with the life

histories of the red alge. Apart from asexual monospores the

reproduction takes place only through carpospores produced in

simple cystocarps. There are in the life histories no groups of

tetrad cells to suggest the position of reduction divisions which

explains the reasons why our understanding of alternation of

generation in the Rhodophycew has come chiefly through higher

forms exhibiting the tetrasporie phase. The first cytological re-

search on a life history among the forms lacking tetraspores was

the study by Wolfe? of Nemalion. This investigation brought

out a number of important discoveries: the presence of a nucleus

in the trichogyne, the mitosis which gives two sperm nuclei after

the fusion with the trichogyne, the proliferation of the cells

terminating the sporogenous filaments of the cystocarp to form

successive carpospore mother cells, the structure of the chroma-

tophore. With respect to the chromosome count in the life his-

tory Wolfe concluded that vegetative mitoses presented about 8

chromosomes and that 16 is probably the number shown by the

mitoses in the developing cystocarp. The period of chromosome

reduction was placed in the terminal cells of mature sporogenous

filaments, one of the daughter nuclei, following a reduction

mitosis, passing into a carpospore mother cell. The other re-

duced daughter nucleus was assumed to become established in

the cell from which successive carpospores are produced by pro-

liferation. The cytology of Nemalion is admittedly difficult and

in the absence of a conspicuous group of four nuclei character-

istic of reduction divisions (such as is found in the tetraspore

mother cell) it is natural that critics of Wolfe’s conclusions

should have suggested other possible positions in the life history

for the critical period of chromosome reduction. Thus it was

suggested that reduction might take place at the germination of

the carpospore, but Lewis" in a short note reports that the first

and later mitoses in the sporelings are vegetative mitoses with

about 8 chromosomes. <A second possibility, that reduction takes

place in the first division of the zygote nucleus in the fertilized

carpogonium, is strongly debated by Svedelius in his recent

paper on Scinaia

13 Wolfe, J. J., “cytological Studies on Nemalion,’’ Ann. of Bot.,

XVIII, 607-630, 1 1904

14 Lewis, I. F., ‘‘ The Germination of the Spore of N: emalion multifidum, ”?

Science, XXXV, 154, 1912.

510 THE AMERICAN NATURALIST [ Vor. L

Svedelius’® gives a cytological study of Scinaia furcellata

which suggests some fundamental readjustments of our concep-

tion of the cystocarp in those red alge which lack the tetrasporic

phase in their life histories. Monosporangia and spermatangia

occur on the same plants growing out of and becoming cut off

from mother cells at the surface of the thallus with evidence that

successive crops may be formed from the same mother cells.

Both monosporangium and spermatangium have a nucleus with

about 10 chromosomes and the similarity of their structure and

manner of development indicates that they are homologous or-

gans. The 3-celled carpogonial branch requires a somewhat de-

tailed account since its history presents phases not previously de-

scribed for the red algw. The terminal cell becomes the carpogo-

nium and a mitosis gives a nucleus to the trichogyne as has been

described for Batrachospermum, Nemalion, Polysiphonia, Rhodo-

mela, Delesseria and Griffithsia. From the second or hypogenous

cell is formed a group of 4 auxiliary cells rich in protoplasmic

contents. The third or basal cell of the carpogonial branch de-

velops finally the envelope of the cystocarp. The nuclei of the

earpogonial branch including the female gamete nucleus in the

carpogonium have 10 chromosomes which is the haploid number

of the Scinaia plant. After the fertilization of the carpogonium

the zygote nucleus, with 20 chromosomes, and consequently

diploid, passes into one of the auxiliary cells which have become

more or less fused together. In the auxiliary cell the large zygote

nucleus prepares for and passes through a heterotypic mitosis

with apparently clear evidence of diakinesis shown in the presence

of 10 pairs of chromosomes. A second mitosis gives the tetrad of 4

nuclei but only one of these becomes concerned with the develop-

ment of the gonimoblasts, passing back into the carpogonium

from which the gonimoblasts arise. The other 3 homologous.

nuclei of the tetrad together with the nuclei of the auxiliary cells —

take no further part in the history of the cystocarp. Nuclei of —

the gonimoblasts have 10 chromosomes and this number is passed

to the carpospores which are formed successively in rows; some —

of the gonimoblasts remain sterile and develop into long fila-

ments resembling paraphyses.

A most unexpected outlook upon the life histories of the red

alge will be opened if these conclusions of Svedelius on Scinaia —

are confirmed and if studies on ohare Batrachospermum, ete., —

15 Svedelius, N., isch-Entwi hichtliche Studien

Scinaia PETR Nor ov. Act. Reg. Soc. Scien. Upsal., IV, No. 4, 1915. i

No. 596] NOTES AND LITERATURE 511

should also determine the position of chromosome reduction to be

at the first mitosis of the zygote nucleus. The cystocarps of these

plants would then be interpreted not as a diploid sporophyte

generation but as a special haploid phase and the carpospores

would have the same value as monospores. There would be no

antithetic alternation of generations in such forms of the red

alge. Svedelius proposes the term haplobiontic for red alge

with this type of life history to be contrasted with diplobiontic

forms (Polysiphonia, Rhodomela, Griffithsia, and Delesseria)

where the gonimoblasts have been shown to be diploid in charac-

ter and chromosome reduction finds its place in the terasporangia

of an asexual generation. The diplobiontie type of life history is

naturally conceived by Svedelius to arise from the haplobiontic

by a delaying of the reduction divisions so that gonimoblasts

carry forward the diploid number of chromosomes to the carpo-

spores. The carpospores being diploid develop an asexual gene-

ration and their diploid sporangia becoming the seat of the re-

duction divisions take the characters of tetrasporangia, a new

type of reproductive organ. The occasional suppression of re-

duction divisions in the tetrasporangium may be meters at

times to transform this cell to a monosporangium, as in

phyllum, showing close analogies between the two structures.

In connection with the conclusions of Svedelius on Scinaia it

should be remembered that Allen has placed the reduction period

in the life history of Coleochete at the first mitosis of its zygote

nucleus. If this view is correct the cellular body developed in

the oospore of Coleochete, preliminary to the formation of a

crop of zoospores, is a special haploid structure and may be

compared with the gonimoblastie development in Scinaia as in-

terpreted by Svedelius. The haplobiotic red alge of Svedelius

then, if clearly established, would have the same type of life his-

tory as Coleochete and bearing in mind the resemblance of the

young oogonium of Coleochete to a carpogonium and trichogyne,

the resemblance of its antheridia to a cluster of spermatangia

and the branched filamentous structure of its thallus, the view

that Coleochete represents a type from which the red alge might

have arisen takes on added strength.

It is interesting to note how far removed from the theories of

Schmitz and Oltmanns are the more recent interpretations of the

life histories in the red alge. The secondary or double fertiliza-

tion assumed by Schmitz as an explanation of the union of-

ooblastema filaments with — cells was s discredited gi suck

512 THE AMERICAN NATURALIST [Vou. L

manns’s work and has received no support from the cytological

studies of Yamanouchi, Lewis, Kylin and Svedelius. Oltmanns’s

view that tetraspores have no fixed place in ontogeny and are

without relation to a sporophyte generation has been over-

whelmed by the cytological work of the authors mentioned above.

And now Svedelius argues that even the old view that cysto-

carps represent a sporophytic phase can not be correct for a

group of the red alge which he terms haplobiotie. The situation

as it now stands may be summarized as follows: An antithetie

alternation of generations may be expected wherever tetrasporic

plants are found in a life history and in these forms the gonimo-

blasts also constitute a phase of the sporophyte. There is no

sporophyte in the ‘‘haplobiotie’’ red alge (e. g., Scinaia with

Nemalion, Batrachospermum, ete., not yet studied from this

standpoint) and the gonimoblasts of these forms represent a

haploid development of the zygote in position upon the plant,

carpospores being equivalent to monospores. Fusions with auxil-

iary cells are merely cytoplasmic in character and associated

simply with nutritive functions. The diplobiotie red alge have

come from the haplobiotic, which, carrying forward the reduc-

tion divisions through the gonimoblasts and carpospores to a new

generation, the tetrasporic plants, have established the reduction

divisions in the tetrasporangium. In theory these views are

simple and logical. For the antithetic alternation of sexual and

tetrasporic plants the evidence is considerable and convincing.

What will be the conclusions for the ‘‘haplobiotic’’ types? Will —

future studies establish their existence?

RADLEY Moore Davis

UNIVERSITY OF PENNSYLVANIA,

July, 1916

Special Articles:

i ‘The Proces of Feeding in the Oyster: Professor Caswell

The American

Naturalist

S intended for publication and bocks, etc., intended fer review should =

sent to the Editor of THE AME

Short articles containing summ

N T, Garrison-on- Hueson, New Yor

es of research work bearing on the

epee of — evolution are papeciatly welcome, and wiil be given preference

ic heaved reprints of peerings are supplied to authors free of charge.

folie peertiata will be supplied at c

Subscriptions

and advertisements should be sent to as e

The

subscription price is four dolla ear. Foreign postage

ian postage twenty-five rs additional. The a for single copies is

forty cents. The advertising rates are Four Dollars for

THE SCIENCE PRESS

Lancaster, Pa. Garrison, N. Y.

_ NEW YORK: Sub-Station 84

, April 2, 1908, at th

t Lancaster, Pa., under the Act of

Congress of March 3, 1879.

aMi HISTORY OF MAN

,000 Specimens i

ties, i , Carvings, T

Guaranteed localities and Type Collections fo

Museums a Specialty. 2e eS —

OLDMAN, 77 Brixton Hill, London, England

The Diatomaceae of Philadelphia eae

LES S. BOYER, A.M.,

cription of the A Ekm ee ‘flora of Philadelphia

ara ea ovine er Soc: occurring in ra

al 3 in th

laware has just been — i

=e raf eee Com 143

More than 700 original on u pargas times,

complete index. = x12. Bound in Art Vellum cloth.

Sent postage pai

Charles S. Boyer, <5

1640 Columbia Avenue Philadelphia, Pa. —

SCIENCE

i WREKLY JOURNAL DEVOTED TO THE ADVANCEMENT OF sorENcE

Fripay, Auausr 4, 1916 —

The Method of Growth of the Lymphatic System: Professor |

Florence R. Sabin.

eaa. De fa o Vay. The

i raa Peninsula: Albert B. Reagan. Nomenclatorial

_ Facts: Morgan Hebard. Sylvester and Cayley: Professor

G. A. Miller.

Scientific

Books: —

Biarigon end Lendis on the Fundamental C Conceptions

of Modern A. Miller. Curtis on

; Eros unai Vad the Circulation of the Blood:

Dr: Perey M. Di

Ths American Amocistion of Museums: Dr. Paul M. Rea.

THE

AMERICAN NATURALIST

Vou. L. September, 1916 No. 597

THE STATUS OF THE MUTATION THEORY,

WITH ESPECIAL REFERENCE TO

(ENOTHERA?

DR. HARLEY HARRIS BARTLETT

THE more or less controversial literature of the muta-

tion theory is so scattered and involved that few except

specialists have been able to follow it with any degree of

satisfaction. A recent book by Gates, ‘‘The Mutation

Factor in‘ Evolution, with Particular Reference to Gino-

thera,’’ will help the general biologist to an understanding

of the situation, but even in the short time since it went

to press there have been developments of such signifi-

cance that I am glad of this chance to review the subject.

The crux of the whole controversy is this: Are the dis-

continuous variations which oceur in cultures of Œno-

thera true mutations, which might appear either in pure

lines or in hybrids, or are they segregates from geneti-

cally impure lines? The mutationist and his Mendelian

critic give diametrically opposite answers to this ques-

tion. Some geneticists are beginning to feel that there is

justification for taking a middle ground, that the muta-

tions are non-Mendelian, but that they are nevertheless a

phenomenon of hybridism, and occur only in impure lines.

The most extensive researches upon mutability have

been carried out with @nothera Lamarckiana. This

1 Papers from the Department of Botany of the University of Michigan,

No. 150. Read before the American Society of Naturalists at Columbus,

Ohio, 30 Dec., 1915. Based in part upon unpublished experiments conducted

at the U. S. Bureau of Plant Industry and published with the-permission

of the Secretary of Agriculture. ne

513

514 THE AMERICAN NATURALIST [ Vou. L

plant gives rise to mutations of two main types, those in

which the chromosome number differs from that of the

parent form, and those in which it does not. The most

striking ones belong to the former category; among them

we need mention particularly only Œ. gigas and Œ. lata,

which have, respectively, 28 and 15 chromosomes, instead

of 14, the normal number in the species. Gates has espe-

cially emphasized the fact that in the mutating Œnotheras

the pairing of the chromosomes, previous to the reduc-

tion division, is very loose, a condition that would favor

irregularities in the distribution of the chromosomes to

the gametes. By irregular reduction divisions, gametes

with a greater or less number of chromosomes than 7

might easily be formed. From these irregular gametes

there would be derived, in turn, zygotes with irregular

chromosome numbers. Mutationists are now pretty well

agreed that the characters of certain mutations are corre-

lated with an unusual complement of chromosomes. If

one believes at all that the chromosomes provide the

mechanism of Mendelian inheritance, it is hard to escape

the conclusion that the cytological studies of Lutz, Gates

and others give a firm basis for removing at least part of

the mutation phenomena from the domain of Mendelian

segregation.

Even Davis, who has been one of the chief opponents of

the mutation theory, has admitted that some of the vari-

ants from @/nothera Lamarckiana and other species are

probably due to irregularities in chromosome distribu-

tion. He makes the point, to be sure, that the unpaired

condition of the chromosomes previous to reduction in

Œ. Lamarckiana is itself exceptional, and presumably the

result of a hybrid constitution. Davis himself has shown

that in at least one of the numerous strains of Œ. grandi-

flora which occur at the type locality of that species m —

Alabama, the chromosomes, in preparation for the reduc-

tion division, become associated definitely and closely into

ring-shaped pairs. Moreover, this same strain shows a

negligibly low pollen and seed sterility. Davis views it as :

essentially a genetically pure strain, whether ro wo , :

No. 597] THE MUTATION THEORY 515

its morphology or by its physiological behavior. In 1912

de Vries and the speaker visited the locality from which

Davis’s, strains came, and found such a confusion of dif-

ferent types growing together that it was impossible to

doubt that the entire Œnothera population was hybridized

to a greater or less degree. Some of the forms belonged

to the series of Œ. grandiflora, being large-flowered and

open-pollinating, whereas others were small-flowered, self-

pollinating types, showing obviously the effects of hy-

bridization with Œ. Tracyi, another southern species. It

is curious that the particular @nothera to be put for-

ward as probably genetically pure, judged by Davis’s

cytological criterion, or by Jeffrey’s pollen test, should

come from a locality where hybridization is so prevalent.

that one would hardly expect to find among the open-

pollinating forms a single unhybridized plant. It will be

remembered that Jeffrey has attacked the mutation theory

from the point of view that pollen abortion necessarily

indicates hybridity. By this criterion the small-flowered

practically cleistogamous species which self-pollinate

generation after generation must be adjudged highly im-

pure, although the evidence is all to the eontrary. There

is every indication that pollen abortion is a frequent con-

comitant of mutation, as well as of hybridization. It

seems not unlikely, therefore, that the unpaired condition

of the meiotic chromosomes may have a causal relation-

ship with pollen abortion as well as with the production.

of those types of mutations which have a chromosome

number different from that of the parent species.

It is ordinarily supposed that a mutation is determined

when the reduction division takes place. This may be the

case with the mutations in which there are irregular

chromosome numbers. Even in such cases it may be that

the germ plasm has undergone at some other point in the

life cycle a premutative modification of such a nature as

really to predetermine the kinds of mutated gametes

which will subsequently appear. A physiological premu-

tation might, for instance, bring. about the condition

which results directly in loose chromosome pairing, and

516 THE AMERICAN NATURALIST [ Von. L

indirectly in the formation of several different types of

mutated gametes. Many mutations are not concerned

with such obvious changes as the shifting of chromo-

somes, but seem rather to depend upon physico-chemical

and chemical alterations of the germ plasm. Obviously

some such alterations would result in physiological as

well as morphological mutations. Just as certain mor-

phological changes are not advantageous, or even dis-

tinctly harmful, so certain physiological changes might be

harmful and lead to sterility. Some premutations (and

by a premutation I mean the inauguration of an unstable

condition in the germ plasm) might be of such a nature

that the nutrition of the spore mother cells would be inter-

fered with. These might fail to develop, fail to undergo

the reduction division or might give rise to defective

daughter cells.

Thus mutation, equally as well as hybridization, may

account for sterility. There are several groups of plants

in which sterility has apparently come about without the

possibility of hybridization. Perhaps Davis’s genetically

pure Œnothera grandiflora, with perfect pollen, provides

a case of hybridization without subsequent mutability.

Those who assert that germinal instability comes about

only by hybridization can bring forward no proof of their

assertion. Conversely, the mutationist can not prove

that any plant in existence has had an unmixed ancestry.

The most that he can do at present is to show that muta-

tion takes place in strains which are genetically pure, and

that the purity is of relatively long standing. One can

only conclude that Davis’s and Jeffrey’s suggested cyto-

logical and morphological evidences of hybridity, if veri-

fied, will merely substitute hybridization for premutation

as a cause of germinal instability. They will not in any

way afford support to the Mendelian conception of

mutation. .

Nothing could be more obvious than the paths which

are marked out for the student of the Enothera problem

by the interesting cytological clue afforded by the un-

paired chromosomes of Œnothera Lamarckiana and other

No. 597] THE MUTATION THEORY 517

species. Some one must find out whether or not the un-

paired condition occurs in hybrids whose parents do not

show it. Strange as it may seem, after all the discus-

sion of hybridity as a possible cause of mutability, no one

has yet shown, or tried to show, that mutability occurs

in any hybrid between non-mutable parent species. This

would seem to be one of the most crucial experiments that

could be performed, and one of the easiest. It would be a

very attractive problem to attempt to produce the un-

paired chromosome condition by hybridization and then

to prove it definitely correlated with the particular types

of mutability which are characterized by disturbances of

the chromosome mechanism.

Even if it were possible in the time at my disposal to

review the evidence that the chromosomes provide the

mechanism of inheritance, it would hardly be necessary

to do so. The brilliant work of Morgan and his students

on the association in Drosophila of groups of characters

with definite chromosomes is well known to every one. In

(Enothera the investigations of Gates, Lutz and others

have shown a connection between chromosome alterations

and the characters of certain mutations so obvious that

it can not reasonably be disregarded.

There are still a few geneticists, however, who believe

that the chromosome cycle has no fundamental signifi-

cance in connection with Mendelian phenomena. Heri-

bert-Nilsson would ascribe as much weight to a change

from flat to crinkled leaves, to choose an example at

random, as he would to a change from the 2x to the 4x

chromosome number. Such an attitude is forced upon

one who attempts to explain all mutations as Mendelian

segregates, as this author does. He has made much of a

case, observed by Geerts and Stomps, in which the chro-

mosome number of a hybrid between Gnothera gigas and

Œ. Lamarckiana became reduced from 21 to 14, probably

through the agency of an irregular reduction division,

without the loss of the gigas characters. This case is

cited to prove that the characters of this mutation do not

depend upon the supernumerary chromosomes. Nothing

518 THE AMERICAN NATURALIST [ Vou. L

is more probable, however, than that the chromosomes

are qualitatively different, and that the gigas characters

depend upon the duplication of some, not all, of the chro-

mosomes. An irregular reduction division might well

result in the retention in duplicate of those particular

chromosomes upon which the gigas characters depend.

The very characteristic aspect of @nothera lata has been,

ascribed to its single supernumerary chromosome. Miss

Lutz has shown that many mutations with 15 ‘chromo-

somes do not have at all the characteristic lata appear-

ance, which must therefore be attributed to the duplica-

tion of a particular chromosome, rather than of any

chromosome.

The production of mutations with irregular chromo-

some numbers is not confined to Hnothera Lamarckiana.

Two other species have given rise to mutations with 28

chromosomes, and in one case, that of Œ. stenomeres, the

gigas mutation is entirely comparable in its characters

with @. gigas de Vries. Its wood structure has been

compared with that of the parent species and has been

found to present deviations as great as those which are

apparent in the external aspects of the two plants. The

differences concern not only the relative size of the ele-

ments, but also their shape, and, to a certain extent, their

distribution. In typical Œ. stenomeres the medullary

rays are sometimes 140 cells high, whereas in mutation

gigas they are typically less than 25 cells high, and as

far as we have observed, never over 50. It is very sig-

nificant indeed that striking structural alterations in the

most conservative tissues of a plant may be instituted by

a single mutative evolutionary step. :

As far as the mutations with modified chromosome

numbers are concerned, there is the best of evidence that

the processes of mutation and Mendelian segregation are

absolutely distinct and independent. The evidence is not

only cytological, but also genetical, for no mutations of this

class show Mendelian inheritance when crossed with their

parent forms. Their significance in evolution is illus-

trated by many widely separated groups of plants in

No. 597] THE MUTATION THEORY 519

which species or genera are set apart from their allies by

the possession of a different number of chromosomes.

Such variations in chromosome number have been found

in many cases among the Rosacex, a family noteworthy

for the complications which it presents to the systema-

tist. Certain Japanese species of Viola exhibit variations

in chromosome number which give a clue to the way in

which the numerous forms of this complex genus have

evolved. Similar variations occur among the Orchid-

aceæ, one of the largest families of flowering plants.

There remain to be considered a large number of muta-

tions in which the chromosome complement has not been

shown to differ from that of the parent form. Such mu-

tations are frequent in @nothera. De Vries has ob-

served them in the case of Œ. Lamarckiana and more

recently he and other workers have observed them in

other species, belonging to the small-flowered, self-polli-

nating portion of the genus. In connection with these

mutations it will be necessary to consider more in detail

the criticisms brought against the mutation theory by

Bateson, Davis, Heribert-Nilsson and others.

In the past, most objections to the mutation theory

have been based upon the supposition that Gnothera

Lamarckiana is a hybrid of garden origin. I-am forced

to admit that I am not satisfied with any evidence thus far

offered that this species, in the form familiar in cultiva-

tion, is or ever was a wild constituent of our flora. Never-

theless I venture to predict that it will eventually come

to light in some obscure locality and that its character as

a natural species will be established. Whether it is a

natural species or a product of floriculture is of relatively

little importance, however, in view of the fact that none

of the mutation phenomena are peculiar to it. Several

other species are known which are equally mutable and

which are now elements of our flora. Moreover, they are

small-flowered, self-pollinating forms, and therefore better

suited to mutation studies than large-flowered, open-pol-

linating forms such as @. Lamarckiana, which in nature

must frequently be hybridized.

520 THE AMERICAN NATURALIST [Von L

Before mutation studies had been extended to other

species of the genus from Œ. Lamarckiana, Davis began

a series of experiments with the object of reproducing

the latter species as a hybrid of known origin, His first

experiments, involving Œ. grandiflora as one parent,

were unsuccessful in producing a plant that bore more

than a superficial resemblance to Œ. Lamarckiana. Some

of the hybrids showed mutability, but none were obtained

which did not show obvious segregation in addition to

the mutability. Moreover, the mutations were not shown

to have been induced by hybridization, since none of the

parent strains were tested for constancy. As de Vries

suggested, the mutability was probably an inherited tend-

ency from one or both parents.

Later hybrids, between Œ. franciscana and Œ. biennis,

were much more successful, in that they bore a much

closer resemblance to Œ. Lamarckiana. The writer saw

hybrids last summer in Davis’s garden that would surely

have been placed by any except the most ultra-critical

systematist under Œ. Lamarckiana. They are being car-

ried into another generation, and the results will be looked

forward to with much interest. A true synthetic Œ. La-

marckiana must show mutability, but must otherwise

come true from generation to generation. Moreover, it

must give twin hybrids in certain crosses with other spe-

cies. Even if Davis’s later hybrids fulfill these condi-

tions, they will not demonstrate the origin of mutability

through hybridization, for one of the parents, Œ. biennis,

has been shown by de Vries and Stomps to be a mutable

species, and the other, Œ. franciscana, has not been tested.

To have much weight, an experiment such as Davis’s

must show the origin of mutability de novo in a hybrid

from non-mutable parents.

A more recent phase of the effort to prove Œ. La-

marckiana a hybrid dates from the publication, in 1914,

of a paper by O. Renner. This author proposed a simple

Mendelian hypothesis to account for the twin hybrids and

high seed sterility of Œ. Lamarckiana. It is well known

that in this species about half of the seeds are empty or

No. 597] THE MUTATION THEORY 521

do not contain normal embryos. Moreover, when crossed

with certain species, the first hybrid generation consists

of two types, the twin hybrids of de Vries. Renner as-

sumes that Œ. Lamarckiana is heterozygous and that it

produces two types of functional gametes. Its progeny

under ordinary circumstances would therefore be ex-

pected to consist of recessive homozygotes, heterozygotes,

and dominant homozygotes in the familiar 1:2:1 ratio.

He further assumes, however, that the homozygotes are

incapable of developing beyond a young embryonic stage,

and that the species is therefore maintained in a hetero-

zygous condition from generation to generation. This

simple hypothesis obviously does not account for the

mutability of Œ. Lamarckiana. It has been amplified

with this end in view by Heribert-Nilsson, whose highly

involved explanation of mutability from the standpoint

of the plural factor hypothesis must receive a brief con-

sideration. For several years this worker has busied

himself in an attempt to demonstrate Mendelian inheri-

tance in Œ. Lamarckiana. In one case he thought he had

found simple monohybrid segregation in crosses between

red- and white-nerved races, and announced that the

nerve color acted as a simple Mendelian character. It

developed later, however, that his ratios were aberrant,

and that the progenies entirely lacked a class of plants

homozygous with regard to the supposed dominant char-

acter. According to Heribert-Nilsson’s interpretation,

the progenies consisted only of heterozygotes and reces-

sive homozygotes. The elimination of the hypothetical

dominant homozygotes he accounted for by assuming that

in certain cases an incompatibility, or, as he puts it, a

prohibition, exists between like gametes. His whole hy-

pothesis is based upon this idea of prohibition. He as-

sumes that the assemblage of characters which we recog-

nize in Ænothera Lamarckiana may be brought about by

many combinations of plural factors. Any one of these

plural factors in the heterozygous condition gives a plant

the Lamarckiana habit, and prohibition prevents the pres-

ence of any of them in the homozygous condition. Segre-

Do THE AMERICAN NATURALIST [Vot. L

gation may lead to the production of pure recessives,

lacking all the plural factors which give the Lamarckiana

aspect. These recessives are the supposed mutations.

Pure dominants, on the contrary, can not be realized.

This, in brief, is the Mendelian explanation of muta-

bility. It involves the important assumption that the

mutations which breed true are Mendelian recessives.

The mutations with irregular chromosome numbers have

been shown not to belong in this category. The remaining

mutations, for many of which the cytological data are

lacking, may conveniently be divided into two classes, (1)

those which come true when self-pollinated, or, at any

rate, do not include the parent species in their progeny,

and (2) those which give a mixed progeny consisting of

the mutational and parental forms. If tkere is any possi-

bility whatever that the Mendelian explanation of muta-

bility is true, it should at least account for the first and

simplest of these two cases. We shall therefore confine

our attention for the moment to mutations which give a

constant progeny.

De Vries found that certain of the original mutations

from @nothera Lamarckiana were of the Mendelian type.

These mutations are assumed by Heribert-Nilsson to be

recessives which have corresponding homozygous domi-

nants, the latter being the strains of Œ. Lamarckiana

which do not give rise in every generation to the muta-

tions in question. Other mutations, isolated by Heribert-

Nilsson himself from Œ. Lamarckiana, are produced in

every generation, and are therefore, according to this

author, recessives which have no corresponding homo-

zygous dominants. If this were the case, they would be

recessive when crossed with @. Lamarckiana regardless

of which way the cross was made. As a matter of fact,

Heribert-Nilsson made his crosses with @nothera La-

marckiana as the pistillate parent, and therefore obtained

the results which he expected. If the crosses had been

made the other way, there is very good reason to believe

that he would have got the most unexpected results, and

would never have advanced his Mendelian hypothesis.

No. 597] THE MUTATION THEORY 523

The speaker has recently observed, in several species of

(Enothera other than Œ. Lamarckiana, the origin of a

large number of different mutations. Several of these

have been found to belong to the type which we are at

present considering. That is to say, they give a progeny

which does not contain the parent species, and the muta-

tions themselves are produced by the parent species in

every generation. In the case of one mutation, deseribed

a year ago as Œ. pratincola mut. nummularia, the chro-

mosome number has been determined as 14, the typical

number in the group. The remarkable fact about these

mutations of Œ. pratincola, as far as work with them has

gone, is that their crosses with the parent species are

identical with the pistillate parent in the first hybrid gen-

eration. Mutation pollinated with parent species yields

the mutation. Species pollinated with mutation yields

the species.

This most interesting state of affairs is absolutely at

variance with the attempted Mendelian explanation. It

can be understood on the supposition that two types of

gametes are produced, which are by no means equivalent.

One type bears most of the characters which differentiate

the different species and forms from one another. The

other type seems to carry characters which are likely to

be common to a number of different species. In the par-

ticular species which gives rise to the mutations under dis-

cussion the gametes of the former class are female, those

of the latter, male. Thus it follows that a mutative modi-

fication of the germ plasm in one of these species might

affect only characters which were borne by one of the two

kinds of gametes. If so, we would have at once a simple

explanation of the behavior of the mutations which give

matroclinic crosses with their parent species.

The same idea may readily be extended to cover the

cases of mutations which give progenies containing both

the mutational and the specific types. Perhaps the muta-

tive change is a reversible one, and certain gametes in

each generation show reversion from the mutated to the

unmutated condition. Or perhaps in some species there

524 THE AMERICAN NATURALIST [ Vou. L

are male and female gametes of both types, but certain

mutative changes are sex limited. In the following dis-

cussion I shall designate the two types of gametes as a

and 8 gametes. The former are those which bear the

most distinctive specific characters of the various forms,

whereas the latter bear the more general characters. The

known facts seem to be accounted for if we assume that in

fertilization the conjugation of an a with a 8 gamete ordi-

narily takes place, but not the conjugation of two 8 gam-

etes. In certain cases it seems that fertilization takes

place by the fusion of two a gametes and it appears likely,

also, that some species produce no £ gametes. Some spe-

cies produce a and 8 gametes of both sexes. Others do

not seem to do so. It sometimes seems to be the case that

-the female gametes are all a. When a mutation takes

place the modified character is perhaps Mendelian if it is

borne by both a and £ gametes, but non-Mendelian if it

affects only the a gametes of a species in which fertiliza-

tion takes place by the fusion of an a with a 8 gamete.

This conception of non-equivalent gametes has been

highly developed by de Vries, in a somewhat different way

from that outlined above. It has many obvious advan-

tages in explaining the Œnothera situation. It explains

seed sterility as well if not better than the Mendelian

hypotheses of Renner and of Heribert-Nilsson, hypotheses

which are based of course upon the idea of gametic equiv-

alence. It explains why certain reciprocal crosses are

alike, and others unlike, why some of them breed true,

whereas others show segregation, why certain crosses

yield twin hybrids, and why the twins are, respectively,

matroclinic and patroclinic. It also explains other com-

plications which are quite unintelligible from a Mendelian

standpoint. I would by no means give the impression that

there are not many phenomena which remain obscure, but

I do wish to emphasize very strongly that a flood of light

is thrown upon the @nothera situation by the conception

of non-equivalent gametes. | |

By way of illustration, let us consider for a few mo-

ments the phenomenon which I have called mass mutation.

No. 597] THE MUTATION THEORY 525

Mass mutation differs from ordinary mutation only in

that the mutations, instead of being produced in small

numbers, are produced in very large numbers. For ex-

ample, the frequency of mutations in Œ. Lamarckiana,

which shows ordinary mutability, is roughly 2 per cent.

In certain strains of Œ. Reynoldsii and Œ. pratincola, on

the contrary, the number of mutations rises to 50 per

cent., or even 100 per cent., of the progenies. According

to Mendelian conceptions, it is impossible to get extracted

recessives in a progeny in excess of 334 per cent., and in

order to get this many we must grant the elimination by

prohibition of the corresponding dominants. What shall

we say, then, of progenies containing 499 mutations out

of 500 plants, a condition which has actually been realized

in my cultures of Œ. pratincola? It is impossible to in-

voke the elimination of a large class of typical plants, for

the typical zygotes are known to be stronger and better

fitted to develop than the mutational zygotes. My own

explanation is that most of the female germ cells of Œ.

pratincola are a gametes and the male, @ gametes. The

phenomenon of mass mutation consists in the wholesale

production of modified a gametes, a’, a”, a”, a’’”’, ete., each

of which corresponds to a different mutation and has

characters which impress a distinctive habit on the zygote

which is formed by fusion with an unmodified 6 gamete.

In accord with this hypothesis the reciprocal crosses be-

tween mutation and parent species are matroclinic. Mu-

tation pollinated with species gives mutation. Species

pollinated with mutation gives species.

Mention has already been made of the mutations which

by self-pollination give progenies containing both the mu-

tational and the specific types. If the mutation is cross-

pollinated with pollen from the specific type, the progeny

is a mixture of two types, just the same as if self-pollina-

tion had occurred. On the contrary, if the specific type

is pollinated by the mutation, only the specifie type

occurs in the progeny. Here, it seems, we have a case

where the modification which results in the production of

a’ instead of a gametes is reversible. Cases of this kind

526 THE AMERICAN NATURALIST [ Von. L

Heribert-Nilsson refers to (I give a literal translation)

as ‘‘heterogamous combinations which are recessive only

in the female gametes, but in the male gametes continuously

heterozygous.’’ As far as I can interpret this vague

statement at all, it involves a decidedly unique concep-

tion, namely, that the individual 2% mutation embodies

two different kinds of germ plasm, a homozygous female

germ-plasm which will give one kind of cells when the

reduction division takes place, and a heterozygous male

germ-plasm, which will give two kinds of cells. I think

that no one will be inclined to adopt this altogether revo-

lutionary and useless hypothesis. It is by no means cer-

tain, after all, that the mutations which show the type of

inheritance in question do not belong to the class with:

irregular chromosome numbers. With one exception they

have not been examined cytologically. Cnothera lata, a

mutation which shows this type of inheritance, has 15

chromosomes. Consequently there is an opportunity for

the formation of two kinds of gametes, with 7 and 8 chro-

mosomes, respectively. The male gametes with 8 chro-

mosomes appear to be eliminated. As a result, zygotes are

formed with 7 + 7—14 and 8 + 7—15 chromosomes. The

former are Œ. Lamarckiana, the latter are Œ. lata. This

beautiful correlation of cytology with inheritance has

been worked out by Gates and Thomas.

In either event, whether the mutations which throw the

specific type in every generation have a regular or an

irregular chromosome number, the mutation hypothesis

provides a far more plausible explanation for their be-

havior than the Mendelian hypothesis.

It must be clear by this time that the speaker finds

incredible the arguments that have been brought forward

in favor of the idea that mutation and Mendelian segre-

gation are the same. Doubtless it often happens that a

mutated germ cell fuses with a typical germ cell and pro-

duces an ordinary Mendelian heterozygote. If the mu-

tated character is recessive, and the dominance is com-

plete, the first hybrid generation will of course resemble

the parental type, and the second hybrid generation will

No. 597] THE MUTATION THEORY 527

show simple segregation. The mutation will appear for

the first time in 25 per cent. of the progeny. De Vries has

recently reported that the dwarf mutation from @nothera

gigas is of the simple recessive Mendelian type. We must

believe, in a case of this kind, that the factor whose modi-

fication results in dwarfness is present in all gametes. It

does not follow, however, that the gametes are all equiva-

lent with respect to the factors for other characters.

In connection with the discussion of Davis’s hybrids

which resembled Gnothera Lamarckiana I mentioned that

the mutability shown by them was probably inherited

from one or both parents. There seems to be some scepti-

cism about the inheritance of mutability as a character.

Much of my own experimental work of the last two years

has involved Œ. pratincola, a mutable species which has

already been refered to several times. There is another

species from the same locality which is rather closely

allied to Œ. pratincola, but differs in enough regards so

that the hybrids between them can be studied with great

satisfaction. The second species, Œ. numismatica, is im-

mutable, as far as my experience extends. At any rate it

is very much less mutable than Œ. pratincola. The cross

Œ. pratincola X Œ. numismatica gives twin hybrids, one

of which is exactly like the pistillate parent except in one

minor pubescence character. The reciprocal cross, Œ.

numismatica X Œ. pratincola, is to all outward appear-

ances the same as the pistillate parent. We have here a

most striking case of matroclinic reciprocal hybrids. I

am inclined to believe that most of the differences between

the two species reside in the a gametes and that the B

gametes are essentially similar. In accord with this hy-

pothesis nothing could be more interesting than to find

that the pratincola-like hybrid is mutable, and produces

the same types of mutations that Œ. pr atincola itself does.

This result, it seems to me, is of the highest significance.

It indicates that the germ plasm of Œ. pr atincola is in a

labile condition, and that this condition 1s not modified

when a zygote is formed by the fusion of its a gamete with

the 8 gamete of a different and stable, or at least relatively

528 THE AMERICAN NATURALIST [ Vou. L

stable, species. We could hardly find better proof that

such mutations in @nothera involve the a gametes, and

are apparent in the zygotes without the need of subse-

quent segregation because the factors involved have no

counterparts in the 8 gametes.

The same crosses, however, afford evidence that certain

characters are carried by both a and 8 gametes, and may

therefore prove to show Mendelian segregation. The buds

of Œ. numismatica have a short viscid pubescence which

is lacking in Œ. pratincola. The matroclinic hybrid Œ.

pratincola X Œ. numismatica can be distinguished from

the pistillate parent only by the presence of this hair-type,

inherited from the pollen parent. When the second hybrid

generation is grown, segregation with regard to this char-

acter takes place, and part of the progeny can not possibly

be distinguished from Œ pratincola.

In these results we have a clue to the segregation shown

in certain hybrids, and the lack of it in others. Most of de

Vries’s hybrids have involved @. Lamarckiana, a species,

according to my interpretation, with very dissimilar a and

B gametes. He has therefore obtained and described

many measurably constant hybrids. Davis, however,

studying @. grandiflora, which may conceivably have but

one type of gametes, has found segregation the rule

rather than the exception. In his later studies, involv-

ing Œ. franciscana and Œ. biennis, he has obtained twin

hybrids within each of which there was a considerable |

degree of segregation. All of these varying results will

eventually become coordinated as we become more used to

distinguishing between non-Mendelian and Mendelian

characters.

Another point which must be mentioned is the fre-

quence with which the various types of mutations give

rise to one another. For example, two mutations of Œ.

pratincola, mut. nitida and mut. fallax, each give rise to

plants of mut. numularia, which are as typical as though

they had been derived directly from Œ. pratincola. As

already brought out, some mutations appear to be re-

versible in that they revert to the parent species in part

No. 597] THE MUTATION THEORY 529

of every progeny. The germ plasm seems to be a system

capable of existing in several different states of equi-

librium. Some of these equilibria may be thought of as

stable, others as metastable, others as labile, to borrow

terms from the physicist. The germ plasm of different

species may undergo parallel transformations, resulting

in parallel variations. All who have dealt with the spe-

cies of large genera know that oftentimes the same series

of variations turns up in one collective species after

another. Many characters have arisen independently, at

so many points in different lines of descent, that they

- have no phylogenetic significance whatever.

It seems to the speaker that the @nothera situation is

clearing up. More and more evidence is accumulating

which shows that although the phenomena are complex,

they are orderly. Probably no two of the workers on the

(Enothera problem look at it from the same point of view.

In this paper I have not hesitated to state freely my

present working hypotheses. Next year they may have

changed, to fit new facts. Even now there are data at

hand which do not accord with the best hypotheses I have

been able to formulate, but neither do they accord with

any others. Under the circumstances, one should not

draw conclusions of too sweeping a nature. It may con-

fidently be stated, however, that the appearance of muta-

tions in nothera is not due to Mendelian segregation,

and that the Mendelian method of attack has been utterly

fruitless. It is freely admitted that the mutation proc-

esses themselves are hardly understood at all, and that

further work must decide whether or not mutation is

always or ever conditioned by previous hybridization.

Bateson has recently described the genetical behavior

of the rogues which occur in certain varieties of peas.

Although he does not suggest that these strange forms

are mutations, his evidence would tend to convince a mu-

tationist that they are. Would it not be a strange turn of

fate if Bateson, the leader of the Mendelian school and

critic of de Vries, were destined to discover mutations of

a non-Mendelian type in the very genus which provided

Mendel with the material for his classical researches?

INHERITANCE STUDIES IN PISUM

I. INHERITANCE OF CoTYLEDON COLOR!

DR. ORLAND E. WHITE

CURATOR OF PLANT BREEDING, BROOKLYN BOTANIC GARDEN

INTRODUCTION

Mopern students of genetics such as Baur, East, Mor-

gan, Emerson, and others classify all variations in ani-

mals and plants into three general categories on the as-

sumption that organisms are made up of unit factors, in

the same way that a chemist thinks of rocks and minerals

as being composed of elements. These three categories

of variation are:

1. Variation resulting from changes in environment.

. Variation due to ‘‘loss’’ or ‘‘gain’’ of new factors

through crossing.

3. Variation due to mutation.

Tue PROBLEM

The present paper has to do largely with data on vari-

ations in Pisum belonging to the first and second cate-

gories mentioned above. An attempt is being made defi-

nitely to work out the Mendelian or factorial constitution

of the genus Pisum. with reference to all those characters

by which its few species and numerous varieties are dis-

tinguished. In order satisfactorily- to accomplish this

object, all or nearly all the known varieties of the genus

Pisum must be considered. In this paper only the inheri-

tance of cotyledon color is considered. Further papers

1 Published as Brooklyn Botanie Garden Contributions, No. 10. These

studies on the genetics of Piswm are being carried on in collaboration with

the Office of Forage Crop Investigations and the Office of Horticultural and

Pomological Investigations t of Agriculture. Based in

part on a paper given at the Twentieth Anniversary Celebration, New York

Botanical Garden, September 9, 1915,

530

No. 597] INHERITANCE STUDIES IN PISUM 531

will deal with other characters and the modifications of

various characters through crossing.

Of more than two hundred and fifty varieties and spe-

cies upon which the writer has been conducting experi-

ments, the great majority have seeds which in the mature

condition possess yellow cotyledons, but in such an array

of varieties, it was soon noticed that the shades of yellow

varied from a light greenish yellow to that of a deep

orange. Roughly one could divide these forms with yel-

low cotyledons into light and deep yellows, but any one

particularly ‘‘keen’’ on forming a series showing con-

tinuous variation, could easily grade the varieties so as

to present a series without breaks from light greenish

yellow to deep orange yellow. All the wild varieties and

species so far examined have yellow cotyledons, which

favors the assumption that yellow cotyledon is the oldest

color character. Many of the cultivated varieties and es-

pecially the so-called blue ‘‘field peas’’ such as Wisconsin

Blue and Prussian Blue and the majority of those known

as ‘‘garden peas’’ have green cotyledons when the seed

is mature. What has been stated regarding the grada-

tions of color in yellow cotyledon varieties is equally true

of those with green cotyledons. Roughly classified, there

are dark and light green forms, but the various varieties

can be arranged in a continuous series representing every

shade from very dark green to light yellowish green.

Among the numerous green and yellow cotyledon varie-

ties, when grown under the same environment, there are,

however, many varieties to which certain distinct shades

of either green or yellow are peculiar. Some varieties

have characteristically deep orange cotyledons, others

have light yellow cotyledons, and still others breed true

to the shades between these two extremes. With the group

of green cotyledon varieties, the same state of affairs

holds true. Classification of yellows and greens is still

further complicated because some varieties with light yel-

low cotyledons grade into the light greens and vice versa,

even though both are grown under the same conditions.

532

THE AMERICAN NATURALIST

[ Vou. L

The following table (Table I) gives the names of varie-

ties of greens and yellows representing the classes dark

‘VARIETIES

TABLE I

OF PISUM CLASSIFIED ACCORDING TO SHADES OF COTYLEDON COLOR

N GROWN UNDER THE SAME CONDITIONS

Yellow Cotyledons

Saraga Variety | Tireta Source

Black-Eyed Marrowfat | 14 | Vaughan Seed Co.

yellow ae of | 22 P. Henderson & Co.

etit Pois 25 P. Henderson & Co.

Späte G old 29 Haage & Schmidt

sargue s Perfection Sugar 60 Henderson & Co

147 S.P.1. 22036

Adm viral 159 S.P.I. 29323

Khaba 176 S.P.I. 20380

Yellow Mummy 1 H AE Si, dian Eng.

White Marrowfat 23 | P. Hen & Co.

mrin ea 31 | Haage T Schmi idt

Wachs Schw 32 | Haage & Schmidt

ee von Blocksberg 34 |A. D. Darbishire

omardi’ | 40 | Cambridge (Eng.) Bot. Gard.

T. potter 41 | Cambridge (Eng.) por Gard.

Ais erity | 71 | P. Henderson

Pois géant sans parchemin 107 | Vilmorin & Cie

sth Black 132 | §.P.1.,2 Dept. of Agriculture

formosum "137 |H. Winkler

ÄN aw 208 | S.P.I. 25439

Archer 209 =| S.P.I. 22037

Light oldké 30 | Agr z Pohmidi

yellow ea humile ? 33 |A. Sut

Benton 138 -4 SPE pee

Killarney 150 | S.P.I. 22078

Khauaka 1 | S.P.I. 31808

Green Cotyledons

rT Variety Bhi Source

Green to Alaska 15 Vaughan Seed Co.

light green elephone 26 P. Henderson & Co.

to yellow- Laxtonian 27 P. Henderson & Co.

ish green Acacia (wrinkled) 38 W. Bateson

Everbearing 62 Henderson & Co

(Fade Yorkshire Hero 65 Thorburn & Co.

easily) Duke of Albany 93 Sutton & Sons

Hundrediold 97 Sutton & Sons

| Alfred 140 EPE 12

| Blue Prussian 154 S.P.I. 19787

Yellowish | Acacia (wrinkled) 38 teso;

green | Duke of Albany 93 Sutton & Sons

| Hundredfold 97 Sutton & Sons

No. 597] INHERITANCE STUDIES IN PISUM 533

Dark Market Split Pea 35 mys -n ow. Markets

green | Acacia (Smooth) B06 Bi

| Velocit 59 Vaughan ‘Seed Co.

| Braunschweiger 88 ge & Schmidt

| French Grey 149 TY 27003

Rosenberg 161 S.P.I. 10274

Alaska 193 S.P.T. 29366

Scotch Beauty 198 S.P.I. 27004

Wisconsin Blue .207 S.P.I. 22049

Green | 20 | A. D. Darbishire

| s it's 's Excelsior 21 | P. Henderson & Co.

Alde 28 | Haage & Schmidt

yellow, yellow, light yellow, dark green, green, light green,

and yellowish green, when these varieties are all grown

under approximately the same conditions. Any one can

distinguish between dark green and dark yellow, but one

well acquainted with the color of cotyledons in Pisum

would have difficulties in distinguishing between light yel-

lowish greens and light yellows. The classification made

is admittedly arbitrary, though based on the same sort of

acquaintanceship with these colors as that of a nursery-

man with varietal differences in bulbs or varietal charac-

ters in leafless nursery trees. The point which it is de-

sired to emphasize by the foregoing remarks is that these

shades of cotyledon color are distinctly varietal charac-

ters, and are always characteristic of the respective vari-

eties when these varieties are all grown together under

any one of the several specific? environments in which the

pea cultures at the Brooklyn Botanic Garden have been

grown.

THe RELATION or ENVIRONMENT TO CoTYLEDON COLOR

One other perplexing factor enters into the study of

cotyledon color in Pisum—the difficulty of being certain

that all varieties under observation mature their seed

under as nearly as possible identical environments, a

factor that many geneticists experimenting with other

plant forms are prone to neglect. Some varieties of peas

28.P.1. oe for Office of Foreign Seed and Plant Introduction, U. 8.

Department of Agriculture, to which I am very much indebted for help

in Ree eg a. and species of the genus Pisum. |

3 These environments will be described in detail in a later paper.

534 THE AMERICAN NATURALIST [Vou. L

gradually change from green to yellow when maturing,

while others appear to change very suddenly, but only if

plenty of sunlight and no over-supply of moisture is

present. This is particularly true of some of the deep

orange varieties, such as Späte Gold (P 29). The seed

of this variety remains very dark green until the general

appearance of the vine leads you to suppose it is ripe, but

if plenty of sunlight is present and not too much moisture,

and the pods are allowed to remain for a few days, the

dark green changes to a very deep orange, and this deep

orange is characteristic of Späte Gold when grown com-

mercially.

With the green cotyledon varieties, one is bothered by

fading of the green to a sort of washed-out yellow in

many varieties, if the vines are not harvested at exactly

the right time. Express, Velocity and many of the

wrinkled sorts (see Hurst, 1904) are particularly subject

to change under these conditions.

The above mentioned difficulties regarding the proper

maturing of pea seed have been considered quite fully by

Bateson, Darbishire, Lock, Tschermak and other workers

in genetics. Hurst (1904) and Lock (1905) particularly

have studied the tendency of certain varieties such as

Telephone with green cotyledons to fade easily, even when

harvested carefully, but left exposed to light. Bateson

and Kilby (1905, p. 58) have studied the so-called ‘‘pie-

bald” peas (peas with green or yellow cotyledons partly

spotted or tinged with both colors) and find them to

largely result from environmental conditions such as

failure to ripen properly ar from bleaching after ripening.

‘*Piebald’’ peas are characteristic of certain varieties of

peas, in which the green fades much faster upon exposure

to light or moisture or to both than in ordinary green

cotyledon types. ‘‘Piebald’’ peas of one pod, according

to Bateson and Kilby, are always tinged on the same sur-

face. Injuries causing the death of the cotyledon tissue

(Bateson, 1905) (Tschermak, 1902) also are a cause of

yellow spots on peas from green cotyledon varieties.

No. 597] INHERITANCE STUDIES IN PISUM 535

THE PIGMENTS or CoryLepon COLOR IN PISUM

Bunyard (see Darbishire, p. 131) has shown that both

yellow and green cotyledon varieties have a yellow and a

green pigment in their cotyledons when the seed is im-

mature, but the yellow cotyledon varieties possess a fac-

tor (an enzyme perhaps), which causes the green pigment

to fade on the maturity of their seeds. Thus green pig-

ment is epistatic to yellow pigment, since, when both are

present, only the green is in evidence.

THE GENETICS OF CoTyLEDON COLOR IN PISUM

Historical

As early as 1729, according to Darwin (1876, I, p. 428)

white (yellow cotyledon) and blue (green cotyledon) peas

were found in the same pod and these results were under-

stood to be due to chance crossing. Wiegmann, Goss

(1824) and others observed that varieties of Pisum breed-

ing true to blue peas when crossed with pollen from vari-

eties breeding true to white peas, always showed a direct

and immediate effect of the pollen parent. Gärtner

(1849) and later hybridists incorrectly regarded this

phenomenon as xenia, believing that tissues of the parent

generation were affected so that the color of the seed was

changed. The fact that the change in color was due to an

embryonic character of a new hybrid generation seems

never to have occurred to them. The true significance of

these facts were never understood by Knight, Goss, Gart-

ner, nor any of the hybridizers before Mendel’s time.

Knight distinguished between cotyledon colors and seed

coat colors, and Goss and others had observed practically

everything regarding crosses between green cotyledon —

and yellow cotyledon peas except the numerical propor-

tion of one to the other in the F, generation. Darwin

(1876, p. 348) mentions some observations of Masters,

which, if authenticated, show a complex state of affairs

in the inheritance of cotyledon color, since Masters claims

to have obtained both yellow (white) and green (blue)

536 THE AMERICAN NATURALIST (Vou. L

peas from a certain pea plant and when these two kinds

were planted separately each continued to produce the

two kinds through four generations, that being as far as

the experiment was carried. In the light of the data I

present below his observations may be correct, he having

possibly secured one of the yellow forms such as I have

ound. —

Mendel (1865) found when peas with yellow cotyledons

were crossed with green cotyledon forms that the first

generation offspring all had yellow cotyledons, but each

one of these yellow cotyledon F, plants produced F,

seeds, approximately three fourths of which had yellow

cotyledons and one fourth green cotyledons. Either color

of parent could be used as the seed or female parent, and

the result was the same. Further, the F, greens in F,

only produced greens, while the F, yellows when planted,

in some cases gave only yellows, in other cases both yel-

lows and greens in the proportion of 3Y:1G. The actual

data by which Mendel supported these statements are as

follows: fifty-eight crosses on 10 plants were made, and

in every case, yellow was dominant to green in the F,

generation of these crosses. 258 F, plants produced 8,023

F, seeds of which 6,022 were yellow and 2,001 had green

cotyledons, an actual ratio of 75.1 yellow to 24.9 green or

3.01 Y:1G. Mendel is careful to call attention to the wide

variability in the ratio of yellows to greens when the F,

peas of each F, plant are considered separately, the vari-

ation ranging from 32 Y:1G on one plant to 20Y:19G

on another. Between these extremes, there were some

among the 10 F, plants of which he gives the ratios, that

closely approximated the theoretical 3:1 ratio. I call

attention to this great variability that Mendel found be-

cause some geneticists of late, apparently not having

noted that Mendel himself observed these same facts,

have referred to this as a new phenomenon. Only aver-

age ratios from large numbers were considered by Men-

del, as small numbers tended to obscure the significance

of the facts. Of the 8,023 F, seeds secured by Mendel,

No. 597] INHERITANCE STUDIES IN PISUM 537

519 seeds with yellow cotyledons were used to grow an

F, progeny. Of these, 166 F, seeds bred true or pro-

duced only seeds with yellow cotyledons, while 353 pro-

duced both yellows and greens in the proportion of 3Y:1G.

353 to 166 gives a ratio of 2.13 to 1. Mendel (p. 327)

especially calls attention to the difficulties involved in

classifying the two colors of seeds, and notes, as I have

done in the preceding paragraphs, that the seeds of pure

green varieties and of segregate greens, have a tendency

to bleach, another fact that several critics of Mendelian

methods seem to have overlooked or forgotten.

Mendel’s work has been substantiated by a large num-

ber of trained investigators, as well as by a host of teach-

ers and amateurs. The results for cotyledon color in

Pisum obtained by seven well-known geneticists are given

below (Table IT).

TABLE II

| | Percentage

Hybrid Generation Observer | Yellow Green | of Green

Second... Mendel | 6,022 2,001 24.9

Corrers | 1,394 453 24.5

Tschermak | ,080 1,190 24

Bateso: 11,903 3,903 24.7

1,310 445 25.4

Lock 1,488 514 26.2

Darbishire 1,089 354 24.9

Var) a a. Correns 1,012 344 | 25.5

Tschermak 3,000 959 24.2

ck 3,082 1,008 24.6

Darbishire 5,662 1,856 24.7

Fourth: 222), 225 70 23.7

Lock 2,400 850 26.1

Ta 56,064 42,117 13,947 24.9

These results approximate very closely the ratio of

3 Y:1G demanded by Mendel’s theory. Darbishire (1913,

p. 62) in testing out 140 F, progeny with yellow coty-

ledons, secured 98 F, plants heterozygous for green and

yellow cotyledons, and 42 breeding true or homozygous

for yellow cotyledons, a proportion of 2.3 heterozygous

F, plants to 1 homozygous F, yellow. Many varieties

gathered from all over the world were used in’ these —

4These data are taken from Darbishire (1913).

538 THE AMERICAN NATURALIST [ Vou. L

studies and all gave similar results. With these facts

before us, there can be no denying the validity of Men-

del’s law as regards inheritance of cotyledon color in

Pisum. The criticism has sometimes been made that the

F, segregate yellows and greens were a little less green

and a little less yellow owing to the association of the

unit factor materials of the two pigments in the F, genera-

tion. In other words, segregation was not complete; the

TABLE Ila

Crosses oF DOMINANT YELLOW WITH GREEN (F, GENERATION) 6

Cotyledon eo

Yellow Gree!

Pele tne ae 34

Pere Pili ee 129 41

Pot Ke ON a Be) 60 16

(eE PEAS ar he, 23 7

(PGA PIANS os sik 16 0

OR BEY ck 32 10

PE PER eee sy 34 6

PRES PL ee, 33 10

Psik POA i ere te kl 48 6

Pakre eR a 37 22

(PORE RPL Ce a 46 17

CPOE SPSS Fy o 11. 4

(PIS RIS a 65 16

(Pota PMT aa, 103 34

PESK aaa -9 4

(PIES PI-4)-8 oe 49 16

(P30-1 x P6849) ee 63 17

(POAT K PSEA oak 91 30

(P40-1 X PMS a a. 68 29

(Pat PIE 103 30

Paai PAD oo aon 81 25

(P621 SPUS oo 40 17

(P821 x Pira- oe 74 22

PEA xPHS A o 71 29

Total actually obtained ........... 64 543

a ee — owes 1,642.4 547.5

Ratio (theoretical) i206.) o.l.. 75 yellow25 green

Ratio ‘aleiatis obtained) ......... 75.2 yellow 24.8 green

® Pedigree numbers such as —1, -2, ete., following the pedigree stock num-

ber of the variety as, e. g., P28-1 refer to plant numters. P28-1, e. g.,

is progeny plant No. 1 of stock variety P28. P28-1-1 is plant No. 1 of

e second generation from pure inbred stock P28. The seed or maternal

parent in a cross is always atin first, e. g., P28-19 x P29-1 g.

No. 597] INHERITANCE STUDIES IN PISUM 539

determiner for green pigment was not able to produce as

dark a green in F, green segregates as in peas of the

green cotyledon parent race. This is true undoubtedly

in some few cases, but in still others, Hurst (1904), Darbi-

shire (1913) and myself have been unable to find any dis-

tinction in shading by comparing the segregates with the

grandparental seeds of both colors. In those cases where

there has been found a difference, the observers probably

failed to take into account all the environmental factors.

New Data

In my own investigations’ on the heredity of cotyledon

color, the F, and F, generations from over 79 crosses in-

volving combinations of 40 varieties and species of Pisum

have given results similar to those secured by other work-

ers except in the case of crosses involving a variety of

German pea, ‘‘Goldkénig,’’? obtained from Haage &

Schmidt. The data for most of these crosses are given

in Tables IIa, IIIb, IIc.

TABLE IIIb

CROSSES OF DOMINANT YELLOW AND RECESSIVE YELLOW (F, GENERATION )7

Cotyledon Color

Troeen. ; Yellowish

Yellow Green Green

(P22-3-1 XP30-A-5)=l;. omr- eee rripa enio 26 7

(P22-3-1 XP3I0-A-5)-2.... 12 ien) 18 6

(P22-3-1 XP30-A-5)-3. cui reniir ee eee 11 3

(P22-6-1 XP30-A-3)-1-.. ourse eeraa 17 5

(PIOI XPI- ien a a ee 73 16 2

(P302 XP h, ar a a a 51 5 6

(P30-2 XPI- a e 62 8 3

(P30-3 XPS82-1)--1 5 i cee Fe ee ee 52 ae 3

(P: XPI0-4)-1.. i-se ee cae ee eee ee 15 7

(Pari XP8O-B Yi ee ea eee 49 12 3

Total—actually obtained. . ... +... rse- 457 87 22

109

Total—theoretically expected ee E 459.2 yellow : 106.2 green

Ratios... ara sine Buta Weekes vee 13 yellow: 3 green

5 All os of peas have been inbred for at least two generations and

, all the ordinary precautions against differentiating environmental factors,

insect palatin, etc., in use by geneticists have been employed.

7 The investigations of Mendel, Bateson, Lock, ‘Tschermak "e

540 THE AMERICAN NATURALIST [ Vou. L

TABLE IIIc

Cross oF GREEN X RECESSIVE YELLOW (F, GENERATION)§

Yellow or Yellowish Green

2+ 1? 2

Cross

(P21-15-1 X P30—-A-2)-I1 .........

(P21-15-1 X P30—A-2)-2 ......... 4+1? 22

(P30-5—-4 x P38-20-1)-1P ........ 15+ 1?

(P30-5-4 X P38-20-1)-2P ........ 15 13 + 4?

(P30-5-4 X P38-20-1)-1 ......... 2+ 2?

(P30-5-1 X P38-20-1)-1 .......-. 7+ 1? 26

(P3541 X P30-3)-4 rA eee, 9 22

(PISI KP 30-3) 2 ao i cn Ses wes 0 20

(P35-9-1 x P30-5-4)-1 2.2... 22... 1 6

(P35-9-1 X P30-5-4)-2 ~......4.-. 29

(P35-9-1 X P30-5-4)-3 ........... 1+ 1? 8

(P35-10-2 K P30-5-6)-1 ........-. 12

(P35—10-2 X tea pines isis 1+1? 26

(P35-10-2 a P30-5-6)-3. .......... 10

ran tually obtained, 326.... 70 256

Total kiii expected . 81.5 244.5

A e a a e ciate a mms 1 yellow : 3 green

The variety ‘‘Goldkénig’’ breeds true to yellow coty-

ledons and wrinkledness. When crossed with varieties

breeding true to green cotyledons, the F, generation was

invariably green. In most of the crosses, the seed parent

was the green variety, but reciprocals have been obtained

in two cases.

The crosses were:

Goldkénig X Acacia and reciprocal

Goldkénig X Market Split Pea and reciprocal

have always shown oR eee a in peas to be inherited independently

of roundness and wrinkledness of potas The data given in Tables

IIIb and IIIc give reason nh believe there is linkage or partial coupling

involved. Round yellow X ee yellow U. gives practically

only three classes in F,—round yellow , wrinkled yellow, and round

n. All four classes Torm in only one cr on: where a single

wrinkled green was obtained. Round green X nkled vee (Gold-

gave all the expected classes except round ese which was absent

m the F, progeny of all the eight crosses examined. Wrinkled yellow

(Goldkénig) X wrinkled green and reciprocal Sony wrinkled yellows and

wrinkled greens approximating the expected ra

8 The cotyledon colors of the peas conce et in Takles IIIb and Ile

and these a when compared, differed but slightly, and only in

very few cases. r help in these determinations, I am indebted to Mr.

Montague Free of bel Brooklyn Botanic Garden staff

No. 597] INHERITANCE STUDIES IN PISUM 541

Nott’s Excelsior X Goldkénig

Aldermann Xx Goldkénig

Scotch Beauty X Goldkénig

Sutton’s Main Crop X Goldkénig

An F, generation has been grown from the first three

of these crosses with the results (see Table IIIc) ap-

proximating a ratio of 3G:1Y or the reverse of the com-

mon result. Practically all of the green seeds are distinct

greens, but among those classed as yellows are several

doubtful cases, and these are marked questionable. An

F, generation is being grown which will decide whether

I have erred in considering these doubtful cases as yellow

cotyledon peas in which the greenish color results pos-

sibly from lack of enough sunlight during the ripening

‘period.

When the ‘‘Goldkénig’’ yellow was crossed with other .

varieties having yellow cotyledons, the F, progeny all had

yellow cotyledons, but in the F, generation, a certain pro-

portion of peas with distinctly green cotyledons appeared,

the F, ratio in the progeny of the ten different F, plants,

showing considerable variation, but averaging 13 yellow

seeds to 3 green seeds, provided all yellows having any

considerable amount of green pigment are classified as

greens. The number of F, generation progeny obtained

was small, totaling only 566, of which 457 had yellow

cotyledons, 87 distinctly green cotyledons and 22 seeds

had yellowish green cotyledons. All were grown under

conditions insuring their maturity, but under these con-

ditions (greenhouse cultures) the amount of moisture

present is such as possibly to cause some of the true

greens to bleach. Further justification for classifying

the 22 doubtful greens as true greens comes from the fact

that all the varieties having yellow cotyledons used in

these crosses with the exception of Goldkénig are varie-

ties with distinctly bright yellow cotyledons, which grown

under the same conditions and often side by side with the

crosses, never show any green coloring matter in the coty-

542 THE AMERICAN NATURALIST (Vou. L

ledons of their mature seeds. The crosses of dominant

yellow with Goldkénig were:

“Pisum Jomardi’’ X Goldkénig

Goldkénig X ‘‘Mummy Pea’’

Goldkonig Xx Wachs Schwert

‘‘ Pisum elatius” X Goldkénig

First of All X Goldkönig

Gold von Blöcksberg X Goldkénig

Späte Gold x Goldkénig

Benton X Goldkönig

All the yellows other than Goldkénig yellow gave the

ordinary Mendelian ratios when crossed with varieties

having green cotyledons. No greens were obtained from

crosses between varieties having yellow cotyledons, other |

than those with Goldkénig.

THEORETICAL INTERPRETATION

Interpreted in Mendelian terms the above data are

brought into accord with other data on the inheritance of

cotyledon color in Pisum by regarding all varieties of

peas, both with yellow and with green cotyledons, as pos-

sessing a factor for yellow pigment (Y), while the domi-

nant yellow varieties possess a factor for green pigment

(G) and a factor (I) which causes the green pigment to

fade on the maturity of the seed. The German variety

_ “Goldkénig’’ may be regarded as lacking both the factor

for causing green pigment and the factor for causing that

pigment to fade on the maturity of the seed, while the

green varieties lack only the factor (I). Green pigment

masks yellow pigment, hence may be regarded as epi-

static to yellow pigment.

Regarded thus:

(1) YYGGII—dominant yellow varieties

(2) YY¥ggii—recessive yellow varieties

(3) YYGGii— green varieties

No. 597] INHERITANCE STUDIES IN PISUM 543

Crossed with each other these give:

(1x 2) =YYGgli (F,) yellow, (F) 13Y:3G4

(1x3) =YYGGIi (F,) yellow, (F,) 3Y:1G

(2X3) = YYGgii (F,) green, (F,) 1Y:3G4

The hereditary substances responsible for yellow pig-

ment, of course, have not been isolated and may take the

orm of more than one factor, but I have represented

these as Y, to make my interpretation clearer. The essen-

tial point in the interpretation is that all the hereditary

differences in cotyledon color in Pisum so far discovered

may be pictured as due to the presence or absence of tw:

genetic factors.

TABLE IV

FACTORIAL COMPOSITION OF F, PLANTS OF THE THREE CROSSES AND THE

APPEARANCE OF THE F, PROGENY

Dominant Yellow X Recessive Yellow and Reciprocals

F2 Ratio 13 ¥ °3G Character of Fs Prog

T YOGA ss Breeds true to yellow saiid

oY GOI 2 x oy2 0

9 yellow

2 YY Gelt snes Breeds true to yellow but heterozygous for G

SY GQ S40 o.; 13 S:a

3 greon a YYOGI ose Breeds true to green cotyledons

2 VIGI eras Ti pG

bY 2 oe ies Breeds true to yellow cotyledons

3 yellow ;

eG e's LE en Breds true to yellow but heterozygous for I

l yolow 1 YYggu a Breeds true to yellow cotyledon:

Dominant Yellow X Green and Reciprocals

F: Ratio 3Y:1G Character of Fs Progen

1 YOGGIE as Breeds true to yellow setae

3 yellow :

ZXGGU -2 1:16

1 grees 1 YYGGH . Breeds true to green cotyledons

Recessive Yellow X Green and nna

3G Charac' :

1 Solow ; YYggii a Breeds true y péra le

3 YOGE ce: Breeds true to green cotyledons

Oe eave 1¥:36

Regarded thus, the F, plants of all crosses so far made

in Pisum, involving cotyledon color, can be represented

by the gametic formule given in Table IV. The char-

544 THE AMERICAN NATURALIST [Vou. L

acter of the F, progeny, providing this interpretation of

the facts regarding cotyledon color in Pisum holds, is

also indicated in this table.

Additional data on the inheritance of cotyledon color

in Pisum will be given in a succeeding paper.

CONCLUSIONS AND SUMMARY

Variation in cotyledon color in Piswm belongs to all

three of the categories of variation mentioned in the fore-

part of this paper, although there are no definite data as

regards the origin of the green cotyledon and the ‘‘re-

cessive’’ yellow cotyledon varieties.

1. Variations in cotyledon color due to environment -

re:

(a) Yellow cotyledon varieties producing seeds with

green cotyledons, because of immaturity, absence of suf-

ficient sunlight, excess moisture at the period of ripening

of the seed, ete.

~ (b) Green cotyledon varieties, especially those with

wrinkled seeds, producing seeds which fade or bleach to

yellow or yellowish green owing to excess of moisture and

sunlight after the seed has matured.

2. Variations due to innate or hereditary differences

probably arising as mutations are:

(a) Different degrees or intensities of yellow and green

coloring in the different varieties of Pisum. These differ-

ent intensities are characteristic of particular varieties

when all varieties under consideration are grown under

approximately the same environment.

3. Hereditary distinctions as regards cotyledon color

in Pisum may be represented by the presence and absence

of two factors, a factor (I) causing green pigment to

fade when the variety matures its seed, and a factor (G)

causing the production of green pigment. All varieties

of Pisum so far experimented with, have yellow pigment

in their cotyledons and the determiner or determiners re-

sponsible for this pigment may be graphically repre-

sented by (Y). As the presence of green pigment masks

No. 597] INHERITANCE STUDIES IN PISUM 545

yellow pigment, green may be regarded as epistatic to

yellow.

4. The majority of varieties with yellow cotyledons

when crossed with varieties having green cotyledons,

have yellow cotyledon F, offspring, the F, generation

breaking up into yellow and green cotyledon plants in the

ratio of 3Y:1G.

The yellow cotyledon variety ‘‘Goldkénig’’ when

crossed with green cotyledon varieties has green coty-

ledon F, offspring, in F, giving a ratio of approximately

LY:3 G just the reverse of the ordinary result.

‘‘Goldkonig’’ crossed with other varieties having yel-

low cotyledons has yellow cotyledon F, offspring, in F,

giving a ratio of approximately 13 yellow seeds: 3 green

seeds.

= With these facts in view, ‘‘dominant yellows’’ may be

represented by the formula YYGGILI, ‘‘recessive yellows”?

(Goldkénig) by the formula YYggii, and green cotyledon

forms by the formula YYGGii. These formule account

for all the facts so far discovered in experiments on the

inheritance of cotyledon color in Pisum, except the data

on linkage or coupling, referred to in page 539, note 7.

These results will be discussed when more data are

available.

LITERATURE CITED

1. Bateson, W., and E. R. Saunders, and others

1902-1908. T to the Evolution Coniaities of the Royal So-

cie .

2. Bateson, w.

1909. Mendel’s Principles of Heredity. E (Eng.), Uni-

versity Press, pp. ix + 396. Fig. 37. Pl.

. Baur, E.

pa Einführung in die experimentelle Vererbungslehre. G. Born-

träger, Berlin. S. 293, Fig. 80. Tafn. 9.

4. Castle, W. ;

1912. The Inconstancy of Unit-characters. AMER. Nar., 46, pp. 352-

. Collins, ze eg and Kempton, J. H.

1911. Tulevitaste of Waxy Endosperm in Hybrids of Chinese

: Maize. IV° Conf. Internat. de Génétique, pp. 547-557.

. Correns, C.

1900. Gregor Mendel’s Regel über das Verhalten der Nachkom-

menschaft der Rassenbastarde. Ber. d. d. M Gesell. 18, S.

158-168.

546 THE AMERICAN NATURALIST [Vou. L

1902. ka den Modus und den Zeitpunkt der Spaltung der An-

a aA bei den Bastarden vom Erbsen-Typus. Bot. Zeitung 60,

SIA p. 66.

8. Astute si

1913. bresain and the Mendelian Discovery. Cassell & Co., Ltd.,

ork. Pp. vi+ 282. Fig. 34. Pl. VI.

9. Darwin, g

1876. Animals and Plants apa riaa DE 2 vols. 2d ed.

pleton & Co., New

10. East, E. M.

1912. The TOK Notation as a Description of Physiological

Facts. AM AT., 46, pp. 633-655.

11. Gärtner, C. F.

1849. Bastarderzeugung.

12. Goss, J.

1824. On the Variation in the Color of Peas, occasioned by Cross

Impregnation. Trans. Hort. Soc., 5, pp. 234-236

13. Hurst, C. C.

1904. n aa in the Heredity of Peas. Journ. Roy. Hort.

oc., 28, pp. 483-494.

14. Johannsen, W.

1911. The Genotype Concéption of Heredity. AMER. NAT., 45, pp-

129-159.

15. i

1913. Elemente der exakten Erblichkeitslehre. Zweite auflage.

Jena. S. vii + 723

16. Knight, T. A.

1799. Philosophical Trans., V, pp. 195-204.

17. Lock, R. H.

1905. Studies in Plant-breeding in the Tropics, II. Annals Roy.

Bot. Garden, Peradeniya, 2, pp. 357-414.

18.

1908. The Present State of Knowledge of Heredity in Pisum.

Ibid., 4, pp. 93-111.

19.

1906. Recent Progress in the Study of Variation, Heredity and

Evolution. J. Murray, London. Pp. x + 334.

20. Mendel, G. J.

rünn, Bd. 4, S. 3—47.

1865. Versuche über Pflanzen Hybriden. Verh. Naturf. Ver. in

. Sturtevant, A. H.

1915

e Behavior of the Tea as Studied through Link-

age. Zeitschr. f. ind. Abstamm. u. Vererbungs., Bd. XIII,

Heft 3/4, S. 234-287. Taba. ra

. Sutton, A. W.

1911. Experiments in Crossing a Wild Pea from Palestine with

Commercial Peas with the Object of Tracing any Specific

Identity. between tħis Wild Pea and the Peas of Commerce.

IV° Conf. Internat. de Génétique, Paris, pp. 365-67.

No. 597] INHERITANCE STUDIES IN PISUM 547

23. Swingle, W. T.

1911. a in First Generation Hybrids (Imperfect Domi-

ance): Its Possible Explanation through Zygotaxis. IV°

Cont. Lite dad. de Génétique, Paris, pp. 381-394.

24, Tschermak, E.

1900. Über künstliche Kreuzung bei Pisum sativum. Zts. f. d.

landw. Versuchsw. in Oesterr. III. Jahrgang, Heft 5.

1901. hiiest oe uber eae der Merkmale

bei Kreuzung von pen n und Bohnen. Ber. Deutsch. Bot.

Geselisch: XIX, Hef

1901. Weitere Beitriige iiber Verschiedenwertigkeit der Merkmale

bei Kreuzung von Erbsen und Bohnen. Zts. f. d. landw.

Versuchsw. in Oesterr. IV. Jahrgang, Heft 6.

27.

1903. Die Theorie der Kryptomerie und des Kryptohybridismus.

Beihefte z. Bot. Centralbl., XVI, Heft 1.

1911. Examen de la théorie des facteurs par le recroisement

mét baat des hybrides. IV° Conf. Internat. de Génétique,

Paris, pp. 91-95. Tab. I-VIIIc.

29. Vilmorin, Pailin

1911 Ive Conf. Internat. de Génétique, p. 51.

30. Weldon, x T R.

1902. Mendel’s Laws of Alternative ser a in Peas. Bio-

peels I, Part 2, pp. 228-254,

31. hie Orland

The la of Teratological Development in Nicotiana on

Theories of Heredity. Amer, NAT., 47, pp. 206-228.

32. ;

1914. Swingle on Variation in F, Citrus Hykrids and the Theory

of Zygotaxis. AMER. NAT., 48, pp. 185-192. `

THE RESULTS OF FURTHER BREEDING

EXPERIMENTS WITH PETUNIA

EDITH R. SAUNDERS

LECTURER, .LATE FELLOW, NEWNHAM COLLEGE, CAMBRIDGE, ENGLAND

In view of the present discussion on the inheritance of

doubleness in Petunia! and pending further investiga-

tion of the evidence in favor of an explanation based

wholly or in part upon selective sterility such as has been |

described by Geerts? for Gnothera Lamarckiana, by

Belling’? for the Velvet ‘‘Bean’’ (Stizolobium deeringt-

anum and other species) and by East for species of Nico-

tiana, it seems desirable to make available the further

results of the breeding experiments which have been car-

ried out since the publication of the earlier data in 1910.°

The two main facts established by the earlier work were

(1) That singles of the cultivated forms of P. violacea,

P. nyctaginiflora and of various garden strains

(Countess of Ellesmere, hybrida grandiflora and

others) give only singles when self-fertilized or

crossed with pollen of other singles.

ae That these same singles give a mixture of singles and

doubles in F, when crossed with the pollen of a

double.

The later experiments carried out on the same lines and

with the same kind of material have considerably widened

1 Frost, ‘‘The Inheritance of Doubleness in Matthiola and Petunia,’’

Am. Nart., Vol. XLIX, ae MK p. 623, Oct., 1915; also Saunders, ‘‘Selec-

tive Partial Sterility as an Explanation of the Behaviour of the Double-

throwing Stock and the Polna ’? Ibid., Vol. L, No. 596.

2 i“ Beiträge zur Kenntniss der Gytologis und der partiellen Sterilitat von

nothera Lamarckiana,’’ Recueil des Trav. Bot. Néerl., Vol. 5, 1909.

3‘íThe Mode of Inheritance of Seanbaterttity is in the Offspring of Certain

Hybrid re Zeitsch. f. ind. Abst. u. Vererbungslehre, Bd. XII, Heft

5, p. 303, ;

+The Phenomenon of Self-sterility,’’ Am. Nart., Vol. XLIX, No. 578,

p. 77, Feb.,

‘pares pes in the Inheritance of Doubleness in Flowers,’’ i.

Petunia. J. of Genetics, Vol. I, No. 1, p. 57, 1910

548

No. 597] EXPERIMENTS WITH PETUNIA 549

the basis upon which these generalizations rest; while the

fact that mixed F, families have now been obtained in the

case of a new wild form as well as with one nyctagini-

flora individual raised from wild seed, though not prov-

ing that the second statement would invariably hold good

for singles of each of ‘these two species, certainly in-

creases the probability that this may be found to be the

case.

With respect to these more recent experiments I was

able in 1911 to obtain a larger series of counts than had

been possible heretofore. For the opportunity to carry

out the work on this larger scale I was much indebted to

Professor Bateson, who kindly had some 5,000 plants

grown for me at the John Innes Horticultural Institution.

The results may be summarized shortly as follows:

Seven crossbred singles out of matings in which the

four strains mentioned above were variously combined,

were self-fertilized. Only singles were obtained, viz.,

1,200, 123, 89, 73, 33, 32, and 12 in the different families.

Total 1,562 singles.

Fifteen singles were tested by crossing and pollen from

8 doubles was used to fertilize them. The single parents

included

One violacea plant (commercial material, new stock).

Twelve F, plants the offspring of 4 singles (Countess

of Ellesmere) which had been crossed with pollen

from various doubles (hybrida grandiflora). Among

these 12 were 4 of those which had yielded all-single

families when self-fertilized.

One hybrida grandiflora plant derived from a mating

between two singles each of which was the offspring

_ of a single crossed with pollen from a double.

One F, plant derived from an F, single out of the mating

nyctaginiflora X hybrida Grier (double), the

F, plant having been crossed back with another

double.

Thus doubleness was known to have been introduced

into the pedigree only once in the case of the first-men-

tioned (violacea) come chs thrice in the case of the

550 THE AMERICAN NATURALIST [ Von. I

last (F,) plant. In the remaining 13 cases it was intro-

duced twice,—in consecutive generations in the case of

the group of 12 plants, with the skipping of a generation

in the remaining instance. Families were raised from

each of these 15 plants and doubles occurred in them all.

The numbers obtained, as was to be expected from the

more extended scale of the experiments, indicate a more

uniform proportion of singles and doubles than was ap-

parent in the earlier results. In many families the num-

bers clearly suggest a ratio of equality; in a few, how-

ever, there was a considerable excess on the side either of

the singles or of the doubles. No connection could be

traced between the proportion of doubles obtained and

the number of times doubleness was introduced into the

pedigree. The numbers in each family are shown below

where the 15 seed-parents are indicated by the capital

letters A to O and the pollen parents by the small letters

a to h.

Parents | X| xX} K XI x] XK] xX] KI xT KL KY] x] xX] x] x

<i A} mi alo A A fj & O| O | O

Fı |

O a 27 (104| 74 | 19 | 27 | 99. 58 | 53 250.237|148 103 113

Douek 36 | 63! 67 | 47 | 22 |100| 56 | 51 (216 25612071 42 | 65) 63/106

S “| 3 a | = | to o “ x x ai.

Parents RA MPR ee X x| x

o|o|m|sle|Mlalalslzlolo; ™™"

E a a ce eat vic 25 |188 23 | 58 | |15 34 | 20 | 25 | 38 | 51 | 52 | 17 1,999

Doubhist. . etosan 40 (101| 25 78 | 14 | 29 | 17 | 30 | 13 | 39 | 39 | 15 1,837

How far these proportions are determined, as sug-

gested by Frost, by a condition of partial selective ster-

ility can hardly be profitably discussed until further

microscopic investigations have been made, but that this

explanation forms a part, though possibly not the whole,

of the explanation appears highly probable. At the time

it seemed advisable to postpone further breeding until

wild material was available for comparison. Though

repeated efforts have been made to obtain seed of wild

plants, they have, in the case of P. violacea, been so, far

quite unsuccessful. In the course of 1912 ad 1913 how-

6 Loc. cit.

ever, through the assistance of the authorities at Kew

and of Sir Reginald Tower in Argentina, to whom I am

much indebted, seed was obtained of wild plants of P.

nyctaginiflora and also of two new unnamed wild forms

(species, both white-flowered), all of which had been col-

lected most kindly by M. Thays, director of the Botanic

Garden at Buenos Aires. An entirely new stock of

double material differing in nature as well as in origin

from that previously employed was also now available.

This was raised from

(1) seed of a plant exhibited at the Conference on Genet-

ics held in Paris in 1911, which had some flowers

double and others apparently of a normal single

structure. This plant had appeared in the grounds

of the establishment of MM. Vilmorin-Andrieux et

Cie at Verriéres-le-Buisson. Some seed harvested

from the single flowers, together with a small

quantity obtained from the doubles was later most

courteously forwarded to me by Dr. Haagedoorn;

(2) seed sent to me by Mrs. Francis of Ventura, Cali-

fornia, of an interesting new strain of seed- prome.

ing doubles which she had succeeded in raising.’

The South American seed samples gave plants of uni-

form type in the case of the two new forms, each presum-

ably being a distinct species. One of the two (referred to

below as P.x.), the seed of which had been collected in

Cordova, was crossed with pollen from a double raised

from the seed of one of the single flowers on Haagedoorn’s

half-and-half plant, and also with two seed-giving doubles

of the Ventura strain. Each F, family showed a mixture

of singles and doubles. In the case of the former cross a

large number of semi-double plants with the supernu-

merary petaloid structures small or few in number, were

also obtained. The numbers recorded are given on page

These 5 seed-parents as well as other individuals tested

proved quite self fertile and set a good quantity of seed

when Boit pona

7 For an account of this strain see Mrs. Myrtle Francis [Shepherd] on

‘“‘ Doutle Seeding Petunias,’’ J. of Heredity, Oct., 1915

552 THE AMERICAN NATURALIST [Vow. L

Single | Double Parent | Fi

Flowers Magenta

Flowers Flowers Magenta AA ; | we

a | | Singles Poses | Doubles

| |

P.x. 1... . Descendant of Haagedoorn’s plant 15 54 | 23

F32 -| $ er aE N E 12 32 15

Ja S.. | A pon 5 Baiat

B a a OO dea Eon a

Flowers Rose Pink | Flowers White with Dark Tube

PRE ‘Ventura ARI SOE OE ore tess ass | 14 pis dai | 15

Peb o T a T E I E es | Doo oe 2

| |

aa aa aa e eee ee

Thus we have evidence that in a third species (for as

such there seems good ground for regarding this wild

type) doubles appear in F, when a cross is made with a

double form. The behavior in this respect of the other

new form has not yet been ascertained.

With regard to P. nyctaginiflora the seed of which was

collected from Punta Ballena, Maldonaldo (Uruguay), it

was noticed that the plants were not entirely uniform in

color, some showing a very definite tinge of purple on

the outer side of the flower tube, others scarcely a trace.

Similar variations had also been observed in the original

(1906) commercial material. Whether this variability is

normal to the species or is an indication of crossing is

not certain. It is somewhat remarkable that in this

original material all the 8 individuals tested proved to be

self fertile and yielded abundance of seed, whilst six

self-pollinated flowers on three of the Uruguay plants

taken at random did not yield a single seed. It seems un-

likely that this result could be due to accident, or to a

difference of conditions due to the fact that the 1906

plants were grown in the open, whereas the individuals

grown from the wild seed were kept in pots in a cool

house. Indoor treatment had not been found to affect the

fertility of the other strains, and it is hardly likely that it

did so here; but this point is now being verified. Only

one Uruguay plant was crossed with pollen from a double

(Ventura 2). From this ecross:the same result was ob- —

No. 597] EXPERIMENTS WITH PETUNIA 558

tained as with the commercial material. F, was mixed,

the numbers recorded being 225 singles and 113 doubles.

So far, then, the material employed has furnished no

exception to the statement that singles crossed with the

pollen of doubles yield some doubles in F, though breed-

ıng true to singleness when self-fertilized or pollinated by

other singles. i

Unfortunately very little evidence is yet available as to

the results of using the double plant as the seed-parent.

Haagedoorn’s plant was of an exceptional character and

the Ventura plants, though more typical, showed con-

siderable sterility. Only 5 plants were raised from the

seed sent by Mrs. Francis. These were all double. Only

one individual was obtained by self-fertilization of these

plants and this was also double. A cross with a single

(Countess of Ellesmere) produced only one offspring and

this plant was lost before flowering. Itis however hoped

by a repetition of this mating to obtain an F, generation

which will throw further light on the relation of the

double to the single.

The following information received from Professor

Bateson concerning a cross made at the John Innes

Horticultural Institution by E. J. Allard unfortunately

only came to hand after the above account had been for-

warded for publication.. In the mating in question a

nyctaginiflora plant, one of a batch raised from a sample

of the same wild seed as that from which my own plants

were grown, was crossed with pollen from a pink-flowered

double, a florist’s highly cultivated type of plant. More

than 20 plants were raised in F, and all were single. If

we consider this number large enough to be taken as con-

clusive we should have in this experiment an exception to

the results hitherto obtained and summed up in the gen-

eral statement formulated above. We must then suppose

that there exist in these Petunia forms certain singles and —

doubles the relation between which is such that double- ;

ness completely disappears in the F, generation obtained

from a cross between them. | :

INHERITANCE OF SEX IN THE GRAPE!

W. D. VALLEAU

SECTION OF FRUIT BREEDING, UNIVERSITY OF MINNESOTA EXPERIMENT

STATION, ST. PAUL, MINNESOTA

Since the discovery of Correns in 1907 that in Bryonia

the staminate plants produce two kinds of gametes with

respect to sex, and the pistillate and hermaphrodites only

one, great advances have been made in the study of sex

inheritance.

Shull (1910, ’11, 14) has shown that in Lychnis dioica

also the staminate plants are heterozygous for the sex

genes, while the pistillate ones are homozygous, but that

the hermaphrodites are heterozygous for the determiner

for femaleness and for the hermaphrodite condition.

These hermaphrodites were apparently developed from

staminate plants as they bear only partially developed

pistils in many flowers, and further, the factor for narrow

leaves is linked with the determiner for the hermaphro-

ditic condition, while in the normal males it is linked with

that for maleness, as pointed out by Shull. Apparently

the determiner for femaleness is carried suppressed and

linked with the determiner for maleness in the staminate

plants. ;

A simpler case of sex inheritance than either of the

above is that of the sweet pea in which Bateson has shown

that genotypically three kinds of plants may be produced;

namely, the normal hermaphrodites, which produce only

hermaphrodites when selfed, the pistillates bearing conta-

bescent anthers, which when pollinated with pollen from

the normal hermaphrodites produce the third type, which

is phenotypically the same as the normal hermaphrodites.

These when selfed produce hermaphrodites and pistillate

plants in a 3:1 ratio, showing them to be heterozygous for

1 Presented before the Society for Horticultural Science, Ohio State Uni-

versity, December, 1915.

554

No.597] INHERITANCE OF SEX IN THE GRAPE 555

the hermaphrodite and female determiners, and showing

further that in the sweet pea only one dose of maleness

is necessary for the production of functional stamens.

Similarly in many animals, it has been proved by

Wilson, Morgan and others that the males are hetero-

zygous for the sex determiner and the females homo-

zygous, and that this condition is correlated with the pres-

ence of two chromosomes in the female which are distinct

from the others and which they called the ‘‘X’’ bodies,

while in the male only one may be present, or if there are

two, the second is sometimes smaller and is spoken of as

the ‘‘Y’’ body. Occasionally the X and Y bodies may be

of equal size and the supposition that they are different

is based upon inheritance studies.

The reverse of the above mentioned condition may

exist, as in Abrazis, pigeons, cultivated fowl, etc., in

which the males are apparently homozygous for the sex

determiner and the females heterozygous.

Strasburger found that in Bryonia there were two

chromosomes which were larger than the others and

thought that these might carry the determiners for sex.

A great deal of evidence has recently been collected

which points to the chromosomes as being the carriers of

factors and as many factors have been shown to be linked

with sex, it seems safe to conclude that certain chromo-

somes carry the determiner for sex. If this is the case,

then in the hermaphroditic plants it must be assumed

that the determiners for maleness and femaleness are

linked or carried in the same chromosome; otherwise

there would continually be produced not only hermaph-

rodites but staminate and pistillate plants as well.

The trend of development in many plant groups seems

to be toward the production of a diecious condition by the

suppression of the stamens in one set of individuals and

of the pistils in another. As an instance of this may be

cited the strawberry, in which staminate, pistillate and

perfect flowers are produced. The grape and maple and

many other plants show a like suppression ın varying

degrees.

556 THE AMERICAN NATURALIST (Vou. L

Because of the number of flower types, and the fact

that most of our cultivated grapes are only one or two

generations from the wild, they would seem to furnish

ideal material for the study of sex inheritance.

Two types of vines are found in the wild, those produc-

ing functionally pistillate flowers, but bearing reflexed

non-functional stamens, and those producing functionally

staminate flowers, but bearing suppressed pistils.

We have then apparently a transitional form, in the

ease of the grape, from a hermaphroditic condition such

as is found in the apple, in which the male and female

determiners are apparently linked, to the strictly diecious

forms, as ashes, willows, ete., in which the determiner for

maleness is completely suppressed in the sex chromosome

bearing the determiner for femaleness, and the female

determiner is completely suppressed in the chromosome

bearing the factor for maleness.

On this hypothesis we would assume that in the func-

tionally pistillate grape flowers the suppression of male-

ness has begun and evinces itself in the production of

reflexed stamens bearing non-functional pollen, i. e., lack-

ing germ pores (Dorsey, 1913) and containing degenerate

generative and vegetative nuclei embedded in apparently

normal cytoplasm? (Gard, 1913). The period of degen-

eration of the nuclei is not at all definite.. Rarely, the

microspore nucleus does not divide. In some cases de-

generation takes place directly following the microspore

division, in others one nucleus only will degenerate at this

time, and in still other cases the two nuclei will appear

normal at the time of dehiscence (Dorsey, 1913). Beach

(1899), Booth (1902) and Hedrick and Anthony (1915)

have shown from pollination and germination tests that

occasionally a few pollen grains borne in reflexed stamens

are entirely functional. There is an apparent lack of

suppression of maleness, occasionally, which allows tie

development of these normal grains.

Similarly it might be assumed that in the staminate

2 Pollen of this type should not be confused with abortive pollen ss

is often produced in hybrids.

No.597] INHERITANCE OF SEX IN THE GRAPE 56T

flowers suppression of the female determiner has taken

place. Booth (1902) and Dorsey (1912) have shown that

in practically all staminate grape flowers suppressed

pistils are found. In some cases under cultivation and in

rare instances in the wild state, the suppression of pistils

is less marked and fairly well developed to perfectly de-

veloped pistils are formed. On individual plants occa-

sionally all gradations from staminate to functionally

hermaphroditic flowers are found.

A third type found under cultivation but which is ex-

tremely rare in the wild, is the functional hermaphrodite

bearing all hermaphroditic flowers. A discussion regard-

ing the probable origin of this type will be taken up later.

Although breeding work has been carried on for the

past twenty-five or thirty years in this country with the

grape, apparently little attention has been given to the

inheritance of the various flower types, although a knowl-

edge of sex inheritance would be of much value to the

breeder. In 1914 Anthony published valuable data on sex

inheritance in the grape, but gave no satisfactory inter-

pretation of the results. In 1915 the data again appeared

in more detail (Hedrick and Anthony, 1915), and, as no

further attempt was made to interpret the results, the

writer wishes to present the following as a probable ex-

planation of sex inheritance in the grape, or at least as a

working hypothesis for the interpretation of further

results which may be obtained.

For the reason that in diecious plants there are appar-

ently definite determiners for maleness as well as female-

ness, while in animals it is supposed that males are pro-

duced when only one dose of the sex determiner is

present, while females are produced if two doses are

present, it seems well to use different symbols to designate

the sex determiners of plants from those of animals.

Therefore, those suggested by Shull (1914, p. 293) in

which “the female is assumed to be a neutral homozy-

gote,’’ will be used in the following discussion, namely

FF to represent a female and FM to represent a male.

558 THE AMERICAN NATURALIST [Von. L

The hermaphrodites would then be designated as FFM

or simply as FH. It will be seen from the following dis-

cussion that the only formulation which will meet the

conditions is that which assumes the female to be a neutral

homozygote.

In the above mentioned paper on the ‘‘Inheritance of

Certain Characters of Grapes’’ (Hedrick and Anthony,

1915) the authors have concluded that the results ob-

tained on inheritance of sex do not conform to the ex-

planation of sex inheritance dependent on one sex being

considered heterozygous and the other homozygous for

sex determiners. It appears, however, that by using the

hypothesis of partial suppression of sex determiners, the

condition in the grape would be in accordance with the

assumption of a homozygous condition for femaleness in

the functional females, a heterozygous condition for male-

ness and femaleness in the functionally male plants, and

a heterozygous condition for femaleness and hermaph-

roditeness in some of the hermaphrodites, while others

would be homozygous for the hermaphrodite determiners.

The authors based their conclusions upon the supposition

that the hermaphrodites bearing upright and those bear-

ing reflexed stamens were of a single type genetically, and

produced only hermaphrodites and no females when

crossed. This assumption seems erroneous.

Using the formule suggested above, let us apply them

to the data given by the authors, which are as follows:

U X Us =180U + 47R RX R? = 16U+16R

U selfed = 673 U + 152 R R selfed=— 94 U + 73'R

Uselfed= 18U+ OR :

Total. is; 871U + 199R Total.: 110 U + 89 R

Rato... 43U: IR Erto... IZU: oR

R X U207U + 206R

Rátio... 1U: iR

UXR?

Hermaphrodite female X pure male = 56 hermaphrodites + 51 males.

$**The pollen parent is always placed last.’’ ‘‘U’? refers to hermaphro-

dites bearing upright stamens which are usually functional. R refers to

hermaphrodites bearing reflexed stamens which rarely produce functional

pollen.

No.597] INHERITANCE OF SEX IN THE GRAPE 559

In the following discussion the term ‘‘female’’ will

refer to the plants bearing morphologically perfect

flowers but having reflexed stamens, the pollen of which

is not functional. ‘‘Hermaphrodite’’ refers to those

bearing perfect flowers, the stamens of which are upright

and produce functional pollen. ‘‘Male’’ refers to those

plants bearing staminate flowers.

The expectation from the cross ‘‘U x U” (hermaph-

rodite FH X hermaphrodite FH) would be 3 hermaph-

rodites: 1 female (FF); the hermaphrodites being of

two types, viz, 2 FH:1 HH. This ratio is very closely

met in the figures 180 upright and 47 reflexed. ‘‘U

selfed’’ should give the same proportions and these are

closely approached in the figures 673 upright and 152 re-

flexed. This assumes the production of homozygous

hermaphrodites (HH), which, when either selfed or

crossed with other types, produce only hermaphrodites.

Apparently the two hermaphrodites which produced 18

hermaphroditic seedlings only, when selfed, are of this

genetic constitution. At the Minnesota Experiment Sta-

tion four hundred seedlings of Beta, open to cross pollina-

tion, produced only hermaphroditic flowers; indicating

that Beta must be homozygous for the hermaphrodite

determiners.*

The crosses ‘‘R X R” and ‘‘R selfed’’ (female FF x

female FF), producing both hermaphrodites and females,

might be explained on the hypothesis already given, v1z.,

that of partial suppression of the determiner for male-

ness in the chromosome bearing the determiner for

femaleness. i

It has already been pointed out that in the pistillate

flowers bearing reflexed stamens, a series of pollen condi-

tions, ranging from those in which the microspore nucleus

does not divide, through those in which the generative

nucleus aborts directly after the microspore division, to

those in which a few normal functional pollen grains are

4In Lychnis dioica Shull has shown that homozygous hermaphrodites are

never produced.

560 THE AMERICAN NATURALIST [Von. L

produced, has been found. This very evidently shows

that variation in the amount of suppression of the de-

terminer for maleness takes place in the determiner for

sex of these normal grains. This normal pollen when

used to pollinate pistillate flowers should give, in some

cases, females bearing reflexed stamens and in others

hermaphrodites, depending upon the extent to which sup-

pression of maleness is lacking in the chromosome bear-

ing the sex determiners of these normal pollen grains.

From the cross “R X U” (female FF X hermaphrodite

HF) should be expected females (FF) and hermaphro-

dites (HF) in the proportion of 1:1. This ratio is met

exactly in the cross ‘‘R X U’’—207 upright and 206

reflexed.

The following analysis, kindly furnished me by Mr.

Anthony of the New York State Agricultural Experiment

Station, of the cross ‘‘hermaphrodite female X pure

male’’ which produced ‘‘56 hermaphrodites + 51 males,’’

shows that both hermaphrodites and females were used as

the female parent and that three kinds of males were

used, namely, wild males, males one generation from the

wild and ‘‘intermediates’’ (males bearing occasionally a

few well developed pistils).

Upright! Reflexed | Males

Hermaphrodite X wild male.. os ya ji 6 9

Hermaphrodite e male ( i pee from’ wid... <2 ... 15 0 14

Female X male (1 generation from wild): ke N A l 10 3 7

Hermaphrodite X intermediate (origin a muknowe) Vise: 6 + 15

Female X intermediate (origin unknown)............... 0 1 3

38 14 48°

The various combinations will be considered separately.

The cross hermaphrodite (FH) X wild male (FM) pro-

duced 7 hermaphrodites, 6 females and 9 males, somewhat

approximating the expected ratio of 1 female (FF) : 1

5 The result of hermaphrodite X wild male.

° These totals do not quite coincide with those given in the published data,

as the parents of 7 of the vines were not certainly known and are therefore

omitted.

No.597] . INHERITANCE OF SEX IN THE GRAPE 561

hermaphrodite (FH) : 2 males (FM) and (MH). The

male MH is an entirely new genotype but apparently can

exist as shown by the next cross, in which a hermaphro-

dite was pollinated by a male derived from this cross,

Fifteen hermaphrodites, no females and 14 males were

produced, the expected ratio being (if a male of the type

HM were used) 2 hermaphrodites (HF and HH) : 2

males (MH and MF). If a normal male of the constitu-

tion FM had been used on a hermaphrodite of the con-

stitution HF we should expect to have produced 1 her-

maphrodite (HF) :1 female (FF) :2 males (HM and FM).

No females were produced. Again we might assume that

the hermaphrodite used was of the constitution HH and

that a normal FM male was used. In this case we should

expect a 1:1 ratio of hermaphrodites and males as before,

but in this case all of the males would be of the new type

HM. It seems, therefore, that the production of this new

male genotype (MH) must be admitted. Further evi-

dence for the production of males of this type is produced

in the cross female (FF) X male (one generation from

wild) which produced 10 hermaphrodites, 3 females and

7 males. If the males used had been of the normal type

FM only females and males could have been expected, as

are found under wild conditions, and no hermaphrodites.

If a male of the type HM were used, however, the expected

ratio would be 1 hermaphrodite HF : 1 male FM. The

presence of three females, although not expected, from the

cross FF x HM can be readily explained, as it has already

been pointed out that the males ‘‘one generation from the

wild’’ are of the two genotypes FM and HM, but of one

phenotype, and therefore could not be distinguished at

the time of pollen collection.

The cross hermaphrodite X intermediate (origin un-

known) which produced 6 hermaphrodites, 4 females and

15 males, throws some light on the genetic constitution of

these intermediates and incidentally upon the suppression

of femaleness. Observations on a number of ‘‘inter-

mediates” produced at the Minnesota Fruit Breeding

562 THE AMERICAN NATURALIST [ Vou. L

Farm showed that certain clusters of a vine may be

entirely staminate, while others of the same vine con-

tain all gradations from staminate to functionally per-

fect flowers, many of which are capable of setting fruit.

There is very evidently a suppression of femaleness in

certain parts of these plants and not in others. This

raises the question as to whether pollen from the pure

staminate clusters can transmit only determiners for

maleness and femaleness, or whether they are able to

transmit the hermaphrodite condition. Mr. Anthony in- |

forms me that the pollen used in the above cross was

‘“‘most certain to have come from such blossoms”? (i. e.,

from pure male clusters). If the two types of gametes

produced by these flowers bear the determiners H and F.

respectively, the cross hermaphrodite intermediate

should produce hermaphrodites and females in a 3:1

ratio and no males, while if these male flowers function

as normal males and the gametes produced carry the

determiners F and M respectively, a ratio of 1 hermaph-

rodite (HF) : 1 female (FF) : 2 males (HM and FM)

would be expected. -A close approximation to this ratio

was actually produced.

The cross female (FF) x intermediate which produced

1 female and 3 males, gives further evidence that the

staminate flowers of the intermediate vines do not pro-

duce gametes bearing the hermaphroditic determiner, but

act as pure males. Otherwise the appearance of the

three males can not be explained.” Although the number

of vines produced from this cross is small, still the appear-

ance of the three males is extremely significant.

It has already been pointed out that in the wild there

are two types of vines, male and female, and that under

cultivation a third type, the functional hermaphrodite, is

common. We are now in a position to discuss the pos-

sible origin of these types. It is clear that both the

staminate and functionally pistillate vines carry the de-

terminers for femaleness and maleness, respectively,

7 Anthony (1914) pointed out the fact that the pollen ea these inter-

mediates ‘‘seems to behave as the pollen of a pure male.

No.597] INHERITANCE OF SEX IN THE GRAPE 563

partially suppressed and therefore, there are two possi-

bilities with regard to the origin of functional hermaph-

rodites. (1) Maleness may express itself fully in one

of the chromosomes bearing the determiner for female-

ness in a pistillate plant, and (2) femaleness may express

itself fully in the chromosome bearing the male de-

terminer in the staminate plant. I think it can be said

definitely that functional hermaphrodites have been de-

veloped in both of these ways. The production of her-

maphrodites from the cross female X female can hardly

be explained on any other basis than entire lack of sup-

pression of maleness in certain gametes bearing the female

determiner, while the appearance of well-developed pistils

in a few flowers of certain male vines must be the result

of lack of suppression of femaleness in at least a portion

of the somatic cells of these males.

As there is an apparent segregation in the somatic

tissue of these vines, whole clusters and occasionally all

clusters on a cane being staminate while others bear

many intermediate and perfect flowers, it seems logical to

assume that the perfect flowers can transmit the her-

maphroditie condition to some of their seedlings through

both the male and the female gametes, resulting in either

homozygous or heterozygous hermaphrodites, all of whose

flowers are perfect.

LITERATURE CITED

Anthony, R. D.

1914. Methods and Results in Grape Breeding. In Proc. Soc. Hort.

Soc., 1914, pp. 81-86.

Bateson, W., Saunders, Miss E. R., and Punnett, R. C.

1908. Sterility of Anthers and Axil Color. In Ropt. Evol. Comm. Roy.

Soc.,

Beach, S. A.

1899. dopey: Self-Sterile Grapes. In N. Y. State Agr. Exp. Sta.

Bul, 169, pp. 331-371.

Booth, N. O.

1902. i Lead of Grape Pollen. In N. Y. State Agr. Exp. Sta. Bul.

, pp. 291-320, 6 pl., 1 fig.

Dorsey, M. pt :

1912. Variation in the Floral Structures of Vitis. In Bul. Torrey

Bot. Club, Vol. 39, No. 2, pp. 37-52, 3 pl.

564 THE AMERICAN NATURALIST [ Von. L

1913. Pollen Development in the Grape with special reference to

Sterility. In Cornell rapid ee Also, Minn. Agr. Exp.

ta. Bul. 144, 60 pp., 4 pl.,

Gard, M.

1913. Les éléments sexuels des hybrides de vigne. In C. R. Acad,

Sci. Paris, Vol. 157, pp. 228

Hedrick, U. P., and Anthony, R. D

1915, ERENT of Certain Characters of Grapes. In Jour. of Agr.

Research, Vol, 4, No. 4, pp. 315-330.

one George Harrison

Inheritance ‘of Sex in Lychnis. In Bot. Gaz., Vol. 49, pp. 110-

125, 2 figs

1911. Reversible Sex-mutants in Lychnis dioica. In Bot. Gaz., Vol. 52,

No. 5, 5 figs.

1914, poe Limited Inheritance in Lychnis dioica L. In Zeitschr. f. ind.

bst.- u. Vererbungsl., Bd. 12, pp. 265-302, 5 figs.

ON PRIMARILY UNADAPTIVE VARIANTS!

JOHN TREADWELL NICHOLS

AMERICAN Museum or NATURAL HISTORY

THs paper deals with vertebrate variants (forms or

species of animals more or less related but differing from

one another) which, although geographical, are not

direct or obvious responses to the environment.

Several types of variants are defined. Representa-

tive forms occupying adjacent regions are designated as

adjacent races or species: Forms intermediate in struc-

ture between adjacent forms and occupying territory

remote from them as foreign intermediates: Related

forms occupying the same territory and contrasted in

superficial characters as complements: Forms separated

geographically and showing greater resemblance (not in-

duced by environmental adaptation) than their degree

of relationship would presuppose, as outcrops.

The hypothesis is advanced that, probably on account

of competition, closely related forms are antagonistic.”

That is, when in touch geographically they tend to force

one another apart in superficial characters. If this

hypothesis, which seems to fit into certain known facts

extremely well, be accepted, it involves a centrifugal

force in Evolution opposed to the centripetal tendencies

of blood relationship.

It is the main theme of the paper to advance the con-

cept of these two forces as the fundamental framework

of evolutionary control, the helm which is swayed by

natural selection or other forces.

Larus marinus (the great black-backed gull), which

1 Paper read before the Section of Biology of the New York Academy of

Sciences, February 14, 1916.

2 See in, ‘‘Origin of Species,’’ p. 57. ‘‘We can dimly see why the

competition thonld be most severe between allied forms, which fill nearly

the same place in the economy of nature.’’

565

566 THE AMERICAN NATURALIST [Vou. L

breeds in the north and reaches our latitude only in

winter, finds its nearest relative not in an equatorial

species, but in Larus dominicensis of the southern hemi-

sphere, below 10° south latitude. This bird is more or

less of an intermediate between Larus marinus (the

great black-back), and Larus fuscus (the lesser black-

back) which occurs with the former in Europe and is

represented by allied forms eastward to the Pacific coast

of North America, the mantle of dominicensis is very

dark as in the larger bird; its size about that of the lesser

ones. These are four in number—Larus fuscus occurs

in Europe east to the Dwina, Larus affinis from the

Dwina eastward across Asia, Larus schistisagus in

northern Japan and Bering Sea, Larus occidentalis on

the west coast of the United States. The four birds of

this series are most readily separated by size and pro-

portions of feet. Affinis resembles schistisagus in the

former, fuscus in the latter; schistisagus resembles affinis

in the former, occidentalis in the latter. That is affinis

and schistisagus are more or less intermediates struc-

turally between the species they separate geographically.

For convenience we will call them ‘‘adjacent intermedi-

ates’’ as opposed to dominicensis which we will call a

‘‘foreign intermediate’? between marinus and birds of

fuscus group.

A species of vertebrate animals distributed over a wide

geographical area often varies in the different regions it

inhabits sufficiently to be separable into different inter-

grading races. Ordinarily no two of these races from

the very nature of their origin will be found inhabiting

the same region, sometimes they mingle in migration.

Often these races are in direct response to different en-

vironmental conditions, but sometimes this response

ean not be traced. Such races form a series comparable

to the gulls of the Larus fuscus group, and we may call

them ‘‘adjacent races’’ and the gulls of the group re-

ferred to ‘‘adjacent representatives,’’ also we may call

Larus dominicensis a ‘‘foreign repr tative’’ of the

No.597] PRIMARILY UNADAPTIVE VARIANTS 567

black-backs as well as a ‘‘foreign intermediate’’ between

the two divisions of that group.*

Many of the smaller gulls have a well-defined dark

hood in the adult. These hooded species may be divided

into two apparently natural groups, the first in which the

hood is black, the second in which it is dark brown. In

the former group Bonaparte’s gull (Larus philadelphia)

has a great deal of white in the primaries, making a

lengthwise band in the wing, conspicuous in life. Bona-

parte’s gull is found in North America. In the brown-

headed Larus ridibundus of Europe and Asia the white

in the wing makes a similar conspicuous mark, so that

no one seeing the two species in life could fail to note the

great resemblance. They show distinctly what is ordi-

narily termed parallelism. Such more or less distantly

related, geographically separated parallels, not environ-

mental parallels, are of not infrequent occurrence. For

convenience we will call them ‘‘outcrops.”’

Probably the closest relative of Larus philadelphia is

Larus saundersi from the inland waters of China and

Mongolia, visiting the coast in winter. This would be a

foreign as opposed to an adjacent representative.

Using, for the sake of familiarity, some of the same

material we have already considered, if we contrast

white-winged Larus philadelphia, which is common in the

vicinity of New York, with its closest relative common

here, the laughing gull Larus atricilla, we will find that

the two are as different as the limits of the natural black-

hooded group of which they are both members will allow.

The white primaries of philadelphia are contrasted with

the unusually dark primaries of atricilla, the mantle of

the former is pale, that of the latter dark, and there 1s

considerable difference in size. Such contrasted co-

existing allies are of frequent occurrence, let us call them

‘‘complements.’’ Pe,

For my next example I will turn to an entirely different

group of animals, the Spanish mackerels, fishes of the

3 See Matthew, W. D., Ann. New York Acad. Sci., 1915, Vol. XXIV, p.

180, ‘‘ Principles of Dispersal.’’

568 THE AMERICAN NATURALIST [Vou. L

genus Scomberomorus, found in warm seas of the world.

Along the Atlantic coast of America S. maculatus is

abundant south to Florida, and S. regalis is abundant

about Cuba, where maculatus is practically unknown.

Regalis is a close ally, we may reasonably say a deriva-

tive of maculatus, from which it differs in the scaling of

the fins, arrangement of the numerous brownish spots

which ornament both species, and in minor characters.

Occurring abundantly between the two and associated

with one to the north, the other to the south, is unspotted

S. cavalla (the kingfish of EAT I will call such an

interposed species an ‘‘intrusion.’’ Incidentally cavalla

is a complement of both maculatus and of regalis.

All the above mentioned cases of variants have this in

common: although more or less geographical they are

not obvious and direct responses to the environment;

probably environment has little to do with them. iAp-

parently we get all the various types of variants as

classified where such control is lacking or not essential.

It is then pretty certain that variants occur when not

induced by environment. Lacking contrary evidence this

hypothesis is accepted, and cases where it apparently

obtains are alone considered in this article, environ-

mentally induced variants being too common and too

generally discussed to need treatment here.

Let us go over the various types of variants, begin-

ning with the simplest, and discuss them in relation to

their probable origin. The admittedly inherent tendency

to variation in any form would be sufficient, where the

form is widely distributed, to break it up into adjacent

races. Adjacent representative species are pretty obvi-

ously derived from adjacent races. Even temporary

isolation so readily explains this slight advance and is so

easily assumed, that perhaps we should look no further

for explanation. So far we have trodden familiar

ground scarcely worthy of mention, and the hypothesis,

that namely, probably on account of competition, closely

related forms are antagonistic, is conservative enough.

No.597] PRIMARILY UNADAPTIVE VARIANTS 569

If this hypothesis which seems to fit into the known facts

with surprising neatness, be accepted, the writer believes

that it will explain a force in direct opposition to the

commingling of blood along a line of geographical de-

mareation of adjacent races, tending to drive those races

apart in structure and making possible the derivation

from them of adjacent species without geographical

isolation. In fact we find in the incompatibility of

closely related forms a powerful centrifugal force for

differentiating species and forcing them apart*—and it

also explains complementary coexisting forms, as it

would tend to make coexisting forms complementary and

would not act to change or eliminate them if they already

were so. In the case of the foreign intermediate we have

a geographically isolated form less influenced by the cen-

trifugal forces, therefore varying less, retaining inter-

mediate or primitive characters. This centrifugal force

should always be considered as balanced against blood-

relationship, doubtless the chief cause of resemblance in

species. In the outcrop we have a case where the centrif-

ugal force is inoperative and the centripetal tendency

brings about a parallelism different in fundamental na-

ture from the more familiar environmentally induced

parallels.

There seems to be an analogy between the outcrop and

homologous rectigradations in Paleoevolution.

aving given this rather concentrated outline of my

hypothesis and the class of facts it is designed to explain

it will not be out of place to mention other widely scat-

tered examples of the important classes of variants al-

luded to. I will begin with the foreign intermediate.

Trichiurus is a long band-like silvery fish with a fila-

mentous tail found along the shores of warm seas, where

also occur numerous representatives of the Scombroid

or mackerel-like fishes. The two are utterly unlike, yet

a clear line of relationship is found through intermediate

forms (for instance Lepidopus) from the ocean depths.

4 See, en Gulick, J. T., ‘“‘ Evolution, Racial and Habitual,’ . 1905,

p. 258. We ae

570 THE AMERICAN NATURALIST [ Von. L

Lepidopus and its confreres are a foreign intermediate

family between the mackerels and Trichiurus, as Larus

dominicensis is a foreign intermediate species between

the two closely related types of black-backed gull in the

north. Please note another analogy. The fundamental

philogenic Scombroid characters of Lepidopus are a

direct response to life at the surface of the mackerel

group. Though an intermediate it is impossible (as it is

difficult in Larus dominicensis) to believe that it occupies

the primitive habitat of the group. Other fishes which

have persisted in regions distant from the center of

ichthyological competition, namely the marine shore line,

as in the deep sea or in fresh waters, are in a sense foreign

intermediates. Before leaving Trichiurus let us glance

at its ancestry. The deep-sea Lepidopus-like fishes

would have little chance in competition with the

mackerels in surface waters. They form in.a sense a

rectigradation (again apologies to Paleontology) de-

velopment from the mackerels of which Trichiurus is the

terminal member; but Trichiurus is so unmackerel-like

that it may and does inhabit the same waters with the

mackerels; it is broadly speaking complementary to

them.

The tree squirrels are connected with the ground in-

habiting spermophiles through the chipmunks (Eu-

tamias, Tamias and Callospermophilus). Our common

eastern chipmunk (Tamias) is a boldly marked animal

in appearance very much resembling the Rock Squirrel

(Callospermophilus) of high altitudes of the northwest.

The resemblance is more striking than the closeness of

blood relationship would presuppose, and Callosper-

mophilus may be characterized in this connection as an

outcrop. The forms of Eutamias inhabiting the same

regions as Callospermophilus and sometimes found asso-

ciated with it, are very unlike that animal, and form a

very good complement with it. ‘The eastern chipmunk is

a foreign intermediate in superficial characters between

Callospermophilus and western Eutamias. Interlocking

No.597] PRIMARILY UNADAPTIVE VARIANTS 571

of foreign intermediate and outcrop in this case is in-

teresting, it may or may not be significant. Inhabiting

more or less the same region as Callospermophilus, as

do western Eutamias, are other spermophiles. They are

however speckled or plain colored, little marked for the

spermophile group, complements of it, whereas widely

separated from it, to the south and east occur sper-

mophiles with sufficient striping to perhaps be considered

foreign intermediates between these two northwestern

types of spermophile. The above discussion will show

how in a single group of mammals, ground squirrels,

there exist the various types of variants which have

been differentiated above in discussion of fishes and

birds.

The most favorable region for fish development ad

the center of evolutionary competition for fishes is the -

tropical coral reef. A great many species have the habit

of seeking protection among the projections and crevices

which there occur in abundance, and these as a rule are

very brightly colored, blue, yellow, gréen, red, or marked

with bold bizarre patterns in endless variety. The

theory has been advanced and quite generally accepted

that the colors harmonized with the brightly colored

corals, etc., over which they occurred, but the fact seems

to be that although with occasional spots of color the tone

of the reef as a whole is comparatively uniform, these

bright-colored fishes are very conspicuous swimming

over it, in fact give it a good deal of its brilliant appear-

ance. Butterfly fish, wrasses, chromids, parrotfish, ete.,

belong to this bright, varied crowd. They are safe from

all enemies among the labyrinths of the reef and in their

evolution have not felt the necessity for concealment.”

Neutral tones are closer the one to the other than the

bright and bizarre, therefore from centrifugal force we

should expect, as we find, the bright and bizarre.

Brilliant birds amidst a luxuriant foliage simulate the

conditions and the security of reef fishes and frequently

5 Reighard, J., Publie. Carnegie Instit., Washington, 1908, No. 103.

572 THE AMERICAN NATURALIST [ Von. L

have similarly bright and striking plumage, especially in

the males, though the sitting females may be dull. The

bright colors of the males may readily have been ac-

quired under control of similar forces. I do not mean

to rob sexual selection of its reputed force, but indica-

tions are that it is not entirely responsible for all that

might be laid to its door. The genus Dendroica of small

active arboreal birds has developed numerous bright,

varied and beautiful colors. Two species perhaps as

closely allied as any others are the Blackpoll Dendroica

striata and Bay-breasted warblers, Dendroica castanea.

It is interesting to find them very fair complements, the

one of the other, in plumage of the breeding male, though

females and young are little different; whereas the male —

blackpoll resembles in color the black and white warbler,

` Mniotilta varia, a distantly related bird of the same

family. ;

Ten years ago the writer had the pleasure of making

the acquaintance of the beautiful grey slender-billed

fulmar (Priocella), on the southern ocean. Sometime

afterwards when crossing the North Atlantic he met with

the northern fulmar (Fulmarus) and was surprised to find

the resemblance between the two so great, even carried

to a light mark on the wing, very useful in field iden-

tifications. Though belonging to the same family the

two species are really not very closely allied and are an

example of the outcrop. Another case which might be

so considered is that of the African true larks or pippits

which simulate our meadow larks of the family Icteride

in color. ‘All passerine birds are so closely allied that

the two are not too distantly related to be outcrops, but

there are other reasons for thinking that this case is not

a very good one, but rather a case of environmental

parallelism. The black breast mark, for instance, is so

common among ground birds that it probably has con-

cealing value.

The isolated islet of South Trinidad in the south At-

lantic is remarkable in that three closely related species

No.597] ` PRIMARILY UNADAPTIVE VARIANTS 573

of petrels (Genus Avstrelata) are indigenous to it.

These are almost identical in size and build but differ

markedly in color. Astrelata arminjoniana is bicolored,

dark above and white below, Avstrelata trinitatis is uni-

formly dark colored, Æstrelata chionofara resembles the

former bird, but the white of the underparts spreads up

the sides of the neck and on to the back, which is largely

white, with dark shafts to the feathers. ‘These birds are

obviously very closely related and it is sometimes debated

whether or not they are distinct species or merely color

phases of one and the same thing. The most convincing

evidence, in favor of the hypothesis of distinctness

perhaps being that arminjoniana and trinitatis, forms

which have long been known to science, seem to breed on

the islet at somewhat different times of the year. Chio-

nofara has only recently been described and is so far

known from the type specimen only.

The writer has been particularly interested in these

birds and has studied them carefully with a view to

forming a definite opinion as to their relationship. The

bicolored type of plumage represented by arminjoniana

is perhaps the most common in the cosmopolitan pelagic

genus Æstrelata of which it is a member, but armin-

joniana is separated from most of the genus by the

greater development of dark color on the side of the neck

forming a sort of dark collar. Several uniformly dark

colored Æstrelata also occur in various parts of the

world, comparable to trinitatis. "The third and recently

described form is complementary in color to trinitatis,

being very white for the genus, and complementary in

pattern to arminjoniana, having the side of the neck

white instead of unusually dark.

If we had three recently evolved forms breeding on the

same islet, complementary plumage is what would be ex-

pected, and that the three types of Æstrelata from there

show such plumages is good evidence that they are bona

fide forms, not color phases.

6 Murphy, R. C., The Auk, July, 1915, XXXII, No. 3, pp. 342-344.

574 THE AMERICAN NATURALIST [ Vou. I.

What color phases are is a matter aside from the

thread of the paper but the consideration of this case

has led so close to the interesting unsettled problem that

the writer hopes to be pardoned for calling attention to

certain things about them which he has noticed and

which seem to hold good pretty well. First, they are

limited to about four manifestations. The timber wolf,

Canis occidentalis we are told is gray, white, black, or red.

The gray squirrel (Sciurus carolinensis) is gray or black.

The black bear (Ursus americanus) black (normal),

white (glacier bear) or red (cinnamon bear). The red

fox (Vulpes fulvus) gray (cross fox), black (silver fox)

or red (normal). The screech owl is gray or red.

Students of heredity have shown that the normal gray

_ coat of a wild guinea pig is a composite of black, white

and red, which has been broken down by breeding into its

constituent parts so that we get black guineas, red

guineas, white guineas and guineas with the colors in

patches. The colors, you will notice, are the same as

those of the color phases occurring in nature and it

seems probable that color phases have come about by a

similar breaking down and reduction of the normal

colors in a species, and have definite limits beyond which

they are not likely to go.

In conclusion one might go on indefinitely demonstrat-

ing the application of the rough classification of non-

adaptive variants proposed in our initial paragraph,

limited only by the number of forms one could call to-

memory. Jf the significance of the classification has not

been exaggerated it is difficult to find another theory —

which fits as well with the existing phenomena as the one

advanced of a centrifugal force, though it is a pure theory

and its acceptance a matter of individual taste.

SHORTER ARTICLES AND DISCUSSION

TABLES OF LINKAGE INTENSITIES

In the July, 1916, Natrurauist, Professor Emerson gives con-

venient formule for calculating linkage intensities. His formula

(I) is especially useful because it is applicable either to cases of

coupling or to cases of repulsion. I had independently worked

out empirical formulae for calculating coupling and repulsion

which are very similar to that given by Emerson ; indeed they are

identical with it, if 1 is substituted for r in cases of coupling and

for s in cases of repulsion. I had failed to observe, what Emer-

son shows, that the two formule may be given a single generalized

form.

TABLE I

THe F, RATIO, 9: 3: 3:1, AS AFFECTED BY COUPLING or LINKAGE, A AND B

ENTERING THE F, ZYGOTE IN THE SAME GAMETE

Ratio, Cross-over to | Proportion Fz Zygotes

Non-cross-over Cross-over “SRT Gs

Gametes Gametes AB Ab aB ab ` Total

Te — 3a2-+2(22+1) |2x+1|2r+1 x? | (27 +2)?

p 1/2 9 3 3 1 16

132 1/3 22 5 5 4 36

1:3 1/4 41 7 ri 9

1:4 1/5 66 9 9 16 100

1:5 1/6 97 11 11 25 144

1:6 [7 134 13 13 36 196

erd 1/8 17. 15 15 49

1:8 1/9 226 Bev § 17 64 324

1:9 1/10 1 19 19 8l

1:29 1/100 29,801 199 | 199 | 9,801 | 40,000

Limiting values? ........ TERE 3 0 0 | +i Ma

I had also found it convenient, for my own use, to make out

and enter in my notebook tables of equivalent gametie and zygotic

series, so that when a suspected case of coupling or repulsion

comes to notice the nearest integral gametic series can at once be ee he

determined by inspection of the table, without making the caleu-

1 No coupli

2 Not distinguishable from the case in which - A and B are » due to a singlo es ne

_genetic factor. :

576 THE AMERICAN NATURALIST [Vou. L

lation anew. With the idea that these tables may possibly be

useful to others, they are given herewith. In making use of such

tables it is necessary only to reduce to the basis of a common total

the observed F, zygotic series and any series of the table with

which a comparison is desired. The total given in the table is in

each case the lowest one which involves no fractions. If one uses

the tables, such formule as Emerson’s (II-IV) will not be found

necessary in estimating the strength of the linkage. Moreover

those formule are less useful than tables in dealing with the modi-

fied dihybrid ratio, 9:3:4, which happens to have been the first

case that I encountered in my own work. The modified ratio as

affected by linkage may be read directly from the table by com-

bining classes aB and ab.

TABLE II

Tue F, Ratio, 9: 3: 3:1, AS AFFECTED BY REPULSION (NEGATIVE LINKAGE),

AND B ENTERING THE F, ZYGOTE IN DIFFERENT GAMETES

Ratio, Cross-over to | Proportion F2 Zygotes

Non-cross-over | C er E EER

Gametes | Gametes AB Ab aB ab Total

liz =o 2(x? +21) +3 |x? +2r| 2?+22} 1 (2z +2)?

1:13 1/2 | 9 3 S4avd 16

1:3 1/3 | 19 8 8 1 36

1:3 1/4 33 15 154 64

1:4 1/5 51 24 24 1 100

1:5 1/6 73 5 35i -1i 144

1:6 1/7 99 48 48 1 196

1:7 1/8 129 63 63 i 256

1:8 1/9 1 80 80 1 324

1:9 1/10 201 99 99) 1 400

1:99 1/100 20,001 9,999 | 9,999 | 1 | 40,000

Limiting vai a 2 1 1 0 $ 4

3 No repulsion;

t Not distinguishable from the case in which A and B are allelomorphs.

W. E. CASTLE

BUSSEY INSTITUTION,

July 10, 1916

VOL. L, NO. 598 OCTOBER, 1916 :

THE

AMERICAN pits

NATURALIST ©

AMONTHLY TOURNAL ==

~ Special Reference to the Factors of Evolution —

The American

Naturalist

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

e Editor of THE AMERICAN NATURALIST, Garrison- on-Hudson, New York.

oa ,

icles containing summari

es of research work beari

blems s ee evolution are especially welcome, and will be given preference

in PES

ndrea reprints of eur etna are supplied to authors free of charge.

pee

Further repriats will be supplied at co

Subscriptions and advertisements Ae = a bat ag _publishers.

subscription A is four dollars

The

"e is fifty cents and

The chargo” for single copies is

THE SCIENCE PRESS

Lancaster, Pa.

N EW TORE

+ a bb.

, April 2

a Post O

Garrison, N. Y.

Sub-Station 84

at th

fice at Lancaster, Pa., under the Act ot

cue of March 3, 1879.

NATURAL HISTORY OF MAN |

mae mn ee

18, ,000 Specimens illustrating the daily life of ion within a radius of on ag tbe kipea eg

Savage Races, many now extinct. Weapons, - | eluding the pre- and post-glacial deposits occurring in the

ments, Deities, Currency, Pols, ete: | DINE Sime etthe suclant Deleware River Dis pHi PES EEE

Guaranteed localities and age. Type Collections for | More than 700 origin (itistrations, enlarged 80 ‘times, with

Museums a Specialty. complete index “size 924 Bound rt Vellum cloth.

‘ t pestage

Address Charles S. Boyer,

OLDMAN, 77 Brixton Hill, London, England | s40 cotumpia stu’ >. POF? Philadelphia, Pa

SCIENCE

A WEEKLY JOURNAL sie ana TO THE ADVANCEMENT OF SCIENCE

, SEPTEMBER 1

FRIDAY,

aa Life and Work of Carl RSS gaan Warren P.

Saar en ek Asset of National Parks. Joseph Grinnell |

2 and Tracy I. Storer.

The Revival of Tntereas in | ird Anato:

aes _ States National Museum, Dr. R. W. rg ah Re

-Univesity and Edacational News

Bieter Selentife Ap

Scien

: song and ‘ca

FRIDAY, SEPTEMBER 22, 1916

The British Association for the Advancement of Science:

New Arenean rere Lights on the Origins of Civilization

in Europe: Sir hur Evans.

ao of Thought: Professor A. rae

Scientific Notes and News.

Darney S RARO News.

opan Aail eee pre sot Dat oe

Early Man: Dr. M. W. Lyon, Jr. ie

tific Boo!

The The Napier Te eo Memorial Volume: protest f

A Sew Bioglan sen am E.G. Fischer.

Petine as uae ae

; as

and other Cereals: Professor L. R. oc S G. Johnson,

S S. Rar. or Use of the Dou

D. Frost and Freda M. Bachmann. E

Zhe St. Louis Academy of £

Greenman,

"State Numbers 15 cots-

THE

AMERICAN NATURALIST

VoL. L. October, 1916 No. 598

A NOTE ON THE INHERITANCE OF EYE

PATTERN IN BEANS AND ITS RE-

LATION TO TYPE OF VINE!

FRANK M. SURFACE

In a recent paper (Pearl and Surface, 1915) from this

laboratory two varieties of yellow-eyed beans were de-

scribed and figured under the somewhat provincial names

of Improved Yellow Eye and Old-Fashioned Yellow Eye.

The type of eye pattern characteristic of each of these

varieties is shown below in Figs. 1 and 2. On the Im-

proved Yellow Eye the colored area covers about one

fourth the area of the bean. The outer border of the eye

pattern is clear-cut and regular, with very little or no

spotting on the remainder of the bean.

The Old-Fashioned Yellow Eye pattern (Fig. 2) is

much smaller in area and is quite irregular in outline but

nevertheless very definite. It consists of at least three

color centers: (1) A posterior? spot covering the caruncle

and extending at least part way around the hilum. Lat-

erally this area is extended into two rather broad wings

which reach as far forward as the micropyle. (2) An |

anterior spot surrounding the micropyle, and (3) an an-

terior stripe which may or may not connect Liege the |

micropyle spot. A

In connection with other work a number of crosses Mars

1 Papers from the Biological Laboratory of the Maine | Ag rieul a

periment Station, No. 99. eas

2The posterior erior en of a bn is that on of the bilom

earuncle lies. ie Se ee EE , oe

BIT

578 THE AMERICAN NATURALIST [Vou. L

been made between these two varieties. Something over

40 cross-pollinated beans have been secured. Of these,

15 have been grown at least as far as the F, generation.

These 15 plants gave a total of

295 F, beans. Except for some

minor fluctuations these F, beans

were all alike, but differed mark-

Ey renege edly from either parent. In the

Yellow Eye color pattern. notes these F, beans have been

designated ‘‘Piebald’’ because of

the very irregular spotted pat-

tern. Fig 3 shows a typical pie-

: bald pattern. In addition to the

eau ame Me a spotting these beans differ from

Yellow Eye color pattern. the Improved Yellow Eye in hav-

ing a very irregular outline to the

colored area. While the pattern

is somewhat variable, there is

never any difficulty in distin-

guishing this from the typical Im-

aloe sadist te proved Yellow Eye pattern.

imeored end bakaa Up to the present time omy

TOTE ail fase a few of these hybrids have been

carried to the F, generation. However, enough have

been obtained to show that these piebald beans give both

parent types and also more beans with the piebald

pattern. It is very probable that only these three types

occur in the F, and later generations.

While the data so far obtained from hand-pollinated

hybrids are not sufficiently extensive to warrant further

discussion, certain other data have been obtained which

have a bearing on this subject.

In 1911 and 1912 the Experiment Station grew a num-

ber of plots planted with different strains and varieties of

beans. Among these were a number of strains of Im-

proved and Old Fashioned Yellow Eye. In some cases

plots of these two varieties were located near each other.

Seed from some of the 1911 plots were planted in‘ 1912.

No. 598] INHERITANCE OF EYE PATTERN Ha

A considerable number of plants in these plots showed

that the seed had been cross-pollinated by bumble-bees the

year before. Among the plants in the Yellow Eye plots

there were a number which bore typical piebald beans

similar to that shown in Fig. 3. Some of these piebald

plants were harvested separately and their progeny con-

tinued in a small way inside a screened cage. By the

spring of 1915 it had been ascertained that a cross be-

tween an Improved and an Old-Fashioned Yellow Eye

resulted in such a piebald pattern. Accordingly a con-

siderable number of these piebald beans were grown in

1915. The following paper is based upon the data from

these natural hybrids.

Table I gives the detailed data relative to the offspring

‘

TABLE I

SHOWING THE SEGREGATION IN THE PROGENY OF PIEBALD BEANS

Pedigree No. Year | Row | Piedad | -1. Y.E, |O.F. Y.E.

1294-5 1912 30 11 $ 4

1913 87 — 3 g

88 5 — —

1914 31 4 — =-

1915 58 2 3 4

š 13 7 3

270 7 2 5

271 8 t 8

272 10 3 1

273 8 — 2

, í 274 > 6 3 6

Total for 1294-5 7i 26. - 30

1311 3

siy

Bes 4 :

4 5

To 3. i

Total for 1817 % A has

1a e Aoa

T tal Tor Ix- : Fig :

me 8

Total for 1818

Gi a nd s

DOU =): THE AMERICAN NATURALIST [Von. L

of 269 piebald beans. It will be understood that each

row was planted from the offspring of a single plant.

Not all of these beans can be considered as belonging to

the F, generation. A portion of these certainly belong

to the F, and later generations. This question will be

considered further in a later paragraph.

From this table it will be seen that only three kinds of

beans were obtained from these piebald seed. These

were piebald, Improved Yellow Eye and Old-Fashioned

Yellow Eye. This fact, in connection with the evidence

obtained from controlled pollinations as noted above,

makes it practically certain that these piebald beans are

hybrids between these two varieties of Yellow Eye beans.

Further, with the exception of three small rows none of

of these piebald beans gave evidence of breeding true.

In each of these three cases some of the piebald beans

have split in later generations. Thus in pedigree No.

1294-5 the 1914 Row 31 is the offspring of one of the five

piebald plants in the 1913 Row 88. It seemed possible

that this line was breeding true. However, the 1915 Row

58 is the offspring of a plant from Row 31 of the year be-

fore, and Row 58 gave all three types, so that both of the

preceding rows must have been heterozygous. If larger

numbers had been grown from the same seed they would

undoubtedly have thrown all three types.

The evidence thus indicates that the piebald pattern is

the expression of the heterozygous condition of the fac-

torial difference between these two types of Yellow Eye

beans. A similar conclusion was reached by von Tscher-

mak (1912). He obtained spotted beans very similar to

our ‘‘piebald’’ from crosses between eyed and white or

eyed and solid color beans. These piebald beans were

always heterozygous, throwing on the one hand a large

eye with regular outline corresponding with our Improved

Yellow Eye and on the other hand a small-eyed bean.

Judging from his figures (p. 208) von Tschermak’s small-

eyed bean had nothing corresponding to the peculiar pat- —

tern on our Old-Fashioned Yellow Eye. However, in

No. 598] INHERITANCE OF EYE PATTERN 581

relative quantity of pigment these beans agree very well.

Von Tschermak assumed a unifactorial difference be-

tween the large and small-eyed beans, with the spotted

pattern as the heterozygote. In the F, generation he ob-

tained a 1:2:1 ratio.

Returning now to our own data as given in Table I it is

clear that if the difference between the Improved and Old-

Fashioned patterns is due to a single factor we should

expect in the segregating generations 2 piebald:1 I. Y. E.

:1 O. F. Y. E. The numbers obtained in Table I will

hardly support this view. 146:53:70 can hardly be

looked upon as a 2:1:1 ratio. It is true that the devia-

tion is not so great, but that these observed numbers might

be chance fluctuations from a 2:1:1 ratio. On the theory

of probability the odds against the occurrence of such a

deviation are about 5 to 1.

Of the more common Mendelian ratios the observed

figures are much more closely fitted by 9:3:4. The ob-

served and expected numbers in this case are

Piebald | I. Y. E. | O.F. Y. E.

Observed N o he | o ty

Expected No. on 9:3:4 ratio........... | 151.3 50.4 | 67.3

Tt is clear that there is a very reasonable agreement.

Further evidence in support of the view that the segre-

gation is not 2:1:1 is found by examining Table I in more

detail. Thus the totals for each of the five pedigrees show

an excess of Old-Fashioned Yellow Eyes over the Im-

proved type. In three of these pedigrees the number of

plants is relatively small. However, the cumulative evi-

dence makes it almost certain that the deviations are not

due to chance.

It was stated above that only a portion of these iati

belonged to the F, generation. In a bifactorial character

considerable difference might be introduced by the com-

bination of data from different generations. From the

records it is known that all the plants in pedigree Nos.

153 X, 1318 and 1321, together with two rows, 104 and 292,

582 THE AMERICAN NATURALIST [ Vou. L

from pedigree 1311, belong to the F, generation. Taking

these plants alone, we have the data given in Table II.

TABLE II

SHOWING THE SEGREGATION IN THE F, GENERATION FROM PIEBALD BEANS

Piebald 1. ¥o8, OFX. Ee

It is seen at once that there is again the same relative

excess of O. F. Y. E. over I. Y. E. that is shown by the

complete data in Table I. The expectation on the 2:1:1

ratio is 50.5: 25.3: 25.3, while on the 9:3:4 ratio the ex-

pectation is 56.8:18.9:25.3. It will be seen that the lat-

ter figures more nearly fit the observed numbers.

A 9:3:4 ratio presumes a bifactorial composition.

However, a moment’s consideration shows that such a

ratio cannot have its usual significance in this case. If

this were the usual bifactorial segregation, one out of

every nine F, piebald beans ought to breed true in the

third generation. Yet out of 15 rows from piebald beans

which certainly belong to the F, or F, generation not a

Single one bred true.

Further, one half of the F, Old-Fashioned Yellow Eye

segregates and two thirds of the F, Improved Yellow Eye

segregates ought to show segregation in the third genera-

tion. In 1915, 43 Old-Fashioned Yellow Eye, F, plants

were grown and every one bred true. At the same time 38

F, plants were grown from Improved Yellow Eye seed.

_ Thirty-seven of these gave typical Improved Yellow Eye

beans, but one plant gave piebald beans. The F, plant

which furnished this latter seed was grown without any

protection from insects in 1912 and it is very probable

that the one I. Y. E. bean which gave piebald seed was due

to insect pollination with Old-Fashioned Yellow Eye

pollen. This is all the more probable because the ratio

1:37 is by no means what would be expected on the usual

bifactorial hypothesis.

The evidence is fairly conclusive that the I. Y. E. and

the O. F. Y. E. segregates breed true and that beans with

No. 598] INHERITANCE OF EYE PATTERN 583

the piebald pattern are always heterozygous. These re-

sults could be very simply interpreted on a single-factor

hypothesis, but the numerical results do not fit the 2:1:1

ratio demanded by that hypothesis.

While the data at hand are not as extensive as one might

desire in order to build a complete theory, yet there is

much to be said in favor of the following provisional

hypothesis. Let Z be a factor which in its homozygous

condition JJ produces the Improved Yellow Eye pattern.

Then Ii will be the zygotic constitution of the piebald

plants and ii that of the Old-Fashioned Yellow Eye pat-

tern. Assume further a lethal factor L independent in

its segregation and of such a nature that LL in the pres-

ence of II produces a non-viable zygote. The complete

F, segregation would then be as follows:

1 JILL Non-viable®

S FLLI

TIL LE

2 libh

4 iLi! Piebald

2L lilr

bee Be ae FP

ogee Et 0R rE

Tiiti:

Such a segregation would result in the ratio 8 piebald :

3I. Y.E.:40.F.Y.E. Testing this ratio against the

total observed numbers in Table I we get

| Piebald T ras orre

red Ko l ee 46 | Ú l b

Expected No. on 8:3:4 ratio...) 143.5 ps | nT

It is seen that there is a very ‘close agreement between

the observed and expected numbers ; much closer, in fact,

than in the case of the 9:3:4 ratio previously used.

3 The same result would be obtained if I in the presence IZ produced a :

non-viable zygote. This = ysl be jamia ae TN cromos Boe

Eroni na E T o oe

584 THE AMERICAN NATURALIST [ Von. L

In the case of known F, plants, as given in Table 2,

the results are

| Piebald s A S 0. e A

Observed No T 27

Expected No. on 8:3: 4 ratio.......... =e eS eee

Here again there is a very remarkable agreement. In

fact all of the data at hand fit into this theory very nicely.

Final proof of its correctness or incorrectness can only

come with more extended crossings between the segregates

and with the parent stocks. Such experiments are now

under way.

RELATION or Eye PATTERNS to TYPE or VINE

Two years ago while going over some data from pure

lines of Yellow Eye beans grown inside a screened en-

closure the writer was struck by the fact that with few

exceptions all of the O. F. Y. E. pure lines had the bush

type of vine, while nearly all the I. Y. E. lines were classed

as short runners. This point was further emphasized

by the observation that in several cases the segregation

from piebald beans showed that all of the O. F. Y. E.

segregates were bush beans and all the I. Y. E. were run-

ners. It was, therefore, of some interest to tabulate the

data relative to type of vine in connection with the eye

pattern.

_The classification of plants with reference to type of

vine has usually been made at the time of harvest. In

some years the plants grown inside the screened cage

have been classified as to vine type shortly before harvest.

In either case the plants were mature or practically so.

The plants grouped under the term ‘“‘bush’’ are those

which show determinate growth, terminal inflorescence,

and lack the ability to twine about supports. The ‘‘run-

ner”? plants show axillary inflorescence and the twining

habit (cireumnutation). All of the runner beans consid-

ered in this paper are of the short runner or short pole

type, rarely reaching a total height of more than 125 cen-

No. 598] INHERITANCE OF EYE PATTERN 585

timeters. Usually they develop few branches. Under

ordinary conditions such beans do not show indeterminate

growth. However, from the investigations of Emerson

(1916) it is probable that they would do so if growth were

not stopped by unfavorable conditions or excessive seed

production.

Data as to type of vine are available from 247 of the

plants given in Table I. Table III shows the distribu-

tion of the type of vine for each of the three eye patterns.

The data for each pedigree number are summarized sep-

arately.

TABLE III

DISTRIBUTION OF TYPE OF VINE FoR EACH OF THE THREE EYE PATTERNS

Piebald TR. a : O.F. Y. E.

Pedigree No. i

Runner Bush Runner | Bush Runner Bush

1294-5. uean 40 19 12 6 0 27

IIE. es! 18 19 4 i] 0 5

106 A E AO 3 10 2 2 0 7

ISIB ssi 0 10 0 3 ( 9

F REEE | 9 6 6 0 0 7

O oS 70 64 24 22 ( 67

The most striking thing in connection with this table is

the complete absence of runner vines among the Old-

Fashioned Yellow Eye beans. Apparently the gene for

bush type of vine is closely associated with the gene for

the Old-Fashioned Yellow Eye pattern. That this asso-

ciation is not absolute under all conditions is indicated

by the fact that I now have two strains of Old-Fashioned

Yellow Beans of unknown origin which for several gen-

erations have bred true to a distinct runner type of vine.

A number of crosses have been made using these runner

types of Old-Fashioned Yellow Eye. It is hoped that

these and other experiments which have been started will

throw some light upon this question.

Attention may be called to the apparent 1:1 ratio of

runner to bush in the case of the piebald and Improved

Yellow Eye beans. Emerson (1904, 1916), von Tscher-

mak (1904, 1912) and others have shown that in crosses _

between tall i and Gwar beans the o S : i

586 THE AMERICAN NATURALIST [Vou. L

ratio is 3 tall to 1 dwarf. Obviously the present data

are of little use in the study of this question because in

the first place it consists of a mixture of F., F, and F4

plants and in the second place the vine characters of the

parents in the different crosses are unknown. It is quite

possible that the parents in the case of pedigree No. 1318

were both of the bush type. The 22 plants in the F, gen-

eration in this strain are all of the bush type.

The only reason for presenting the data in Table ILI

at this time is to call attention to the relation between the

bush type of vine and the Old-Fashioned Yellow Eye

pattern. There seems to be no question but that these

two characters are closely associated.

LITERATURE CITED

Emerson, R. A. 1904. Heredity in Bean Hybrids. Ann. Rept. Nebr. Agr.

Expt. Stat. 1904, pp. 33-68.

1916. A Genetic Study of Plant Height in Phaseolus vulgaris. Nebr.

. Expt. Stat. Research Bull. No. 7, pp. 1-72.

Pearl, R. and Bartica, F. M. 1915. Studies on Bean Breeding. I. Stand-

ard on of Yellow Eye Beans. Ann. Rept. Me. Agr. Expt.

Stat. 1915, pp. 161-176.

von Tschermak, $ 1904. Weitere Kreuzungsstudien an Erbsen Levkojen

und Bohnen. Zeit. landw. Versuchsw. Osterr. eich, pp. 1-102

1912. Bastardierungsversuche an Levko ojen, Erbsen und Bohnen mit

ücksicht auf die Faltorenlehre. Zeit. induk. Abst. Vererb.,

Bd. 7, pp. 81-234.

CHROMOSOME STUDIES ON THE DIPTERA

III. ADDITIONAL TYPES or CHROMOSOME GROUPS IN THE

DROSOPHILIDÆ

CHARLES W. METZ

STATION FOR EXPERIMENTAL EvoLUTION, Coup Spring Harsor, N. Y.

In connection with other work on the Drosophilas I

have for some time been engaged in a comparative study

of their chromosomes, with especial reference to possible

phylogenetic relationships between different species. A

short preliminary report of this study was published two

years ago (Metz, ’14) after five types of chromosome

groups had been found among eleven species. More re-

cently I have studied fifteen additional species of Dro-

sophila, one of Cladocheta and two of Scaptomyza (re-

` lated genera), and have found several more types of

chromosome groups. Altogether twelve main types and

several sub-types have been identified—a series more ex-

tensive, I believe, than any heretofore recorded among

allied species. Of these twelve types all but one are

represented in the genus Drosophila.

The study has not yet advanced far enough to fulfil the

purpose for which it was originally undertaken, but in.

view of the widespread interest recently attracted to the

Drosophilas as objects of genetic research it seems desir- _

able briefly to describe the chromosomes of the species —

thus far examined without awaiting the completion ofthe

original investigation. — In doing this I shall endeavor

merely to give an accurate presentation of the chromo- _ = n

somal data, without dwelling on the theoretical considera- o

tions they may suggest, considerations which can receive o

adequate treatment T after ap more specie:

been examined. _ : oe

Since in almost every case devel or ee

shave ==

588 THE AMERICAN NATURALIST [Von L

the only ones suitable for study, it has been necessary to

breed the various species in confinement in order to de-

termine their chromosome groups. As a result only

about half of those collected have been studied cyto-

logically. Other determinations will be reported in the

future as they are obtained.

The material has been secured from four regions, New

York, Massachusetts, Alabama and Cuba, with the excep-

tion of one species (not found in these localities) from

California and Oregon.! Fourteen of the twenty-nine

species are undescribed and are here given the manu-

script names of Dr. A. H. Sturtevant.? Descriptions of

them are in press.

Most of the chromosome descriptions in the present

paper are taken from pedigree material, either first or

second generation from wild flies, and the results have

been checked up in such a way as to make it very improb-

able that serious errors have crept in. As mentioned in

previous papers the chromosomes of these flies stand out

with diagrammatic clearness when favorable figures are

secured; and since they are uniformly arranged in pairs

and are often of various sizes they offer admirable ma-

terial for a comparative study. This makes it possible

to classify the members of each chromosome group ac-

cording to their size and shape, and to assort the groups

into definite types according to their configuration.

DESCRIPTION or TYPES

The terminology used in describing the chromosomes

will be the same as that used in my previous paper (714).

3 The collections, of course, include only a fraction of the existing species

within these areas, to say nothing of those in surrounding regions. For

more detailed locality references see deseription of types.

2 This investigation would have been practically impossible without the

cooperation of many friends. In addition to Dr. Sturtevant, to whom I

owe all of the identifications, and cultures of several species, I am under

obligation to Professor F. S. Earle, Dr. Carlos de la Torre and Mr. C. T:

amsden for many courtesies shown to me while collecting in Cuba, and to

Messrs, L. L. Gardner and G. F, Sykes for Pacific coast material, including

D. obscura.

No.598] CHROMOSOME STUDIES ON THE DIPTERA 589

Certain recurring kinds will be distinguished in advance

and will be designated by name or letter i in the specific

descriptions. These are as follows:

ong, V-, U-, or dumb-bell shaped chivdlibavnive

attached to the spindle at the median constricted

portion (apex of the V).

r. Rod-like, straight chromosomes, approximately

half the length of V, attached to the spindle at

one end, and radially arranged in metaphase.

c. Short, curved chromosomes, differing from r only

in form and (probably) in having a median at-

tachment to the spindle.

m.® Minute, spherical or slightly elongate chromo-

somes, usually located in the center of the meta-

phase plate.

To these symbols must be added XX and XY, used to

designate the sex chromosomes wherever they have been

definitely identified. In some species they are straight,

in others V-shaped. Identification of the sex chromo-

somes and determination of the XY relations in the males

has been one of the most difficult features of the study,

owing to the extreme scarcity of spermatogonial and

spermatocyte figures; but it has been made in as many

cases as possible and has been of great usefulness in

comparing different groups.

The order or sequence in which the types are described

is a purely arbitrary one, and is not intended to indicate

any genetic relationship.

In order to avoid duplication of figures chromosome

groups which have previously been described and figured

are not reproduced here unless they are of especial in-

terest. The series of types as a whole may best be under-

stood from an examination of diagrams A to L, which

represent schematically, but accurately, the twelve main

types.

3 The term m-chromosome, borrowed from Wilson ( 05), is used here in

a purely descriptive sense, and is not intended to signify any relationship

with the m-chromosomes of the Hemiptera.

590 THE AMERICAN NATURALIST [ Vou. L

MME y: »%

z As ana Ho

Mis Sts fe

il. il. A. IN»

VY 7 YF i L

er am + «=»

2 2e Sf =:ss:s

wr: 1 cM NAA

Type A

Represented by

Drosophila ampelophila Loew. Cosmopolitan.*

(See Stevens, ’08, Figs. 57-60, 80-82, Metz, ’14,

Figs. 4 and 5, Bridwos 716, Figs. 1-4, Metz, 716,

Fig. 19.)

Drosophila amoena Loew. New York. (Metz, 14,

Figs. 1-3, Metz, 16, Figs. 13-16.)

Drosophila busckii Coq. New York. (Metz, 716,

Figs. 17 and 18.)

Drosophila bromeliae Sturtevant mss. Cuba.

(Fig. 1

Drosophila dimidiata Loew. Alabama. (Metz, ’16,

Fig. 20.)

4 With the exception of this species, which has been studied by several

investigators, localities cited are those į in which my cytological material has

been obtained.

No.598] CHROMOSOME STUDIES ON THE DIPTERA 591

i w “2 9;

Jk ae Fe N

s X M

RK SE Sd

PLATE I

Explanation of Figures

All figures were drawn with the aid of a camera lucida, using Zeiss 1.5 mm.

apochromatie objective and compensating ocular number 12, with tube length of

160 mm. The drawings are reproduced natural size. With the exception of

number 17 all are taken from sections cut 5 microns thick.

592 THE AMERICAN NATURALIST [ Vou. L

Drosophila flore Sturtevant mss. Cuba. (Fig. 2.)

Drosophila limbata Will. Cuba. (Metz, 716, Figs.

10-12.

Drosophila procnemis Will. Cuba. (Metz, ’16,

Figs. 34-36.

Drosophila quinaria Loew. New York. (Metz, 14,

Fig. 7.)

Drosophila robusta Sturtevant mss. New York.

(Figs. 3 and 4; Metz, ’16, Figs. 7-9.)

Drosophila saltans Sturtevant mss. Cuba.

Drosophila pallida Will. Cuba.

Scaptomyza graminum Fall. New York. (Fig. 5;

Metz, 16, Figs. 4-6.)

This is type I of my previous paper, and is the prevail-

ing type among the species studied, being represented by

thirteen of the total twenty-nine. It consists of two pairs

of long, V-shaped chromosomes, one pair of straight sex

chromosomes, and one pair of m-chromosomes.

Slight modifications of this type are found in certain

species, usually characterized by peculiarities in the sex

chromosome pair. In D. ampelophila it appears from

the work of Bridges (716) that the Y-chromosome, instead

of being straight as is the X-chromosome, is hook-shaped

or V-shaped, although never large enough to be confused

with the V-shaped euchromosomes. My observations

confirm those of Bridges in this regard with the exception

of two figures in which X and Y appear equal and

straight. Perhaps these are due to accident, but they are

entire figures and normal in other respects.

Fic. 1. Drosophila bromelie Sturtevant mss. .. Spermatogonium.

Fic. 2. Drosophila fore Stt. mss . Ovarian ce

Frcs. 3 and 4. Drosophila robusta Stt. mss., » spermatogonia

min

- 10. Scaptomyza adusta, ovarian ¢ ell.

Figs. 11 and 12. Dr rosophila neglecta Stt. mss., ovarian cells.

Fig. 13. Drosophila similis, ovarian cell.

IG. 14. Drosophila cardini Stt. mss., ovarian cell. The m-chromosomes

t evident in other =e of the same ovary.

IGs. 15 and 16. Cla PRONE ta nebulosa, ovari

Fic. 17. Drosophila repleta, variety b, ovarian a i Taken from an aceto-

carmine smear preparation

Fie. 18.

[=]

þa

Same, spermatogo nium, from a section

Fic. 19. Drosophila caribea Stt. ss., Ovarian cell.

Fic. 20. Same, spermatogonium

No.598] CHROMOSOME STUDIES ON THE DIPTERA 593

In D. amoena and S. graminum spermatogonial figures

indicate that X and Y are unequal and straight (Fig. 5,

and Metz, 714, Figs. 2 and 3); while in D. busckii, D.

flore, D. bromelie, D. robusta and D. limbata no in-

equality is evident. The other species have not been

examined with respect to spermatogonial groups. In D.

robusta the rod-like members (sex chromosomes?) ap-

pear to be hook-shaped and to be attached sub-terminally

to the spindle in much the same manner as the Y-chromo

some in ampelophila.

Another modification or sub-type is represented by D.

flore and D. bromelie, in which the m-chromosomes are

materially larger than in the other species. Indeed, those

of the former are so large as to suggest a transition be-

tween m-chromosomes and r-chromosomes.

Type B

Represented by

Drosophila earlei Sturtevant mss. Cuba. (Figs.

6-9.)

This interesting type consists of one short rod-like pair

and two large V-shaped pairs, one of which is much

longer than the other. No trace of m-chromosomes has

thus far been found in the fifteen or twenty figures I have

studied. Unfortunately I have not yet secured good

spermatogonial figures and am unable to identify the sex

chromosomes.

In many respects type B is of greater interest than any

other type of chromosome group I have studied, for it not

only contains the smallest number of chromosomes thus

far found among the higher flies but each of its three

pairs is conspicuously different from either of the other

two, making possible an individual identification of the

chromosomes not obtainable, with such a degree of cer-

tainty, in any other known species of Diptera. If a

genetic continuity of chromosomes be admitted there can

be no question that here each paternal chromosome as-

5 Two species of Culicide (Anopheles punctipennis and Culex pipiens)

also have three pairs of chromosomes. See Stevens, ’11

594 THE AMERICAN NATURALIST [ Vou. L

sociates with its corresponding maternal mate. (See

Metz, ’16, p. 251.)

Type C

Represented by

Drosophila ornatipennis Will. Cuba. (Metz, 716,

Fig. 21.)

Scaptomyza adusta Loew. New York. (Fig. 10;

Metz, 716, Fig. 22.)

Type C corresponds to type IV of my previous paper

(714), but the single species formerly referred to it has

been trahsferred to type E. Chromosome groups of type

C are composed of one large V-shaped pair, one long,

straight sex chromosome pair, two shorter rod-like pairs

and one small m-pair. In both species spermatogonial

figures show a noticeable inequality between X and Y;

thus identifying the sex chromosomes.

Type D

Represented by

Drosophila tripunctata Loew. New York. (Metz,

14, Figs. 21-26.)

Type D corresponds to type V of my previous paper

(714) and includes only one species. It differs from type

C in lacking the m-chromosome pair, and in sex chromo-

some relations. X and Y are apparently equal in size

and similar to the rod-like euchromosomes. Their iden-

tification is based solely upon. their precocious contrac-

tion in prophase (see Metz, 714, p. 52), and hence may be

held with some reserve, but from analogy with species

of types A and C it seems highly probable that the iden- —

tification is correct.

Many preparations have been made from material of

this species in an attempt to discover an m-chromosome

pair such as is found in most other species of Drosophila,

but although stocks have been obtained from several

localities and various methods of fixation have been used

no trace of the pair has been observed.

No. 598] CHROMOSOME STUDIES ON THE DIPTERA 595

Type E

Represented by

Drosophila melanica Sturtevant mss. (two varieties).

New York, Alabama. (Figs. 11, 12; Metz, 716,

Figs. 23-26.)

The chromosomes of this species resemble those of

type C, with the exception of one of the short pairs, which

is curved or U-shaped instead of straight. Such a differ-

ence in shape is apparently associated with a different

mode of attachment to the spindle, and seems to be a char-

acteristic feature. In my previous paper (714) D. mel-

anica was cited as an example of the type corresponding

to C of the present paper, and the curved shape of this

pair was not considered significant; but more recently I

have examined many additional figures and have become

convinced that the character is normal and sufficient to

distinguish the two types. The few spermatogonial

figures I have examined closely resemble those of female

groups and give no evidence of an unequal XY pair.

The two varieties? of D. melanica, although refusing to

hybridize, are very similar in external appearance and

indistinguishable in chromosomal characters.

Type F

Represented by

Drosophila virilis Sturtevant mss. New York City.

(Metz, 714, Figs. 11-13, Metz, 716, Fig. 2.)

Drosophila similis Will. Cuba. (Fig. 13.)

Drosophila ramsdeni Sturtevant mss, _ Cuba. (Metz,

14, Fig. 10, Metz, ’16, Fig. 3.)

Drosophila rA Sturtevant mss. Cuba. (Fig.

14.)

Drosophila Hodali Sturtevant mss. Alabama.

Drosophila repleta Woll., variety a.’ Cuba, Texas.

(Metz, ’14, Figs. 8 and 9.) |

This type (Type II of the previous paper) differs from

C in having two pairs of rod-like chromosomes in place of

the large V-shaped pair, and from type A in possessing _

6 For, discussion of these see eet m ty Dr. A. H. Sturtevant. A

T See also t :

596 THE AMERICAN NATURALIST [ Vou. L

four rod-like pairs in place of two V-pairs. Next to type

A this type is of most frequent occurrence, being repre-

sented by six of the twenty-nine species.

Spermatogonial figures have been examined in only one

of these species, D. virilis, and here no conspicuous in-

equality between X and Y is to be seen. One pair of

chromosomes appears to be larger in nearly all figures of

either sex, and a slight difference in length between the

two members of this pair may be seen in some male

figures, but it may be purely accidental.

Type G

Represented by

Drosophila funebris Fabr. New York, California,

North Dakota. (Metz, 714, Figs. 14-17, Metz,

716, Figs. 27-33.)

This interesting type (type III of the previous paper)

is apparently a modification of type F, but differs from

it in the relative proportions of all of the chromosomes.

The m-chromosomes are so minute in most specimens as

searcely to be visible, and for this reason were entirely

overlooked at first. Their conspicuousness doubtless

varies with the amount and kind of stain, and with the

fixative used, but even after making full allowance for

this there can be no doubt that the pair is much smaller

here than in most other species. Otherwise the type is

characterized by the smaller size of the short, rod-like

chromosomes and the greater length of the longest (sex

chromosome?) pair. As in the preceding case no con-

spicuously unequal XY pair is to be found in the males,

although a noticeable difference between the two large

chromosomes may be seen in some of the figures.

Type H

Represented by

Cladochaeta nebulosa Coq. Cuba. (Figs. 15 and

16.) 3

This species—the only known member of the genus—is

8 Fig. 14 (Metz, 14) and Fig. 27 (Metz, ’16) are from the same cell;

the latter drawn after the m-chromosomes were discovered.

No.598] CHROMOSOME STUDIES ON THE DIPTERA 597

included in the present review because of its obvious rela-

tionship to the true Drosophilas. The type of chromo-

some group which it represents is the only one of the

twelve not thus far found in some species of Drosophila,

and its general similarity to some of the Drosophila types

is marked. Female groups consist of three similar pairs

of long, V-shaped chromosomes and one small pair of

m-chromosomes. Unfortunately the species breeds very

poorly in confinement and no male preparations were

secured. It is almost certain, however, that one of the

long pairs is the sex chromosome pair.

; Type I

Represented by

Drosophila repleta Woll., variety b. New York,

Massachusetts, California. (Figs. 17 and 18;

Metz, 716, Figs. 39-41.)

In my 1914 paper D. repleta was referred to the type

corresponding to F of the present study, but it is now

evident that two very similar but distinct varieties of the

species occur, characterized, among other things, by de-

cidedly different sex chromosomes. In one, the sex

chromosomes are short and rod-like in the female and

presumably so in the male, while in the other they are

long and V-shaped in the female and markedly unequal in

the male. The latter represents the present typeI. The

difference between the two may be readily appreciated

by an examination of diagrams F and J. Although it

relates only to the sex chromosomes it is very striking in

the females, and easily separates the two varieties into

distinct types. The fact that the two forms can not be

induced to hybridize lends support to the chromosomal

evidence of their distinctness, but externally they are

astonishingly similar and are referred to the same Aaah

by Sturtevant. d

9 It may be noted that these are not the ‘‘light and dark’’ varieties de-

scribed by Sturtevant (’15), both of which belong to type I.

598 THE AMERICAN NATURALIST [Vou. L

Type J

Represented by

Drosophila obscura Fall. California and Oregon.

(Metz, ’16, Figs. 44-50.)

Ovarian cells of D. obscura contain three rod-like eu-

chromosome pairs, one small m-chromosome pair and one

very long, V-shaped sex chromosome pair (diagram

J, 2). In the male the sex chromosomes are very dis-

similar, Y being straight and only about half as long as X.

Type K

Represented by .

Drosophila affinis Sturtevant mss. Alabama.

(Metz, 716, Figs. 42 and 43.)

In general this type resembles the last, but differs in

having two S-, or hook-shaped pairs in place of rod-like

ones. Apparently this peculiar shape is due to a sub-

terminal attachment to the spindle, although I have been

unable to get figures actually demonstrating the attach-

ment. In some cases one or both pairs extend radially

from the center of the figure as if they were attached

terminally, but usually their position is characteristically

that described above. In any event the two pairs are

readily distinguished from any others of the group, un-

like those of D. obscura.

Type L

Represented by

Drosophila caribea Sturtevant mss. Cuba, Panama.

(Figs. 19 and 20.)

This type is radically different from any of those de-

scribed above, and like the two preceding is represented

_ by only one species. Female (ovarian) groups are com-

posed of four long V-shaped pairs of chromosomes, one,

of which is shorter than the other three. In the male one

pair is conspicuously unequal, much as in types J, J and

K, but I have been unable to determine with certainty

whether this is the small pair or one of the large ones.

It is represented as a large one in the diagram (L, 8).

No.598] CHROMOSOME STUDIES ON THE DIPTERA 599

Several of the species contained in this survey have

been, or are being used in genetical studies. With the

exception of the well-known D. ampelophila these are in-

cluded in the following list, together with references to

literature dealing with them, so far as known to me:

D. repleta, type H., Ria 15, Hyde, 19.

D. affinis, type J. Hyde, 715.

D. tripunctata, type D. Metz and Metz, 715.

D. virilis, type F. Metz and Metz, 715.

D. similis, type F. In press.

D. obscura, type I. In press.

LITERATURE CITED

Bridges, C. B. 1916. Non-disjunction as Aapa of the Chromosome Theory

of Heredity. Genetics, I, p.

Hyde, R. R. .1915. The Origin of a New p Color in Drosophila repleta,

and Its Behavior in Heredity. AMER. NAT. 83.

1915.. A Wing Mutation in a New Species of Drosophila, Thid., p. 185.

Metz, C. W. 1914. Chromosome Studies in the Diptera I. A Preliminary

Survey of Five Different Types of ee ae Groups in the

Genus Drosophila. Jour. Exp. Zool., 17, p. 4

Metz and Metz. 1915. Mutations in Two Species of Drosophila. AMER.

Nam, 49, p. 187.

AT of Chromosomes in a oer and Its Signifi-

nee. Jour. Exp. Zool., 21, p.

Stevens, N. Me 1908. A Study of the ee ee of Certain Diptera, ...

Jour. Exp. Zool., 5, p. 359.

Stevens, N. M. 1911. Further Studies on Heterochromosomes in Mos-

quitoes. Biol. Bull.,

Sturtevant, A. H. 1915. A Sex- linked Gharda in ee al repleta.

Amer, Nart., 49, p. 190.

Wilson E. B. 1905. Studies on Chromosomes. II. Jour. Exp. Zool., 2, p.

50.

THE SHAPE OF THE STERNUM IN SCORPIONS

AS A SYSTEMATIC AND A PHYLO-—

GENETIC CHARACTER

ALEXANDER PETRUNKEVITCH, Ph.D.,

` Assistant PROFESSOR OF ZOOLOGY IN THE SHEFFIELD SCIENTIFIC SCHOOL

(From the Osborn Zoological Laboratory of Yale University)

Ir is generally recognized that the shape of the sternum

furnishes one of the important characters for the distinc-

tion of families in recent scorpions. The small family

Bothriuride is the only one in which the sternum is com-

posed of two transverse plates. This family includes

seven genera, six of which occur in South America, while

the seventh (Cereophonius) is an inhabitant of South

Australia and contains a single species. The families

Scorpionide, Vejovide, Cherilide and Chactide have a

distinctly ‘‘pentagonal’’ sternum with more or less par-

allel sides. The Cherilide belong exclusively to the old

world. The Chactidæ are divided into three subfamilies,

the European Euscorpiine and the neotropical Megacor-

mine and Chactine, which have Mexico for their northern

limit of distribution. The Vejovide are unevenly dis-

tributed between the Old World and the New. One of the

eight genera composing this family is found on the shores

of the Mediterranean (Iurus, with a single species I. du-

foureius), another (Scorpiops) with about eight species

in East India. Of the remaining six genera, two occur

in South America, while the other four belong to the

southern and western United States and to Mexico. The

family Scorpionidx, to which some of the largest scor-

pions belong, has representatives from various countries

and regions. It is usually divided into five subfamilies.

Of these the Urodacin» are Australian; the Scorpionine

Asiatic and African; the Hemiscorpionine Asiatic; the

600

No. 598] THE STERNUM IN SCORPIONS 601

Ischnurine African, Asiatic and neotropical (Opistha-

canthus elatus is the single neotropical species) ; the Dip-

locentrine neotropical and Asiatic (a single genus with

two species from eastern Asia). ;

All scorpions belonging to the large family Buthidæ

have a distinctly ‘‘triangular’’ sternum with converging

sides and truncated apex. The family is naturally di-

vided into two large subfamilies. Of these the Buthine

may be regarded as belonging to the Old World, since

of its 14 genera a single genus and species (Ananteris

balzani) is found in South America. The subfamily Cen-

trurine includes four genera. The genus Isometrus is

characteristic of the Old World but its commonest species,

I. maculatus, is cosmopolitan and occurs in Florida, Ha-

walian Islands, South America, ete. Zabius is South

American. Titius is neotropical, although one species, T.

floridanus, occurs in southern Florida. The largest

genus, Centrurus, is represented by some of the com-

monest species in the southern United States and the sub-

tropical and tropical America.

Let us fix our attention for a moment on the distribu-

tion and characters of two genera of scorpions common

to the United States. One is Vejovis (of the family

Vejovide) and is represented in this country by six spe-

cies; the other is Centrurus (of the family Buthide) and

is represented by seven species. Vejovis belongs more to

the southwest and west. It is distributed through Cali-

fornia, Nevada, Utah, Arizona, New Mexico and Texas

and extending northward into Idaho and Nebraska. A

single species, V. carolinus, is found in the southeast. It

occurs as far north as South Carolina and spreads south-

ward to the Gulf states and Texas. Except possibly this

species, the other species of Vejovis occur also in Mex-

ico where the genus is represented by four additional

species which do not occur in the United States. I have

besides, in my private collection, a new species of Vejovis

from Terra del Fuego. Centrurus belongs more to the

southeast. A single species (C. ewilicauda) occurs in

602 THE AMERICAN NATURALIST [Vou. L

California. C. nigrescens, a variety of the more south-

ern C. gracilis, has been reported from Texas. Four

species occur in Florida, but of these C. gracilis and C.

margaritatus belong to a more southern fauna, the latter

being the most common scorpion of Mexico and Central

America. One species, C. infamatus, has practically the

same distribution as Vejovis carolinus, spreading north-

ward into South Carolina and southward into Texas and

Fic. 1. Sternum, genital opercula, combs and coxæ of an adult Centrurus

insulanus Thorell from Jamaica, W. I. The sternum is “ triangular ” with con-

verging sides, but still shows traces of its original, pentagonal shape.

northern Mexico. C. infamatus is the common scorpion

of the southeast and south. Several other species of Cen-

trurus are characteristic of tropical North and South

America.

No. 598] THE STERNUM IN SCORPIONS 603

Vejovis carolinus is a small scorpion, its length not ex-

ceeding 34 mm. The width of the carapace at the pos-

terior edge is slightly less than the length. The center of

the eye tubercle is about 2 of the entire length of the

carapace from its anterior edge. The sternum is pentag-

onal; the comb has 13-14 teeth; the central and inner

rows of plates in the comb are beadlike. The fingers are

rather short, being either as long as or only slightly longer

than the hand.

Centrurus infamatus is usually about 45 mm. long, but

large specimens measure up to 70mm. The width of the

carapace at the posterior edge is equal to or even slightly

exceeds its length. The center of the eye tubercle is

about 2 of the entire length of the carapace from its an-

terior edge. The sternum is triangular. The comb

usually has 18-19 teeth, although their number may reach

25. The central rows of plates is not beadlike, but com-

posed of five plates the limits between which are difficult

to ascertain. The inner row is beadlike. The fingers are

rather long, being more than 14 times as long as the hand.

Let us now consider the Paleozoic scorpions. The

sternum of the Silurian Proscorpius osborni Whitfield is

unfortunately not preserved. Its nearest European rela-

tive, the Silurian scorpion Paleophonus hunteri Peach,

has, according to. Pocock, a pentagonal sternum. The

sternum of the carboniferous scorpions is fairly well-pre-

served in many specimens. The family Isobuthide dif-

fers from all other fossil and recent scorpions in the posi-

tion of the fourth pair of coxe which abut against the

genital opercula. The sternum is either triangular

(Paleobuthus), rhomboidal (Isobuthus) or oval (Eobu-

thus). The family Cyclophthalmide has a ‘‘pear’’-

shaped sternum, the family Eoscorpionide a distinctly

pentagonal one with parallel sides. If the pear-shaped

and rhomboidal impressions of sterna do not owe their

shape to poor preservation or displacements, then these

types of sterna must have disappeared completely, as has

the type of triangular sternum found in Isobuthide. Of

604 THE AMERICAN NATURALIST (Von. L

preserved fossils there remain then only the Silurian

Paleophonide and the carboniferous Eoscorpionide hav-

ing a sternum and arrangement of coxæ similar to that

in recent scorpions. But the Silurian scorpions pos-

sessed other characters of their own and have either dis-

appeared completely or perhaps have changed gradually

into carboniferous forms. In the absence of Mesozoic

fossils any attempt to trace relationships between recent

and Paleozoic scorpions can be only conjectural. Thus

in my ‘‘ Monograph of Paleozoic Arachnida’’ I arrived

at the conclusion (p. 26) that ‘‘ the family Eoscorpion-

ide shows many relations to the recent Scorpionide and

Vejovide and represents probably two or three families

thrown together for lack of distinctive characters.’’ In

formulating this opinion I was guided chiefly by the shape

of the sternum, in several cases remarkably well pre-

served. Since that time I have made an observation, in-

_ significant in itself, but one which affords an insight into

the past history of recent scorpions possessing a trian-

gular sternum and suggests a closer relationship between

the Eoscorpionide and the Centrurine than between the

former and the Vejovide. This observation was made

by pure chance. While studying the problem of seg-

mentation in Arthropods, I was examining a frontal sec-

tion through a recently born Centrurus insulanus from

Jamaica (Fig.2). To my amazement the sternum proved

to be beautifully pentagonal. An error of identification

was excluded. I myself collected the material and pre-

served the adult females with the young carried on their

back. Yet if objection should be raised, a final proof is

offered by the fact that late embryos, too, have a pentag-

onal sternum and such embryos are easily obtainable

from adult, gravid females. (Scorpions are without ex-

ception viviparous.) The young of Centrurus infamatus

also have a pentagonal sternum, as have probably the

young of all other species of the genus Centrurus. It is

strange that no one has noticed this before, since there

must be dozens of specimens in every museum. I find

No. 598] THE STERNUM IN SCORPIONS 605

an interesting confirmation of my observation in Fig. 10

of McClendon’s paper on the nervous system of Cen-

trurus infamatus.: The adults of his material were iden-

tified for him by no Jess an authority than Kraepelin, yet

in the figure in question McClendon draws a distinctly

pentagonal sternum in a surface view of a late embryo.

The case is the more interesting because McClendon him-

self is unaware of the value of his observation, nor is

Frontal section cigs to the ventral surface showing sternum,

ery young Centrurus insulanus. The

gen iat real: combs and co: a

sternum is distinctly pi g "in parallel sides.

there the slightest reference to it in the text. He simply

drew the sternum as he saw it, without so much as men-

tioning it. An examination of the adults of Centrurus

insulanus (Fig. 1) as well as of other species of Cen-

1 Biol. Bull., Vol. 8, 1904, p. 51.

606 THE AMERICAN NATURALIST [ Vou. L

trurus, Shows that their sternum retains throughout life

traces of its origin from a pentagonal prototype. Only

the triangular apex is comparatively small and more or

less hidden between the cox, and the sides of the ster-

num are strongly convergent, not parallel. The sternum

in adults of various species of Buthine of the Old World

shows also a small triangular apex hidden between the

Fig. 3. A very young Centrurus infamatus (C. 1. Koch) — (carolinianus

Palisot de Beauvais) drawn with the aid of the Abbe apparatus. The sternum

is pentagonal. The hand has much longer fingers than would be the case in an

adult of the same species. Abdomen distended by yolk.

. Fie. 4. scorpius typicus Petrunkevitch, from Mason Creek, Illinois.

Specimen No. 37986 of the U. S. N. M. (Fig. 6 of the Monograph.) :

cox, and in the absence of material I venture the predic-

tion that the embryos will be proved to possess a pentag-

onal sternum. r

Keeping this in view, let us now compare Centrurus

infamatus with Eoscorpius typicus of the Pennsylvanie

from Mason Creek, Illinois, a rather well-preserved Palæ-

ozoie scorpion of approximately the same size. Fig. 4

No. 598] THE STERNUM IN SCORPIONS 607

is taken from my monograph and represents the holotype

of Eoscorpius typicus as it appears on the reverse. Fig.

3 is a careful drawing of a very young Centrurus infama-

tus made with the aid of an Abbe apparatus. This figure

is much more enlarged than is Fig. 4, to facilitate com-

parison. The shape of the abdomen, as is well known,

is the least constant character and we will leave it out of

consideration. To avoid criticism, however, I would state

that the great distension of the abdomen in the young

Centrurus is due to the presence of embryonic yolk.

When this disappears, the abdomen becomes much thin-

ner. In older specimens the distension of the abdomen

is frequently due to the growing embryos. Yet, in dis-

secting what I supposed to be gravid females, I was sur-

prised to find no embryos in them and only small ovaries.

The fact is that the distension of the abdomen is also often

due to the condition of the liver.

The carapace has a fairly constant shape and is similar

in both scorpions. For figure and measurements I refer

to my monograph. I do not reproduce them here be-

cause the carapace of Vejovis, except for its size, is also

of similar shape. The rows of granules on the caudal

segments are not sufficiently well preserved in the fossil

specimens to allow of a conclusion as to their exact num-

ber. On the other hand, the transverse row of granules

at the anterior end of the first caudal segment, present in

Eoscorpius and wanting in the adult Centrurus, is clearly

defined in the young. The most interesting character is

represented in the hand with its fingers. As a rule the

ratio between the length of the fingers and that of the

hand is a fairly constant one for each species. In the

holotype of Eoscorpius typicus it is approximately 2:1;

in the adult Centrurus infamatus it is 1.6:1, but in the

late embryos and just born young it also approximates

2:1. With other words, Centrurus developed from an

ancestor with relatively longer fingers and the trend of

evolution was toward reduction in their length.

One character presents a difficulty. This is the comb.

608 THE AMERICAN NATURALIST [Von L

Specimen No. 37987 of the U. S. Nat. Mus. of Eoscorpius

shows a comb which is very broad at the base, and has,

apparently, a single median plate, a beadlike „inner row

and 19 teeth. I identified this specimen as Foscorpius

typicus for the reason that ‘‘the general appearance of

the specimen, the shape ‘‘of the tergites, especially of the

seventh,’’ strongly resembled the holotype. The speci-

men is incomplete and about twice as large as the holo-

type. Perhaps No. 37987 is after all of a different spe-

cies. The shape of the comb in recent species of Cen

trurus is not always the same as in infamatus. The comb

in C. junceus and C. agamemnon is twice as wide at the

base as in the middle.

Taking all characters and the geographic distribution

into account, we can not fail to notice the greater simi-

larity between the young of Centrurus infamatus and

Eoscorpius typicus than between the latter species and

Vejovis. What advantage, if any, Centrurus has derived

from the shortening of the fingers and the change in the

shape of the sternum, is a totally different question which

may possibly be answered by studying the functions and

uses of these organs in recent species.

THEORIES OF HIBERNATION

ANDREW T. RASMUSSEN

CORNELL UNIVERSITY

AN examination of the literature on almost any par-

ticular natural phenomenon often reveals the fact that

many different theories have been advanced to explain it.

Some of these explanations may be mere opinions based

upon no or but few scientific facts. One is also frequently

struck with the immense literature that has been produced

and the great gap that still intervenes between the ac-

cumulated facts and a clear understanding of the proc-

esses which they aim to elucidate, even after more than

a hundred years of experimental work, which has usually

been preceded by a much longer period of speculation by

the great thinkers of the past. So that while we con-

gratulate the last few generations upon the rapid growth

that has been made in scientific knowledge, there yet re-

main phenomena that are almost as unintelligible to-day

as they were a hundred years ago—the most earnest and

often tedious experimentation and observations of sev-

eral generations having shed but little light on the fac-

tors and mechanism involved.

The extremely interesting fact of hibernation (called

‘‘Winterschlaf’’ by the Germans, ‘‘sommeil hivernal’’ by

the French and ‘‘letargo’’ by the Italians) illustrates well

the above point. As is well known, during this dormant

state the vital processes are greatly reduced. The changes

that occur are especially marked in certain mammals,

since they apparently undergo a rather sudden transfor-

mation from the warm-blooded (homoiothermal) type to

the cold-blooded (poikilothermal) type. In the latter

state such mammals are able to endure cold, deprivation

of food, confined air, effects of many drugs, and other

conditions that would be fatal at other times. Naturally

609

610 THE AMERICAN NATURALIST [Vou. L

such profound physiological changes, in some respects

almost as striking as the latent vitality of the seeds of

plants and the spores of lower organisms, has aroused the

attention of a great many observers. In fact, the liter-

ature on hibernation dates back to the time of Aristotle

(384-322 B.c.), though real experimental work for the

purpose of understanding the nature and cause of this

torpid state, commenced with Conrad Gessner’ (1551).

From that date to the present there has accumulated a

vast amount of data, the bibliography of which is now

very accessible, due to the extensive works of Raphael

Dubois,? published in 1896, and of Osvaldo Polimanti,®

published in 1912.

As the exciting cause of so-called winter-sleep, cold has

naturally received by far the greatest share of attention.

A rapid survey of the subject shows that much difference

of opinion has existed in regard to the manner in which

cold acts and what other factors are involved. Buffon*

(1749) and Lacépéde® (1829) thought that the blood

simply becomes cold when the small amount of heat pro-

duced by hibernating animals is not aided by the sur-

rounding temperature. The cold blood then produces the

changes characteristic of torpidity. Spallanzani® (1787),

however, considered that he had experimentally demon-

strated that the cold acts on the solid tissues of the body

and not on the blood. According to him the lethargy is

due either to the stiffening of the muscles or to the deple-

tion of the cerebral blood vessels. On the other hand,

Alibert? maintained that the cold diverts the blood from

the periphery to the vessels of the brain and the resulting

congestion causes torpor. But Serbelloni® (1866) claims

` to have found the vessels of the brain nearly empty in the

case of three marmots in full hibernation. Haunter? (1775)

and Serbelloni explained that the cold causes the animal _

to lose its appetite and in the absence of hunger, which 1s

a stimulus, the animal retires.

A long list of authors, Daubenton”? (1760), Geoffroy,’

Cleghorn,'? Allemand,'? Carlisle* (1805), Barkow’®

No. 598] THEORIES OF HIBERNATION 611

(1846) and others, have emphasized also the necessity of

confined air or diminished respiration, Cleghorn and

Allemand maintaining that this is the principal cause.

Reeve'® (1809) said that such a condition favors winter-

sleep, while Bert (1868) first concluded that lack of

oxygen and then later!® (1873) that the accumulation of

CO, in the surrounding air might be the cause of dor-

mancy. Mangili!® (1807), however, denied that vitiated

air has anything to do with this torpid state and Dubois”

(1896) says that confined air is not necessary, for animals

hibernate perfectly in well-ventilated places.

Marshall Hall? (1832) believed that the cold caused

ordinary sleep, which diminishes respiration, and less

heat is produced. Lessened respiration causes the blood

to lose its arterial character and hence its power to stim-

ulate the heart. The heart, however, changes its irrita-

bility so that it does not stop. This change in the irri-

tability of the heart, then, is the important factor in

hibernation. To him winter-sleep is something entirely

different from the torpor produced by cold. To Edwards*?

(1824) and Legallois?* (1824) sleep and cold are so bound

up with heat production that a failure to keep up the body

temperature causes torpidity to ensue.

Throughout the literature of the last one hundred years

there is a strong tendency to consider hibernation as dif-

fering from ordinary daily sleep only in degree. Ed-

wards?? (1824), Dugé?* (1838), Hall?! (1832), Blandet?

(1864), Patrizi? (1894), Dubois?” (1896, 1910), Brunelli?*

(1902), Claparède? (1905), Allen Cleghorn*® (1910) and

Salmon?! (1910) make definite statements regarding the

striking similarity between ordinary daily sleep and

hibernation. Gemelli*? (1906) used the facts obtained by

him from hibernating marmots, in disproving Salmon’s

theory of sleep. Indeed, it has been the hope of many of

the students of hibernation to be able to throw some light

on the process of diurnal sleep in man and other animals,

by a study of what they have considered to be merely an

extreme example of this physiological condition. The

612 THE AMERICAN NATURALIST [Vou.L

discussions on sleep that appeared in the British Medical

Journal in 1913 and the comprehensive treatise by

Pieron®* (1913) on the physiological problem of sleep,

clearly indicate how little has been accomplished in this

direction. Buffon (1749), on the other hand, thought

that ordinary sleep and hibernation were something en-

tirely different. Monti** (1905) even now believes that

these two forms of sleep have entirely different physio-

logical meanings and that hibernation in its phylogenetic

study should be compared with the dormancy of lower

forms, as well as with ordinary sleep.

In reply to the question asked by the French Academy

of Science over a hundred years ago as to the cause of

this lethargy and why it pertains to certain animals,

Saissy** (1808) stated that the cause fundamentally is to

be found in certain anatomical peculiarities such as the

enlarged character of the heart, central blood vessels,

thorax, abdomen and cutaneous nerves, and the smallness

of the peripheral vessels and lungs. To these he also

added as important features the liquid quality of the

blood and the sweetness of the bile. The diversion of the

blood from the surface towards the center of the body, as

a result of the external cold, dilates the heart and blood

vessels of the thorax, and this interferes with respiration,

thus decreasing heat production. As a consequence the

animal becomes cold and numb. Prunelle** (1811), Bar-

kow (1846), Serbelloni® (1866) and Blandet?® (1864)

similarly believed in the importance of such—largely

imaginary —morphological features.

Many investigators have associated hibernation with

the nervous system. Claude Bernard’? (1855-76) thought

that the cold acts on an unusually well developed periph-

eral nervous system, and by slowing respiration cools the

body. This is a loss of stimulus to the heart and muscles

and torpor results. Reeve'® (1809) stated that cold acts

on a special organization of the nervous system, which

causes diminished respiration, etc.; while Quincke®

(1882) interprets the facts he and others have observed,

No. 598] THEORIES OF HIBERNATION 613

in connection with the marmot, as indicating the existence

of a heat center in the brain through whose influence on

the various organs of the body, metabolism and heat regu-

lation are so affected as to produce winter-sleep. The

altered respiration and circulation are secondary results.

Dutto? (1896) is also inclined to believe that hibernation

strictly depends upon the regulative influence of the nerv-

ous system upon metabolism and thermogenesis. He

further considers that the marmot has the power to emit

more heat than has the rabbit, so that torpor may be

based upon the difference in the power of the integuments

of hibernating and non-hibernating animals to lose heat.

Merzbacher*? (1904), after reviewing much of the liter-

ature dealing particularly with the rôle of the external

temperature, food and the nervous system in the produc-

tion of winter-sleep, concludes that the external cold is

only a secondary aid. Cold, like abstinence from food,

immobility, slower respiration and lack of oxygen, simply

makes it easier for the animal to cool off and remain cold,

and tends to make the sleep more profound. The essen-

tial characteristic of the hibernating animal, as compared

with the non-hibernating animal, according to him, is its

ability to change at a rather definite period and in a com-

paratively short time from the homoiothermal to the

poikilothermal type and again at the end of hibernation

to return rather abruptly to the former condition. The

explanation of both of these alterations, he thinks, is prob-

ably to be found in a certain nervous mechanism in the

mid-brain and medulla which is capable of intluencing res-

piration, circulation and metabolism, and, in short, the

production and loss of heat. The other changes charac-

teristic of the lethargy are natural consequences of and

adaptations to the hypothermic and hypofunctional con-

ition.

In addition to other internal factors there is, accord-

ing to Barkow'® (1846), a special susceptibility to the ex-

ternal cold due to a rather primitive organization of hi-

bernating animals. Noè“ (1903) thinks that a primitive

614 THE AMERICAN NATURALIST [Von. L

structure of the organism is the important cause of the

lethargy; but it acts as a mechanism of economy by in-

creasing the resistance of the animal to cold, rather than

to starvation, and thus prevents histolysis from reaching

a dangerous point. An inefficient heat-regulating mech-

anism has been considered the true explanation of winter-

sleep by such men as Dugés** (1838), Marés*? (1889), and

Polimanti** (1904). Simpson** (1911) in this laboratory

has shown that the woodchuck can not be said to ever

have a normal temperature in the sense that a homoio-

thermal animal has. Merzbacher*® (1904) cites many

cases similarly indicating the weak thermogenic organi-

zation among winter-sleepers. Recently Polimanti*

(1914) has explained his views concerning this labile

thermogenic organization. To him it is due to the fact

that at some remote period all animals then existing

periodically fell into lethargy. With evolutionary de-

. velopment most mammals and all birds lost this ability.

Hibernating animals, however, are still able to return to

this cold-blooded type when the heat-producing reflexes

fail, which they are apt to do when the cold becomes ex-

treme. Marés*® (1913) holds fundamentally this same

view—a view he advanced in 1889. He says that the

cause of hibernation is in the organism itself. He re-

gards the facts presented by Pembrey*? and Babak** and

others concerning the poor heat-regulating mechanism of

the newborn in man and other animals, as well as those by

Merzbacher* on the return of the nervous system to a

more segmental type during winter-sleep, as strong evi-

dence in favor of the theory, and as indicating that hiber-

nating animals merely revert to a more primitive type in

which there is no specific sensibility to the outer cold, i. e.,

in which no specific heat-regulating reflexes are called

forth by the external temperature. He further thinks

that since the weakness is in the nervous system, it ought —

to be possible to bring about some of the conditions of

torpor by means of hypnotic suggestions. He and Hel-

lich"? (1889) succeeded by this means in getting a fall of

No. 598] THEORIES OF HIBERNATION 615

2.5° C. in the body temperature of a hysterical woman.

Others have gone much farther in this regard. Thus

Liébeault** (1866) and Forel” (1877), especially the latter,

consider hibernation similar in nature to hypnotic sleep.

To Marés, however, the initial cause of winter-sleep is the

ability of the nervous system to loose its specific sensitive-

ness to the external cold. This sensitiveness, he thinks,

does not belong to the fundamental properties of the

nervous system, since it is not found in the young unde-

veloped animal. It is a property acquired slowly onto-

genetically just as it was slowly acquired phylogenetically

by the two highest classes of animals. A similarity be-

tween the hibernating and fetal states was noted long ago

by'Pallas®* (1778), Prunelle** (1811), Tiedermann™ (1815)

and Edwards?? (1824). Tiedermann claimed that in both

states there is merely a vegetative existence, hardly any

appreciable difference between the appearance of the

venous and arterial blood, much serum and little clot

when the blood coagulates, a low body temperature, an

enlarged thymus (he included the hibernating gland as

part of the thymus) containing a fluid, and a secretion of

bile. He therefore considered winter-sleep as a periodic

return to a fetal state. Pembrey and Hale White®®

(1896) regard the evolution of hibernation, not as the

acquisition of a new power, but as a retention of one al-

ready present, as is evident from the condition of young

mammals and birds in whom the heat-regulating power

is inefficient.

Many observers have questioned the value of cold as a

factor in the production of this dormant state. Quincke**

(1882) thought that rest and an appropriate temperature

generally, though not always, cause torpor, and yet he

said: that there seems to be some relationship between

degrees of lethargy and external temperature. Blandet?

(1864) considered that winter was only occasionally, if at

all, the cause; while Horvath®® (1872-81), with whom

Bunge (1901) seems to agree, said that hibernation is

not sleep at all and that winter has nothing to do with it.

616 THE AMERICAN NATURALIST [ Vou. L

Hahn® (1914) concludes that the torpid condition is not

dependent upon cold weather, although his thirteen-lined

ground squirrel usually hibernated with each cold spell

and woke up with the return of warm weather. Experi-

mentally it was early shown that cold will not induce

typical lethargy. Thus Buffont (1749) in the case of the

hedgehog, Daubenton®® in the hamster, Hunter? (1775)

in the dormouse, Mangili?® (1807) and Bossi® in the mar-

mot, Horvath®* in the spermophile and hedgehog and

Marés?? (1892) in the spermophile, have failed to induce

true hibernation by exposure of the animal to low temper-

atures. Saissy®® (1808) is supposed to have produced

winter-sleep by continued cold and confined air; but like

some other reported cases of artificially produced torpor,

it is not clear that the experimentally produced state was

the same as true hibernation. Sacc! (1858) after eight

years of observation on the marmot could see no relation-

ship between the condition of the atmosphere and winter-

sleep. Mills*? (1892) found that while bats could be

worked like a machine by varying the temperature, mar-

mots, on the contrary, showed a surprising indifference

to the surrounding temperature. Berthold®* (1837)

claims that dormice became dormant in a room kept warm

(16° C.) all winter, though torpidity was delayed two

months. Merzbacher*®: (1904) mentions similar experi-

ences of his own with a bat, as well as several other com-

parable cases. Mangili!® (1818) saw a dormouse fall into

lethargy in the month of June and not wake up till the

middle of July. Forel®? (1887) records that two dormice

which remained awake and active all winter, became tor-

pid in May and remained in this condition till August in

spite of the great heat. Marés*? (1892) found that some

spermophiles and hamsters may hibernate in September

at 16° C. while others remain awake all winter although

the thermometer falls below zero. Hence he concluded

that cold does not cause winter-sleep. Valentine®* (1857),

Horvath®* (1881) and Quincke®* (1882) have observed

marmots become dormant during the summer. Hence

No. 598] THEORIES OF HIBERNATION 617

Pembrey® (1898), while recognizing that want of food

and cold seem to be the most important factors in hiber-

nation, says that some other condition yet unknown is

necessary to explain such lethargy during the summer.

As a result of the uncertain action of cold, certain other

external conditions have been considered the real excit-

ing cause of hibernation. The food factor was empha-

sized by Mangili!® (1807), who believed that neither cold

nor vitiated air has anything to do with the production

of this torpid state. He thought fasting was necessary

because, of several animals under the same external con-

ditions, those animals that were fed did not become dor-

mant, while the non-fed ones did. Marshall Hall?! (1832)

stated that the lack of food predisposes the animal to

torpor by rendering it more susceptible to cold. Sace*?

(1858) concluded that, while he could see no relationship

between the atmosphere and torpidity, he could see some

connection between the fatness of the animal and the

length and profoundness of winter-sleep. He, therefore,

concluded that obesity, in connection with fatigue, is the

cause of hibernation. Claparéde®® (1905) and Forel?

(1887) think that the amount of fat may be an important

factor, while Beretta®® (1902) opposes this idea. Simp-

sonë? (1912) finds that feeding woodchucks greatly inter-

feres with winter-sleep, at least in captivity. Albini®*

(1901) in case of the marmot, and Reeve’* (1809) in con-

nection with dormice and hedgehogs, also confirm the ob-

servation of Mangili on the rôle of food in preventing

hibernation. Yet it appears that these animals (mar-

mots) may go into winter sleep while plenty of food is

available. Thus Mills*? (1892) found that during the

winter of 1890-91 a marmot hibernated in a cage provided

at all times with plenty of food; but during the two fol-

lowing winters two other marmots, kept in the same room

and in the same cage under similar conditions, did not

hibernate at all, though the temperature got low enough

to freeze the water in the cage. It is also a common ob-

servation that some of these animals naturally retire

618 THE AMERICAN NATURALIST [ Von. L

while food is plentiful. Allen Cleghorn? (1910) ques-

tions the lack of food as a factor in producing lethargy

because spermophiles and marmots hide away for winter

when their food supply is at its best. In British Colum-

bia he finds that these animals retire a month earlier in

the lowland than at the timber line because, he thinks, in

the latter region they have not had time to acquire enough

fat, since at the timber line they come out of hibernation

later in the spring. Thus it is not clear exactly what part

food plays in the production of this dormant state.

Treviranus®® (1802) said that the cause of torpidity

during winter lies in the ability to live with all the vital

processes at a minimum. This is an acquired character

resulting from the habit of sleeping during winter, as is

evident, he thought, from the fact that it is lost in mar-

mots kept in captivity. The earlier opinion of Barton”

(1799) was that it is an accidental circumstance and not

a specific character. The general idea, however, that

some sort of instinct, in connection with other factors, is

involved, was held by Reeve** (1803), Barkow’ (1846), |

Claparède? (1905) and others. Desjardine’! (1843)

thought that the need for sleep in rodents is. as great as the

necessity of migration in birds. Blandet (1864) de-

scribed winter-sleep as a relic—an echo from remote

periods when this phenomenon was general, having de-

veloped as a result of winters so severe that unless this

conserving process was resorted to, the animals would

have perished. Hibernation is thus, according to this

author, the effect of habit and annual periodicity. It still

persists in certain animals, but will soon become extinct.

Brunelli?s (1902) believes that this tendency is the result

of a long period of evolution favored by the nature of the

burrow, ete., where hibernation takes place. But accord-

ing to Albini®s (1894) the factors aiding this evolution are

not remoteness or other conditions of the burrow, but the

immobility of the animal. Carlier’? more recently (1911)

classifies hibernation with estivation (summer-sleep) and

migration. Winter-sleep in mammals like the instinct

No. 598] THEORIES OF HIBERNATION 619

of migration in birds, he thinks, may have developed in

remote ages, the prime cause being want of food, and not

cold.

Dubois*® (1895) has developed a carbonic auto-narcosis

theory according to which hibernation is due to the accu-

mulation of CO, in the blood and tissues of the animal.

This excess of CO, is supposed to cause a form of nar-

cosis as seen in the torpid condition of the hibernating

animal. When the CO, reaches a certain concentration

the respiratory center is excited, respiration accelerated,

and the muscles become hyper-irritable. These culminat-

ing results are responsible for the awakening from dor-

mancy. The author claims that he can induce typical

hibernating sleep by causing the active marmot to breathe

a mixture of air (42 per cent), CO, (45 per cent) and

oxygen (12 per cent). Torpid marmots remain dormant

if supplied with this mixture. By increasing the propor-

tion of CO, respiration is accelerated, and if the supplying

of CO, is continued the hibernating marmot wakes up.

The CO, is supposed to act principally on a nervous

center for sleep situated in the mid-brain, since marmots

deprived of cerebral hemispheres are able to sleep and `

wake up; but with only the medulla intact they are un-

able to awake. Further, Dubois’* (1894) found that CO,

actually accumulates in the blood during hibernation in

the marmot and decreases again when the animal wakes

up. Such an increase in the CO, content of the blood

during hibernation has just been observed in this labora-

tory in case of the woodchuck (Marmota monaa)."°

Upon sufficiently good authority’ to receive the serious

consideration of such an author as Max Verworn,” cer-

tain ascetics of India, known as fakirs, are able to volun-

tarily go into a condition of almost suspended animation

not unlike hibernation in some respects. While in this

condition it appears that these fakirs may be buried three —

or four feet in the ground for days, or may be inclosed for

six weeks without food and with but little air in a tight —

box which in turn is sealed up in some dark inner room. ~

620 THE AMERICAN NATURALIST [Von. L

When disinterred the body is cold and stiff with no signs

of any pulse, and apparently lifeless; but it revives with

no bad after-effects upon the application of warm water

to the head and after being manipulated for a quarter of

an hour. Dubois emphasizes the fact that in order to

induce this state of trance, the fakirs make it a point to

breathe as little as possible. This and much other in-

direct evidence is brought forward by this author in sup-

port of his carbonic auto-narecosis theory of hibernation.

Mosso*® (1899) holds just the opposite view. He thinks

that winter-sleep is due to a condition of acapnia, or lack

of CO, in the system.

It is not strange that in this age of ductless glands and

internal secretions some theory should be brought for-

ward that would involve the ductless glands. In 1905

Salmon” advanced the view that the pituitary body (hy-

pophysis cerebri) is a center for sleep and produces an

internal secretion which by virtue of some vasomotor or

autotoxic power acts on the nervous system and thus pro-

duces normal sleep. His view has been further elaborated

in later publications®® (1910) in which he states that hi-

bernation may be explained upon an analogous mechan-

ism involving especially the so-called hibernating gland

—a structure which has lately received renewed attention

by physiologists. Salmon seems to favor the old idea

that a depletion of the cerebral blood vessels offers the

best explanation of the lethargy characteristic of the hi-

bernating state. The rdle of the hibernating gland, how-

ever, is very uncertain. This structure is now generally

regarded as reserved food. Vignes®! (1913), however,

considers it probable that it plays some important physi-

ological role, particularly in hibernation, since its ex-

tirpation in the white rat, where the operation is at-

tended with little difficulty, is nearly always fatal. He

finds that this structure modifies the action of certain

toxic substances such as adrenalin, chloroform, tetanus

toxin and cobra venom, retarding the action of some and

accelerating that of others. He further maintains that

No. 598] THEORIES OF HIBERNATION 621

this gland contains lipase, and while it does not convert

starch to sugar, its extirpation diminishes the amyloptie

power of the serum. It also has an antitryptic power.

Thus he conceives that it might serve as an economizer

of proteins by insuring the utilization of reserve carbo-

hydrate and fats during the long period of winter-sleep.

Salmon’s view on the rôle of the hypophysis cerebri in

the production of sleep was soon criticized by Gemelli®*

(1906), who argued that if this hypothesis were correct,

the pituitary body would show signs of increased activity

during hibernation, since, as has already been stated, hi-

bernation is considered by many to differ from ordinary

diurnal sleep only in degree and duration. But on the

contrary, he found that the cyanophil cells of this gland

in the marmot decreased during winter-sleep and that

they increased again simultaneously with the appearance

of numerous karyokinetic figures after the animal wakes

up in the spring. Gemelli interpreted his findings as in-

dicating that the anterior lobe of this organ cooperates

with other ductless glands in neutralizing toxins which

are produced in increased quantity when the animal be-

comes active, and hence is not to be regarded as a center

of sleep. A later contribution to the relationship be-

tween the pituitary body and hibernation is by Cushing

and Goetsch®? (1913). Asa result of observations on the

hypophysis of the woodchuck, in which they confirm in

general the findings of Gemelli on the decrease in size and

histological changes during winter-sleep, these authors

suggest that hibernation may be ascribed to a period of

physiological inactivity, possibly of the entire ductless

gland series, but certainly more especially of the pituitary

gland, because during the dormant period this structure

diminishes in size and shows profound histological

changes and because deprivation of this gland in the

human subject and in experimental animals causes a train

of symptoms comparable to those of hibernation. Mann“?

(1916), however, found demonstrable changes in the

pituitary body and other ductless glands of a large num-

622 THE AMERICAN NATURALIST [ Vou. L

ber of ground-squirrels (Citellus tridecemlineatus) to be

absent or so inconstant, especially at the critical period—

at the onset of hibernation—that the assumption of any

theory ascribing the phenomenon of hibernation to a lack

of function of all or any one of the ductless glands is not

justified.

From this general summary it will be seen that great

diversity of opinion prevails regarding the immediate

cause of this extremely interesting condition, and of the

sudden transformation from the homoiothermal to the

poikilothermal state (and vice versa) so characteristic of

hibernating mammals. It is not the author’s object,

however, to discuss the relative merits of the various

theories. Suffice it to say that all of them are based upon

insufficient data. To say which of the various conditions

associated or occurring simultaneously with winter-sleep

are concerned with the production of the lethargy and

which are the results of this or some other condition, is

extremely difficult. Until certain causal relations are defi-

nitely established between the factors concerned, many of

these theories are of very little value except as a stimulus

to further research. It is thus very evident that we are far

from having any adequate explanation of the mechanism

of this phenomenon, to say nothing of how it was estab-

lished as a more or less variable character in certain

animals. |

If hibernation of mammals is only an extreme form of

ordinary diurnal sleep of man and other animals, it is

especially to be hoped that this subject will continue to be

investigated by more modern and adequate means, for no

entirely satisfactory theory has yet been advanced to ex-

plain the physiological cause of ordinary sleep. Since

winter-sleep may also be attended with total abstinence

from food and drink for many months, the facts derived

from a study of the various conditions associated with

this dormant period are of interest also in connection with

the subject of inanition in particular and metabolism in

general, as is plainly indicated by the frequent reference

No. 598] THEORIES OF HIBERNATION 623

to and comparison with the observations on hibernating

animals found in the literature on inanition.

REFERENCES

(A short article by L. Hoffman, Monatshefte f. å. naturw, Unterricht,

1914, 196-202, was not accessible, nor could any review of it be found.)

c. Hist. an. de. quadrup. vivip., Zürich, I, 802; ibid., Frank-

- Dubois, R. Physiologie — de ny marmotte, Paris, 1896.

. Polimanti, O. Il Letargo, Roma,

Buffon. Hist. nat., 1749, XVI rite XVII; G. L. Leclere comte de

Buffon, Œuvres conipisies, nouvelle édition, Garmier Fréres, Paris,

II, 636; Daubenton, Œuvres compléte de Buffon, nouvelle édition,

Paris, 1824, Mammiféres, V, 122 and 202; Lasiaide, Œuvres com-

pléte de Buffon, nouvelle édition, Paris, 1817, VI, 392.

Œuvres complétes de Buffon, 1829, XIII, 360.

Spallanzani, L. oe de phys. anim. et végét., Traduc. de Sénebier,

Genéve, 1787, I, ; Mémoire sur la respiration, Sénebier, Genève,

1803, 109; ibid., paas translation, Edinb., 1804; Eloge de Spal-

lanza i par M. Abert, Société d’Emul., II (année).

zi FREER to Serbelloni

erm P. Atti del ’accad. fis. med. statist., Milano, 1866. XXII,

Be w

9. Sea J., Animal Oeconomy, 1786, Owen’s ed., London, 1837,

10. Deshemtan: V. Hist: nat. gen., 1760, VIII, 228; Suppl. à ae oh

184; ibid., 1776, , 125.

11. Adeotding to Dau benton eee. ott.)

12. Ace eg to Mangili (see reference 19).

Low

13.

14. pea A. Phil. T nas 1805, 17.

15. Barkow, he Winterschlaf nach seinen Erscheinungen im

Thierreich ean. meats 1846.

16. Reeve, H. An Essay on Sox Torpidity of Animals, London, 1809.

17. Bert, P. Compt. rend. . biol., 1868, ser. IV, V, 13; Legons sur la

physiologie comparée de ots respiration, Paris, 1870, 507.

18. Ibid. Compt. rend. soc. biol., 1873, ser. 156; i

19. Mangili, G. Saggio d’osservazioni per servire alla storia die mammi-

feri soggetti a periodico letargo, Milano, 1807; Ann. d. muséum

d’hist. nat., 1807, IX, 106; ibid., 1808, X, 435; Reil’s Arch. f.

Physiol., 1807-8, VIII, 427; Fünf Mitteilungen über den Winter-

schlaf, Pavia, 1818

20. Dubois, R. Physiologic comparée de la marmotte; Paris, 1896, 23.

21. Hall, Marshal Phil. Trans., London, 1832, pt. II, 335; Todd "g Cyclop.

f Anat, and Physiol., 1838, II, 764.

22 Bawards, W. F. De l'influence des agents physiques sur la vie, Paris,

23. Legallois. (Œuvres, Paris, 1824.

24. Dugés. Physiologie comparée, 1838, I, 468.

624 THE AMERICAN NATURALIST [Vou. L

25. Blandet. Compt. rend. acad. sci., Paris, 1864, LIX, 656.

26. Patrizi, M. L. Atti della R. PAE della sci., Torino, 1894, XXIX;

Arch. ital. de biol., 1894, XXI, 9

27. Dubois, R. Physiologie comparée de la marmotte, Paris, 1896, 21;

Arch. internat. d. physiol., 1910-11, X, [69].

28. Brunelle, G. Riv. baer di sie nat., Henk: te anno XXI.

29. Claparède, E. Arch. d. , 1905, IN,

30. Cleghorn, A. Po spe rae TRECA ay te 1910, ees 356.

31. Salmon, A. La fonction du sommeil, Paris, 1910, 84 and 1

32. Mari A. Arch. p. le sc. med., 1906, XXX, 341; ee "1906, i,

33. Pieron, H. Le oe physiologique du sommeil, Paris, 1913.

34. Monti, R. Rend. del s re Lombardo, 1905, ser. IL, XXXVIII,

va Arch. ` fisiol., MED

35. Basy, J. A, E A sas ae anatomiques, chemiques, etc.,

sur la physique des animaux mammifères hybernants, Paris et Lyon,

1808; Reil’s ge f. d. Physiol., 1815, XII, 293; Memoire de 1’ Acad.

de Tirin, .1811,

36. Prunelle. Ann. du pais d. Hist. Nat., Paris, 1811, XVIII, 20 and

302.

chaleur animale, Pari 1871; Leçons, 1872, IX, 45; “rare aki la

chaleur ERX ras 374,

38. Quincke, H. Arch. f. exper. Pathol. u. Pharm., 1882, XV, 1.

39. Dutto, U. Rend. della R. Accad. dei Sean 1896, ser Va, V, 270;

biol., 1897, XXVII, 210; ibid., ;

40. Marzbacher, L. Ergeba, d. Physiol., 1904, III, pt. 1, 214

41. pi J. Recherches sur la vie oscillante, Essai de Pe ESN eo Paris,

can, 1903.

42. “aa F. Sbornik — 1889, II, 458; Compt. rend. soc. biol.,

1892, ser. IX, IV,

43, Pebuaatl o. eE deta R.A ccad. med. di Roma, 1904, XXX,

fase. Bee Arch. ital. de biol., 1904, XLII, 359; I Letargo, Roma,

1912, 120.

44. Simpson, S. Amer. Journ. Physiol., 1911-12, XXIX A

45. Polimanti, O. Arch. Í. d. ges. Physiol., 1914, CLVII,

C . CLIX, 320.

47. Pembrey, M. S. Journ. of Physiol., 1895, XVIII, 363 (older observa-

7 z . . by

Arch. f. d. ges. Physiol., 1902, LXXXIX, 154.

49. Merzbacher, L. Ibid., 1903, XCVII, 569.

50. Marès, F., and H ellich, B. Compt. rend. soc. biol., 1889, ser. IX, I,

410.

51. Liébeault. Du sommeil et des états analogues, Paris, Masson, 1866.

52, ea by Revue de seman! > I, 318; Zentralbl. f. Physiol.,

8 (literature of 1887 3} I

53. sleigh - 8. Novae species eames e glirium ordine, Erlange,

1778, 118,

No. 598] THEORIES OF HIBERNATION 625

54.

- Pembrey, M. S., and Hale White, W. Journ. of Physiol., bie ax,

for)

aan

w po

83.

Tiedermann, F. Deutsches Archiv fiir die Physiologie, 1815, I, 4

- Horvath, A. Centralbl. f. d. med. Wissensch., 1872; Verhandl. d. phys.

med. Gea ellsch., irk zburg, N: P., Ra oo 139; ibid., 1879, XIII,

60; ibid., 180, ‘XIV, 55; bid, 1881,

. Bunge, G. (von). Lehrbuch die ste os Menschen, 1901,

Popular cs Monthly, N. Y., 1914, LXXXIV, 147.

a oid

i peeti to Saissy (see reference 35).

. According to Mangili (see reference 19).

ace. Revue te magaz. de zool., Paris, 1858, ser. III, X.

. Mills, W. Trans. Royal Soe. of Canada, 1892, 49; Trans. Pan-Amer.

, 1274,

Congr., Washington, 1893 (pub. in 1895), pt. I

Med.

. Berthold. Arch. f. Anat., Physiol. u. wissensch. Med., 1837 , 63.

. Valentin, G. Untersuchungen zur Naturlehre des VIR und der

i I

Thiere, von Jac. Moleschott, 1857, II, 1.

Pembrey, Text-Book of. Physiology, edited by E. A. Schäfer,

1898, I, 79

. According to AENEON Il Letargo, Roma, 1912, 119.

. Simpson, 8. Proc. Soe. a aper. Biol. and Med., 1912, IX, 92.

Albini, G. Rend. ra accad. d. sci. fisiche. e math. (huoi della Soe.

Reale di rege oe ser. II, VIII, 15; ibid., 1901, ser. III, VII,

18 and 127 1903-4, ser. ITI, IX,

; eons ue Göttingen, 1802-22, v, hay neg 275.

arton, B

Trans. Amer. Phil. Soc., 1799,

. Desjardine. Am. des sci. nat., 1843, ser. II aoe oe 249.

- Carlier, E. W. rigar (bainpkist received in 1911).

. Dubois, R. Compt. rend. soc. biol., 1895, ser. X, II, 149, 514 and 830;

Physiologie comparée de on aaia o 1896, 246; Compt. rend.

soc. biol., 1901, LIII, 229; Compt. rend. acad. sci., 1909, CXLVIII,

. Ibid. Compt. rend. soc. biol., 1894, ser. X, I, 821.

r Jonen: Physiol., 1915, XXXIX, 20.

. Ame

- Braid, J. Observations on Trance or Human Hybernation, Aa

and Edinburgh, 1850; Der Hypnotismus (a translation into Germa

of all but two of the works of James Braid, by W. Preyer), Berlin,

1882; Medical Times, 1845, XII, 437 and 1850, XXI, 351, 401 and

416; "Vea: 1845, II,

325.

- Viowdik, M. Allgemeine Physiologie, 6. Aufl., Jena, 1915, 151.

- Mosso, A. Fisiologia dell’uomo sulle Alpi, Milano, 1899.

Salmon, A. Sull’origine del sonno. Studio delle relazioni tra il sonno

e la funzione della glandula Brea Florence

, 1905.

. Ibid. Riv. de med., 1910, XXX, 765; La fonction ite cee e

logie, psychologie, pathologie, Paris, 1910.

. Vignes, H. Compt. rend. soc. biol., 1913, LXXV, 360, 397, and 418.

ushing, H., and Goetsch, E. Proc. Soc. Exper. Biol. and Med., N. Y.,

1913, XI, 25; Journ. Exper. Med., 1915, XXII, 25.

SHORTER ARTICLES AND DISCUSSION

VARIATION, CORRELATION AND INHERITANCE OF

FERTILITY IN THE MAMMALS

THE purpose of this review is to give an outline of the prob-

lems of fertility in the mammals (exclusive of man) which may

be solved by the application of biometric formule to statistical

data, to furnish an index to the available statistics, and to indi-

cate the results to which the statistical analyses of the raw ob-

servations have led. Many of the biometric constants are pub-

lished here for the first time.

TYPE AND VARIATION IN FERTILITY

The most fundamental biological questions which can be asked

concerning series of data on fertility considered quite independ-

ently of any other characteristic of the organism, its ancestry or

its environment are three:

(a) What is the typical and average fertility of different spe-

cies or races?

(b) What is the variation, within the race, in reproductive

activity as compared with that of variation in the degree of

development of somatic characters?

described by mathematical curves?

With more comprehensive data concerning other character-

istices of the individual organism, its ancestry or environment,

more varied problems may be investigated, but none of more

fundamental importance.

With respect to (a) it need only be said initially that biolog-

ically a knowledge of the number of offspring characteristic of

a species has the same importance as a knowledge of any other

of its peculiarities. That species may differ widely in fertility

as in other characteristics is obvious without the collection of

extensive statistics or the application of mathematical formule.

It is only in a consideration of relatively slight differences in

fertility in nearly related species or races or in individuals of Ss

the same race existing under various conditions, that biometrie

626

No.598] SHORTER ARTICLES AND DISCUSSION 627

work became indispensable. It is just here, too, that the purely

descriptive significance of fertility gives way to genetic, eco-

non ie and sociological sources of interest.

As yet information on these subjects is all too meager. Lloyd’

has emphasized slight differences in fertility in species formation

in the rodents. Donaldson? has brought together the available

data for fertility in the rat. For swine, Rommel? and Bitting*

have given extensive data for different breeds and periods. Fur-

ther records are available for swine from the studies of Went-

worth and Aubel to be discussed below. Equations for theo-

retical curves of distribution of number of young per litter in

ommel’s series have been worked out by Surface.” Large

masses of statistics have been extracted from the herd books for

sheep by Rietz and Roberts. Taken altogether, only a beginning

has been made in a field that has not merely great biological in-

terest, but in certain of its bearings is of material economic im-

portance.

The most extensive and exact work on differences in fertility

has been done on man, but a discussion of this subject. falls out-

side the scope of the present review.

Since data for the solution of the problems of group (a) are as

yet inadequate, it is idle to attempt any detailed discussion of

those of group (b) and (c). Data for such purposes are, how-

ever, now becoming available much more rapidly than heretofore.

BIRTH ORDER AND LITTER SIZE

Fairly large series of records showing the relationship between

birth order and litter size are now accessible.

Minot’ has given data for the relationship in guinea pigs. The

averages which may be computed from these are :

Order of Litter, F Mean Size

| a A E AE eS A 51 1.96

Setola -P oiua sate r PAST eee 29 2.97

PRG o sa as be Er aA e N 15 2.80

rth

1 Lloyd, R. E., ‘‘The Inheritance of Fertility,’’ Biometrika, 8: 244-247,

1911.

2 Donaldson, H. H., ‘‘The Rat,’’ pp. 22-23, 1915.

3 Rommel, G. M., “The Fecundity of Poland China and Durae Jersey

Sows,” Cire, U. 8. Dep. Agr., Bu. Anim. Ind., 95, 1906.

4 Bitting, A. W., ‘‘The Fecundity of Swine,’’ Ann. Rep. Ind. Agr. Exp.

Sta., 10: 42-46, 1897.

5 Surface, F. M., ‘‘ Fecundity of Swine,’’ Biometrika, 6: 433-436, 1906.

1891,

è Minot, C. S., “‘ Senescence an creat Jour. BAL 12: bie, : “ oe

628 THE AMERICAN NATURALIST [Vou. L

Crampe’ many years ago showed from his extensive data on rats

that the maximum fertility was on the second litter. King and

Stotsenberg® have recently given data which lead to the following

averages :

Order of Litter F ~ Mean Size

Sete ee eee ees 21 6.24

A A ee ie a eg fee 21 FEL

ee ek E T e E 18 7.06

PO a aaa aaa a 15 6.40

Pearson and Weldon have shown? that in mice there is an in-

crease in the mean number of young from the first to the third

litter, thus:

Order of Litter Mean Offspring

P ee ee ec 5.46

as eer. 5.57

ae D Oe a oe ee er ae 5.76

For the rabbit Bailey, fide Hammond,” gives the values:

Order of Litt.r Mean Offspring

WO ae i ees ees 5.58 + 0.32

Pond e a i ee ee a es 7.25 + 0.41

WR: ooa oo sass DIZ ee 7.08 + 0.38

Such data as these are of obvious importance in the physiology

of reproduction in the mammals. They will be of far greater

value when it is possible to determine the influence of the actual

age of the mother as well as of the order of birth upon fertility.

Detailed records of size as well as of number of offspring and of

mortality would also be of great importance.

RELATIONSHIP BETWEEN FERTILITY AND SoMATIC CHARACTERS

The interrelationship between fertility and somatie characters

is a subject which may have a morphogenetic, genetic or eco-

nomic interest.

Reference to some of the earlier literature has already been

*Crampe, H., ‘‘Zucht-Versuche mit zahmen Wanderratten. I. Resultate

der Zucht in Verwandschaft,’’ Landwirths. Jahrb., 12: 389-449, 1883.

and J. M. Stotsenberg, ‘‘On the N. ormal Sex Ratio and the

s King, H. D, and

Size of the Litters in Peay Albino Rat (Mus norvegicus albinus),’’ Anat.

_— 9: ei

e Gi a

x ei On Some Factors Influencing Fertility in Domestic

Animals,’’ Your Agr. Bci., 6: 263-277, 1914.

No.598] SHORTER ARTICLES AND DISCUSSION 629

made™ in a memoir dealing with plant materials and certain

special problems more minutely analyzed on further sets of

data.”

In the mammal, the relationship between fertility and so-

matic characters may be determined from (a) the somatic char-

acters of an individual mother and her fertility, or (b) the

characteristics of the progeny which serves as the measure of the

fertility of a mother. Obviously, these two methods of operation

are biologically not at all comparable.

e economic importance of the possible correlation between

bodily characteristics and fertility has naturally given rise to

many popular beliefs concerning the existence of such a relation-

ship. Wentworth and Aubel? have, however, found no evidence

of such in a comparison of the mean litter size in ‘‘large type”

and ‘‘small type’’ hogs.

Pearson has shown" from Captain Lloyd’s data‘ that there is

a sensible and almost linear relationship between weight of

mother and number of young in litter in Poona and Belgaum

rats. The intensity of the correlation is, however, low, of the

order r —.160

Data for the full interpretation of such relationships are much

needed but not as yet available. The problem is evidently one of

great complexity. As Pearson points out, at certain stages of

pregnancy the number of young might actually influence, by its

own contribution, maternal body weight.** Furthermore, in

these rodents growth continues notwithstanding pregnancy, and

one might expect some correlation between weight of mother and

size of litter as a resultant of the relationship between the age of

the mother and her own weight and the age of the mother and

the size of her litter. |

Minot found that the over-gain in weight of pregnant guinea

11 Harris, J. Arthur, Biometrika, 8: 52-65, 1910.

12 Harris, J. Arthur, Amer. Jour. Bot., 1: 398—411, 1914.

18 Wentworth, E. N. and C. E. Aubel, Jour. Agr. Res, 5: 1148, 1916.

14 Pearson, K., ‘‘Darwinism, Biometry and Some Recent Biology,’’ I,

Biometrika, 7: 368-385, 1910.

15 Lloyd, R. E., ‘‘ The Relation between Fertility and Normality in Rats,’’

Ree. Ind. Mus., 3: 261-265.

16 Minot (Journ, Phys., 12: 141-145, 1891) has shown that in the guinea

pig there is a relatively enormous over-gain in weight before delivery.

17 Crampe (loc. cit.) has given RN data on the weight of mothers after

the first and second deliveries in the ra’

18 Watson, 3B The Effect of ia o Bearing of Young upon the PN

630 THE AMERICAN NATURALIST [Vou. L

adduced evidences to show that females which have borne young

are heavier than unmated controls. Whether the effect of bear-

ing young is cumulative in such a way as to influence the corre-

lations in Captain Lloyd’s series is not yet evident.

Taking these various factors into account, there seems little

ground for believing that there is any material correlation be-

tween the fertility of a mammalian female and her measurable

somatic characters.

There is an obvious physiological and morphogenetic interest

attaching to the correlations between the number of individuals

born in a litter and the characteristics of these individuals.

Consider first the correlations between number of pigs in the

litter and number of nipples, in swine. For Parker’s!? and Bul-

lard’s data the values are:

For males, r= .0810 + .0121,

For females, r — .0324 + .0124.?°

These are numerically low, but both are positive, and may pos-

sibly be significant in comparison with their probable errors.

They may indicate morphogenetic relationships between the

` vigor of the mother as indicated by the number of her young

and the characteristies of these young.

These positive correlations for number per litter and number

of nipples are of interest in connection with the negative corre-

lation for number in the litter and mean weight of individuals

suggested many years ago by Minot,” who states that in guinea

pigs the size of the animals at birth depends to a considerable

degree upon the number of young in a litter: the larger the

litter the smaller the animals at birth. Fortunately Minot has

given data from which approximate? values of the correlation

between number of individuals per litter and birth weight may

be compared. The results are:

Weight and the Weight of the Central Nervous System of the Female White

Rat,” Jour. Comp. Neur. Psychol., 15: 514-524, 1905.

19 Parker, G. H. and C. Bullard, ‘‘On the Size of Litters and the Number

of Nipples in Swine,’’ Proc. Amer. Acad. Arts and Sci., 49: 399-426, 1913.

20 Parker and Bullard give the correlation r—=.0035 + .0124 for females

only. This is evidently erroneous. Both values given here have been cal-

culated from their data,

21 Minot, C. S., ‘‘Senescence and Rejuvenation. I. On the Weight of

Guinea Pigs,” Jour. Phys., 12: 96-153, 1891.

23 The only difficulty lies in the fact that his Tables VII and VIII do

not contain the same number of individuals.

No.598] SHORTER ARTICLES AND DISCUSSION 631

For males, faw = — .437 + .039,

For females, fn» ==— .431 + .044,

For all, Tnw —=— .430 + .029.

The results for males and females are in remarkable agree-

ment. Evidently there is a large influence of the number born

upon the weight of the individual.”*

If the results be expressed in terms of regression of weight of

individual upon number in litter the equations are:

For males, w= 87.626 —5.214n,

For females, w= 84.375 — 4.741 n,

For all, w = 86.006 — 4.960 n.

King* has given direct evidence for the influence of weight of

mother in the weight of the young at birth.

Very young females and those that have passed their prime have

smaller litters, as a rule, than females at the height of their reproduc-

tive powers.

And again,

The body weight of a female influences the birth weight of her young

chiefly because it depends on the two more important factors of age and

physical condition.

Finally it may be noted that in the case of sheep the size at

birth and rate of development of twin and triplet as compared

with the single lambs is a problem of very real economic im-

portance. Both Bell and Marshall have considered this phase

of the question. Unfortunately no extensive quantitative data

are available for analysis on this point.

INHERITANCE OF FERTILITY

Biologically all recent studies on the inheritance of fertility

differ from the classic memoir of Pearson, Lee and Branley-

*8 That factors other than number per litter may profoundly influence

birth weight may be seen at once by determining the correlation between

the weight of the individual pigs born in litters of two as given in Minot’s

Table XI. Using symmetrical tables I find for the correlation between the

weight of the two individuals

Tis wa == .686 + .046,

This similarity in weight is probably due in part to hereditary and in part

to environmental factors.

*4King, H. D., ‘‘On the Weight of the Albino Rat at Birth and the

Factors that Influence It,’’ Anat. Rec., 9: 213-231, 1915. :

632 THE AMERICAN NATURALIST [ Vou. L

Moore?’ on fertility in man and fecundity in race horses in that

they deal with the number of young produced at a single birth

instead of with the total young produced during the reproductive

period or the ratio of the number of young actually born to the

number which might have been produced under the circum-

stances.

For Poland China sows Rommel? and Rommel and Phillips”

found values of the correlation between the size of litters in

which dam was farrowed and size of litters produced by daugh-

ters ranging from .1088 to .0032, the values decreasing with

moderate regularity as the daughters became older. For all

ages they find the correlation r—.0601, and conclude that fer-

tility is slightly but definitely inherited.

George (fide Wentworth and Aubel, loc. cit.) worked out four

supplementary series in Poland China swine with the results:

Daughter and dam, r==.0615 + .0390,

Dam and grandam, r= .1147 + .0343,

Daughter and maternal grandam, r—=.0025 + .0392,

Daughter and paternal grandam, r==.0508 + .0392.

All of these values are positive, but they are very small and no

one of them may be considered statistically trustworthy in com-

parison with its probable error.

Weldon and Pearson”: give a series of six relationships, both

parental and grandparental, for size of litter in mice, with the

result that no correlation whatever could be demonstrated.

Wentworth and Aubel? have considered the possibility of the

Segregation of litter size in the two first descendant generations

of matings between boars and sows farrowed in litters of various

sizes by determining the standard deviation of the number per

litter in the so-called F, and F, generations. Let 1 be the num-

25 Pearson, K., A. Lee and L. Branley-Moore, Phil

aii haea hr ve y-Moore, Phil. Trans. Roy. Soc. Lond.,

2° Rommel, G. M., ‘‘ Inheritance in the Female Line of Size of Litters in

Poland China Sows,’’ Biometrika, 5: 203-205, 1906

27 Rommel, G. M., and E. F.,

of Size of Litter in Poland Chin

264, 1906.

* Feen 5, **On Heredity in Mice from the Records of the Late W. F.

. eon. I. On the Inheritance of the Sex Ratio and of the Size of

Litter,” Biometrika, 5: 436-449, 1907.

29 Wentworth, E, N., and O. R. Aubel, ‘‘T

Jour, Agr. Res., 5: 1145-1160, 1916.

Phillips, ‘‘Inheritance in the Female Line

a Sows,’’ Proc. Amer. Phil. Soc., 45: 245-

nheritance of Fertility in Swine, ’’

No. 598] SHORTER ARTICLES AND DISCUSSION 633

ber of pigs in the litter in which an individual was farrowed, d

the number of pigs in the litter in which its dam was farrowed,

and S and D the numbers in the litters in which the grandsire

and grandam were farrowed. Then, the authors reason, if fer-

tility be due to factors which differ in the grandsire, S, and the

grandam, D, and if Mendelian segregation occurs in the fashion

assumed by several of these who have worked on quantitative

characters, one should expect the mean value of the standard

deviation of litter size in cases in which D and 8 differ widely to

be higher than the mean value in cases in which they are closely

similar. There is no conclusive evidence of such greater segre-

gation in the F, from dissimilar grandparents.

Now the data published by Wentworth and Aubel permit the

consideration of several additional questions of considerable in-

terest in connection with the problem of the inheritance of fer-

tility. Thus from the mean litter sizes in their Table II and the

distributions of litter size in the three generations in their Table

IV, it is quite possible to calculate approximately* correct cor-

relations for the relationship between size of litter in different

generations. Thus the formula:

_ 2 (angie) — [2(@)/NIE(ng)/N]

gilt CxO y ;

where the bars denote the means of the y (descendant) litters

associated with particular, x, classes of ascendant litters, leads

to the values :**

Ts = 071 + .023, rpa = -126 + .022,

try = 120 + .022, rp, = .100 + .022.

Superficially considered, these values seem in excellent agree-

ment with those published by Rommel and others, but the fact

that rą has a value which is possibly significant statistically,

should at once arouse suspicion, for surely there is no genetic

reason (excepting possibly non-viability of sperm or the produc-

tion of duplicate twins through an influence of the sperm upon

the egg) why there should be a correlation between the size of the

litter in which a boar was farrowed and the size of litter in which

his daughter was farrowed. Mistrust is heightened by the fact

80 Unfortunately there are inconsistencies in these tables which show the

existence of typographical errors precluding exact constants. :

81 Unfortunately data for the determination of ra; are wanting.

634 THE AMERICAN NATURALIST [Vor. L

that rą is actually though perhaps not significantly lower than

rs;, Whereas on the female side rpar>rp.- Obviously there is no

genetic reason for a correlation between the size of the litters in

which the grandsires, S, and the grandams, D, were farrowed.

But columns 1-3 of Table II of Wentworth and Aubel actually

give:

fsp == .121 + .022,

a value quite as large as those recorded above.

Such a correlation might arise (a) through the existences in

the pens of different breeders of strains slightly differentiated

with respect to fertility, (b) through differences in the condi-

tions in which different breeders maintained their pens, provid-

ing such conditions affect litter size, or (c) through actual dis-

honesty of certain breeders in reporting the size of litters for

herd-book publication.

Such differentiation, if it exists, would also account in part at

least for the correlations hitherto regarded as due to inheritance.

The whole problem is evidently one of great complexity and re-

quiring far more detailed investigation than it has yet received.

The problem of the inheritance ‘of the production of twins in

sheep which has been studied experimentally by Alexander

Graham Bell for the past several years, has recently been in-

vestigated statistically by Rietz and Roberts.*?

There seems to be unmistakable evidence of inheritance, or at

least of ascendant influence,** upon descendant characteristics. ,

This may be most clearly seen by comparing the average number

per litter resulting from certain matings.

Thus for the parental relationship the results are:

Whon sro tad daims ate single’. (6:2... 20.2 02 ee, OS 1.3452 + 0.0059

When sire is single and dam is twin .................... 1.4171 + 0.0067

When sire is twin and dam is single ................+--- 1.3946 + 0.0073

When sire is twin and dam is twin ..............:..2.+- 1.4548 + 0.0088

When either sire or dam is a triplet ................... 1.6076 + 0.0300

Moan Of ah ofapring oii eee 1.3979 + 0.0035

*2Rietz, H. L. and E. Roberts, ‘‘Degree of Resemblance of Parents and

Offspring with Reference to Birth as Twins for Registered Shropshire

Sheep,’’ Jour. Agr. Res., 4: 479-510, 1915

33 In the case of slight relationships between parents or earlier ancestors

and offspring there is always danger of attributing to heredity the influence

of purely physiological factors. ee

No.598] SHORTER ARTICLES AND DISCUSSION 635

Or for the dams and grandams:

When dams and grandams are singles ...........6++e000: 1,3446 + 0.0057

When the dams are singles and grandams twins .......... 1.3689 + 0.0070

When the dams are twins and grandams are singles....... 1.4245 + 0.0071

When the dams are twins and grandams are twins ........ 1.4559 + 0.0078

When either dam or grandam is a triplet .............-. 1.545 - 0.037

Finally for the maternal grandams alone:

When the maternal grandams are singles ........+..+.++. 1.3784 + 0.0045

When the maternal grandams are twins ...........+.++-- 1.4120 + 0.0052

When the maternal grandams are triplets ............... 1,556 + 0.033

It is quite out of the question to review in any detail the

thorough analysis of the numerous interrelationships deduced

from the many thousands of records abstracted by the authors

from the Shropshire record. Their data seem to be free from

the possible objection raised against the swine records above, for

the correlation between sire and dam, which may be deduced

from their Table I, is only r==.0058 + .0070.

The intensity of correlation between the size of litter in which

an individual is born and the size of the litter in which his sire

or dam or grandsire or grandam was born is very low. The

maximum relationships are in fact of the order r= .08.

In the parental relationships the correlation between the size

of litter in which the sire was born and the size of the litter in

which his offspring were born seems to be significant, as well as

that between the size of the litter in which the dam was born and

the number of the offspring. The mean number of offspring.

are:

When the sire was born single .......-.++ee+eeeretrrees 1.3787 + 0.0045

When thé siro was @ twits ook cee ees eee oes ee oe 1.4220 + 0.0057

1.683 + 0.045

When the sire was a triplet .........0:-seeeeeereeereeee

Note the agreement of this result with that obtained by Went-

worth and Aubel. An explanation on the basis of identical twins

induced by the characteristics of the sperm, or of partial im-

potency in certain males, should be sought by those who have

experimental facilities.

There seems to be a significant correlation between maternal

636 THE AMERICAN NATURALIST [ Vou. L

grandams and offspring, but it is impossible to assert any trust-

worthy correlation for the other grandparents.

INFLUENCE OF ENVIRONMENT ON FERTILITY

Marshall** while emphasizing the importance of the hereditary

factor in multiple births in sheep, adduces evidences for the

great importance of feeding as a factor in the production of

twins and triplets. His figures certainly show great and con-

sistent differences in the produce of flocks which have received

different treatment at and preceding tupping time. Unfortu-

nately differences in breed may, but do not necessarily, cast some

doubt on the interpretation of the data. The problem which he

has attempted to solve by the analysis of schedules received from

flock masters certainly deserves experimental study. Such in-

vestigations have actually been begun by Evvard who in a first?

and second? and third®’ report on experiments with swine has

given the results of various feeding upon the vitality of the off-

spring. Discussion of the data as they are presented in these

papers falls outside the scope of a biometric review. Such work

is, however, of great importance at a period of science in which

heredity as contrasted with environment is apt to be assumed to

be an all-important factor. It is a pity that such experiments

as these of Marshall and Evvard can not be carried out in close

cooperation with experts on the physiology of nutrition, so that

differences in rations might be arranged on a uniform scale.

J. ARTHUR HARRIS.

ON A BARNACLE, CONCHODERMA VIRGATUM, AT-

TACHED TO A FISH, DIODON HYSTRIX?

A SPECIMEN of the ‘‘sea poreupine,’’ Diodon hystrix Linn.,

seen swimming near the surface and secured with a dipnet, was

34 Marshall, F. H. A., ‘‘Fertility in Scottish Sheep,’’? Trans. High. Agr.

sit — V, 20: 139-151 , 1908.

ard, J. M., 4 Watrition as a Factor in Fetal Development,’’ Proc.

rie pes A88., 3: 549-560, 1912.

36 Evvard, J. M., t ‘Some Pastors affecting Fetal Development,’’ Proc.

Iowa Acad. Sci., 20: 325-330, 1913.

87 Evvard, J. M., A. W. Dok and S. C. Guernsey, ‘‘ The Effect of Calcium

and Protein Fed Pamani Swine upon the Size, Vigor, Bone, Coat and

wari of the Offspring,’’ Proc. Iowa Acad. Sci., 21: 269-278, pl. 31-35,

1 uli from the Bermuda Biological Station for Research, No. 50.

No.598] SHORTER ARTICLES AND DISCUSSION 637

found to have two living lepad barnacles attached to one of its

erectile spines? upon the ventral surface two centimeters to the

right anteriad of the anus. The Diodon was a small individual,

16 em. long. It was kept under observation in the laboratory for

several weeks.

According to a determination for which I am indebted to Mr.

H. G. Coar, the barnacles belong to the species Conchoderma

virgatum (Spengler), although varying ‘‘a trifle from Gruvel’s

type description, but not sufficiently to correspond to Concho-

derma hunteri R. Owen, 1830, which the specimen approached

slightly, nor to Leach’s (1818) variety chelonophilus of C. vir-

gatum.’? This species has not previously been recorded from the

Bermuda area, though it is known over the Atlantic generally

and (to judge from statements of fishermen) occurs here upon

young turtles. C. virgatum has been found on Mola, ships’ bot-

toms, and various other objects (Pilsbry, 1907, p. 99), but the

present record is somewhat unusual.

Different semiparasitic lepads have quite various hosts, such

as meduse, antipatharians, the spines of echinoids, molluses, crus-

taceans, sharks, teleosts, turtles, the tail feathers of sea birds,

whales, and so forth (Pilsbry, 1907; 1910). Those occurring

on fishes seem, naturally, to affix themselves to some hard part,

for example, the head, as in the case of Tylosurus (Sumner,

Osburn, and Cole, 1913, p. 647). Jordan (1905, p. 341, fig. 226)

figures a flying fish with conchodermas attached to a Penella

growing on the fish, a condition of double parasitism which has

been described for Xiphias. In the present instance, the larger

of the two conchoderma individuals (20 mm. long) was found to

have its peduncle completely surrounding the spine to which it

had become fixed. The second individual was much smaller (4

mm. long) and attached to the peduncle of the first. Both speci-

mens were so oriented that the opening between the valves was

directed toward the head of the fish. The skin of the fish about

the base of the spine was inflamed, and the muscles which nor-

mally control its elevation for defensive purposes had apparently

degenerated. When it was attempted to preserve the Diodon,

2The figure of Diodon hystric, which is used in current ichthyological

handbooks, represents the animal in a semipuffed-up condition and with the

frontal spines erected. Alive, the fish has a quite different aspect, all the

spines being flattened down to the skin unless the creature is much dis-

turbed. When preserved in formalin it assumes the appearance depicted

in the handbooks. 4

638 THE AMERICAN NATURALIST [Vou.L

the spine bearing the conchodermas became detached in the

course of the animal’s self-inflation. It is probable, therefore,

that the spine would soon have been shed under natural circum-

stances.

Several features of the behavior of these conchodermas are of

interest in comparison with those of other barnacles. Some years

ago it was reported by Pouchet et Joubert (1876) that cirripedia

“attached to rocks reacted to shading, while those attached to

floating objects did not; their inference being that to the station-

ary barnacles a shadow signified danger, whereas, to those borne

about at the surface of the water, a fluctuating illumination was

the normal state of affairs. This observation has been regarded

as an instance of adaptation comparable with that of Hargitt

(1909) on the gradual loss of reaction to shading when serpulids

are maintained in the laboratory.

The specimens of Conchoderma attached to Diodon did not re-

act to shadows under any of a number of experimental condi-

tions. They seem, therefore, to be in agreement with the obser-

vation of Pouchet et Joubert. But tests upon lepads found upon

floating timbers and upon Ascophyllum showed that Lepas

anserifera and L. pectinata do respond to shading by retracting

the legs and approximating the valves. From a number of tests

it appeared that neither the legs nor valves are sensitive to shad-

ing, but that the shadow must affect some part of the body within

the shell suggesting that the persisting nauplius eye is the

organ involved. The extent of the response varies with the de-

gree to which the appendages have been extruded: when just

being extruded, they react by complete retraction; when fully

extruded, by a partial retraction; after being fully extruded for

one or two minutes, they react to shading quite promptly and

completely. After completion of a response there must usually

elapse from two to four minutes before another reaction can be

secured.

It seems to me, then, that the supposed adaptation of floating

barnacles is not of the nature which has been supposed. Whether

the non-reaction of Conchoderma to shading is properly to be

considered a direct adaptation is therefore questionable. The

host of these particular specimens is not a surface fish, and the

absence of sensitivity to shading may be due to their deep

habitat. Direct sunlight inhibited the rhythmic movements of

the conchodermas, and they were much more active at night than

in diffuse laboratory light.

No.598] SHORTER ARTICLES AND DISCUSSION 639

The statement is occasionally met with that in barnacles at-

tached to a free-swimming animal the feathery feet are merely

thrust out, not waved about as in the rock barnacles, which must

create food- and respiratory-currents for themselves. Now, it

was observed that when the Diodon bearing the conchodermas

was actively swimming, the legs of the lepads remained extended

for as much as four to five minutes; whereas, when the fish re-

mained stationary, they were alternately extended and retracted

about seven times every minute (at 18° C.), the extension in the

latter case being not so great as when their host was moving.

Lepas anserifera and L. pectinata were then tested as to their

behavior in currents, with this result: when the wood to which

they were attached was stationary, the rhythmic contraction of

the appendages was continuous, but if a gentle stream of water

from a supply jet was allowed to flow past them impinging on

the anterior (concave) edges of the legs, they remained extended

for as long as ten minutes, and were spread farther apart than

in the absence of the current. This was not due to any merely

mechanical effect of the water stream, as the feet could at any

time be caused to contract at a touch. A water stream, striking

the posterior (convex) edges of the legs, led to contraction and

subsequent limited extrusion of these appendages. A more cor-

rect interpretation of the phenomenon described in floating bar-

nacles seems to be, therefore, that when the concave side of the

appendages is stimulated by a water current, the animal responds

by pushing out its legs further than is usual in the absence

of currents, while their rhythmic contraction is inhibited. It

should be noted that the two specimens of Conchoderma observe

were so oriented on the Diodon as to receive the full benefit of

currents derived from its forward swimming; and further, that

this fish is not a vigorous swimmer, so that the currents in ques-

tion are by no means rapid, but rather such as could be efficiently

strained by the barnacles.

W. J. Crozier

AGAR’S ISLAND,

BERMUDA

REFERENCES

colous Annelids. Jour. Exp. Zool, Vol. 7, pp. 157-187.

Jordan, D. S. 1905. A Guide to the Study of Fishes. Vol. I, 4°, xxvi + :

624 pp., 393 figs. New York.

640 _ THE AMERICAN NATURALIST ` [Vou. L

ilsbry, H. A. 1907. The Barnacles ee contained in the Col-

“ Jections of the U. S. National Museum. . U. 8. Nat. Muas. 60,

x+ 122 pp., 11 pl. Wash.

Pilsbry, H. A. 1910, Stomatolepas, a Barnacle Commensal in the Throat *

the Loggerhead Turtle. Amer. NAT., Vol. 44, pp. 304-306, 1 fig.

Pouchet, et Joubert. 1876. La vision chez les Cirrhipédes. C. r. et ii

Soc. Biol., Sér. 6, t. 2, pp. 245-24

Sumner, F. B., Osburn, k C., and Cole, bedi: 1918. A Biological Survey

of the Waters of Woods Hole and Vicinity, Section III. A Catalogue

of the Marine Fauna. Bull. U. S. Bur, Fish., Vol. 31 (1911), PL $

pp. 545-794. ;

: oe

VOL. L, NO. 599

The American

Naturalist

intended for publication and books, ete., intended for review should be

sent to the Zam of THE AMER

rt articles containing

ICAN NATURALIST, Garrison- -on-Hudson, New York.

summari

es of research work b arin on the

lies sof Sdk evolution are especially welcome, and will be given preference

in publ

rea reprints of prpers are supplied to authors free of charge,

maaa

F urther re Saan will be supplied at c

Subscriptions and arisrtiuteaents should be sent to the publishers.

ear. Foreign postage is fift

subscription price is Four dollars a

Canadian posta

yea

-five cents additional.

The

The pei for single copies is

forty cents. The Bootie! Bal rates are Four Dollars for a p

Lancaster, Pa.

THE SCIENCE PRESS

Garrison, N. Y.

_ NEW YORK: Sub=Station 84

plaea matt,

, April 2,

1908, at the

Post Office at Lancaster, Pa., under the Act of

Congress of March 3, 1879.

3 NATURAL HISTORY OF MAN

au localities and age. “Fee Giller irene fx

Museums a

se OLDMAN, 77 Brixton Hill, London, England

The Dia ——- Arco stg ge -< arzi

By CHAR

escription oF the E deve be ‘Phitadetphie

nd the region within a radius of one hundred

ing the pre- s lacial deposits fo ar mgs be

blue elay ofthe ancient Delaware River has just | n issued

ie J. ippi t Company. papt:

complete i index. Size

it pe stage paid $5.

Address Charles S. Boyer,

1640 Columbia Avenue Ca gag Pe

= To Lepidopterists

= W. F. H. ROSENBERG

3 Haversteck Hill, Lames:

sat a very iris stock of

butterflies and Moths from all parts of

world. oe List free. ae 18.)

FRIDAY

‘The British Association for the Advancemen'

E; = i

SCIENCE

A WESEL JOURNAL DEVOTED TO THE

DVANCE HENT OF SCIENCE

, Ocrozer 20, 1916

0 ee a Dr AB: srs

Botany and its Econom:

Prosser: ohn M. Clarke..

E E

‘als ofthe Dearborn Obey Dr. nee :

erell. — f

i t Action asa Diffusion Phenomenon :

eae Bais a et = d

Different requencies : aegis ce

THE

AMERICAN NATURALIST

VoL. L. November, 1916 No. 599

THE EVOLUTIONARY SIGNIFICANCE OF THE

OSMOTIC PRESSURE OF THE BLOOD!

GEORGE G. SCOTT

COLLEGE OF THE CITY OF New YORK

Tur facts of comparative anatomy, embryology and

paleontology form the tripod of evidence on which rests

to a great degree the validity of the doctrine of evolution.

Accepting the doctrine of evolution as a working hypoth-

esis has resulted in clearing up puzzling problems in the

above named departments of biological inquiry. At the

present time, more attention is being paid to physiological

than to morphological problems. In physiology, the

great emphasis is placed on mammalian problems with

especial reference to man himself. Now if the mammals

are the product of a long process of evolution from simple

ancestors, it follows that not only has there been a mor-

phological evolution, but also the present complicated

functions of the higher animals have evolved from the

simpler processes of primitive ancestral forms. In order

to understand the significance of particular physiological

facts, we must therefore view the matter in the light of

evolution. It is not essential that all needful evidence be

at hand to make perfectly clear the significance of the

higher physiological activity. Indeed, it is well worth

while at times to state clearly any of our problems in the

evolutionary form and arrange the evidence accordingly.

1Read by title before The American Society of Naturalists, Philadel-

phia, Dec., 1914.

641

642 THE AMERICAN NATURALIST [ Vor. L

In this way we become aware of the need for information

to clear up the question which inevitably arises.

It is commonly known that the blood and body fluids of

animals possess a certain osmotic pressure. Life proc-

esses are constantly dependent on the passage of ma-

terials in and out of cells and differences in the osmotic

pressure of substances within and without the cell are

held to be one cause of this mutual movement. Varia-

tions in the osmotic pressure of the blood and body fluids

of animals are not so generally known. In the case of

severe hemorrhage it is a common practice to replace the

lost blood by a physiological salt solution which has the

same osmotic pressure as that of the blood. Formerly a

0.7 per cent. saline solution was used. This is isotonic

with amphibian blood. The reason for this was that the

fact was first discovered in a study of frog’s blood. The

saline solution (based on amphibian studies) of the physi-

ological laboratories was considered proper for use in

hospitals as well. Later it was found that a 0.9 per cent.

saline solution represents more nearly the composition of

human blood and this solution is in use at present.

But why does human blood have an osmotic pressure

equivalent to that of a 0.9 per cent. saline solution? In

order to answer this question we must examine all avail-

able data as to the osmotic pressure of the blood and body

fluids of animals in general. When this is done it appears

in many cases at least as though the osmotic pressure of

the blood and body fluids were merely a direct adaptation

to the environment. But in other cases this is not so

clearly apparent, in fact the osmotic pressure possessed

by certain forms shows no evident adaptation to the en-

vironment at all. The terrestrial vertebrates illustrate

this last condition. It is only when we view the entire —

question from the standpoint of evolution that the main

features of the puzzle become apparent.

It might be well to explain at this point the meaning of

osmotic pressure. One gram molecule of hydrogen gas

at atmospheric pressure occupies 22.4 liters space, and to

confine this gas in a space of one liter would require a

No. 599] OSMOTIC PRESSURE OF THE BLOOD 643

pressure of 22.4 atmospheres. A gram molecule of any

other gas under the same conditions has the same pres-

sure. Wan’t Hoff in his theory of solutions established

the fact that a substance in solution behaves as a gas oc-

cupying the same volume as the solution and the laws

which solutions obey are analogous to those which are

followed by gases. Therefore a gram-mol of a substance

dissolved in a liter of pure water would have the same

pressure as a gram-mol of gas, i. e., 22.4 atmospheres.

This pressure property of dissolved substances is called

osmotic pressure. Since the blood and body fluids con-

tain salts and other substances in solution, these fluids

therefore have a certain osmotic pressure. It is well

known that a salt solution has a lower freezing point than

that of pure (distilled) water. The difference is propor-

tional to the difference in concentration. Since the os-

motie pressure depends on the concentration, it follows

that the amount of the depression of the freezing point of

the solution below that of distilled water is a measure of

the osmotic pressure. The osmotic pressure stated in

atmospheres can be readily obtained from the “a: OF

depression of the freezing point by the use of the fol-

lowing formula. Osmotic pressure in atmospheres =

(A X 22.4) /1.85.

The blood of a vertebrate serves two double purposes.

It carries oxygen to tissues and carbon dioxide away.

This is its respiratory function. It also carries nutrients.

to tissues and wastes of metabolism from tissues. We

might call this the nutrient function. But the blood of

the earthworm is mainly a respiratory fluid. The body

cavity is filled with foods absorbed directly from the in-

testine and distributed by the peristaltic movements of

the body to the various tissues. In insects the air is car-

ried directly to the tissues through tracheæ while a so-

called heart lying on the dorsal side of the intestine and

open at its anterior and posterior ends aids in churning

and distribution of food absorbed into the body cavity

from the intestine. The indefiniteness of the term

‘“blood’’ is at once apparent. Most persons in using this

644 THE AMERICAN NATURALIST [ Von. L

TABLE I

SHOWING THE FREEZING Points (A) OF THE BLOOD OF ANIMALS

Species A Blood A Water Locality Observer

I. Coelenterate:

« ACYORINM. .. <...>». 2.195 2.29 Naples Bottazzi

II. Echinodermata:

2. Asteropecten.......... 2.312 2.29 Naples Bottazzi

S Aerie.. nii 2.295 2.29 is sy

4; Balahura eseri 2.315 2.29 bs ag

III. Annelida: .

5. Stpunculus..... 5... 7.31 2.29 Naples Bottazzi

IV. Arthropoda:

6. Homarus vulgaris... .| 2.292 2.29 Naples Bottazzi

7. Maj Madi. ena. 2.36 2.29 pi 3

Maja verrucosa...... 2.13 2.29 a Fredericq

9. Homarus americ...... 1.82 1.80 Woods Hole | Garrey

10. A eee so 1.78 1.76 St. Andrews | Macallum

11. mulus... 1.90 1.82 Woods Hole | Garrey

12. E A A A 2.04 ? ? Macallum

TS. ET oe eo Se. 0.80 DOS Se acne wine a Fredericq

442 Barbus ee kc 0.475 C08 oa AG x

V. Mollusca:

TO: A a 2.31 2.29 Naples Bottazzi

eps T ka, 2.24 2.29 2 ay

Cyclostomata:

7. Polistrotema..:...... 1.966 1.924 Monterey Greene

VII. Elasmobranchii:

3. Mustelus vulg........ 2.36 2.29 Naples Mosso

19. Trygon viol.. ........ 2.44 2.29 =

a. ee aie be 2.378 2.29 és Bottazzi

21. Mustelus laev........ 2.36 2.29 n 5

22. Scyllium stell........ 2.3 2.29 3 4

23. Torpedo ocell......... 2.351 2.29 - x

24. Torp. marmorata..... 2.292 2.29 ip

25. Squatina angelus..... 2.28 2.29 z My

26. Acanthias vulg....... 1.90 1.91 North Sea Dakin

21. Raia clavata.... 2... 1.90 1.91 . 2 vs K:

28. Carcharias lit........| 2.03 1.82 Woods Hole | Garrey

29. A l a8. SESS 1.88 1.82 6 B a

30. Mustelus canis....... 1.8692 1.81 C “ | Scott

31. Squalus acanthias....| 1.84 Te Lei bs bss 4

32. A n eo TO 1.42 New York chy

VIII. Pisces:

33. Acipenser sturio...... 0.76 2.00 Arcachon Rodier

Ot. Chante os 4 Oe 1.040 2.29 Naples Mosso

35. Serranus... 2.2... ..< 1.035 2.29 ” re

36. C Wl er 1.120 2 oly Bottazzi

37. Deutex vulgaris...... 1.022 2.29 e =

orhynchus........ 0.762 1.924 Von’ Greene

39. Pleuronectes fles .| 0.883 1.91 North Sea | Dakin

40. Pleuronectes plat..... . 0.71 1.91 = x 7

41. Lophius... i.. 0.80 2.00 Arcachon Rodier

42. Lump sucker........ 8 1.90 North Sea akin

- Gadi mor... 0.72 (0.64); 1.80 Baltie Dekhuyzen

44, Pleuronectes......... 1 1.80 x :

45. Conger vulg.......... 0.74 1.80 z 5

6. Cottus scorp......... OGG e hoa Amst T

47 ao e a Helder i

2 Mean A of eighty specimens.

No. 599] OSMOTIC PRESSURE OF THE BLOOD 645

Species A Blood A Water | Locality Observer

VIII. Pisc

48. Oaa angel 0.767 1.80 Baltic g

49. Gadus virens........ 0.76 1.80 f M

50. Gadue merl... sson 0.86 1.80 a Et

61. Molwa vuli... oo vre 0.716 1.80 eu i

52. V A byrkel N eee 0.86 1.80 és E

53. Motella Pit.. sop 0.605 1.80 m 4

54. Hippoglossus........ 0.671 1.80 s ái

55. Pleuronectes pl....... 0.672 1.80 P H

56. Pleuron Ce. 0.681 1.80 " M

57. Labrus eas es 0.694 1.80 N pe

58. Labrus mixtus....... 0.681 1.80 ie iy

59. Conger vulg.......... 0.696 1.80 m

60. Salmo trutta......... 0.785 1.80 8 a

61. Labrax lupus........ 0.72 1.80 ze e

62 Nec hirundo....... 0.669 1.80 i 7

OS. ARNTRRE. iaie 0.665 1.80 4 Ge

Agonus cataphr... LOD ase Helder "i

Gb; Zonrees rannin Biot eee piece at is

66 Tautoga onitis ere a 0.86 1.82 Woods Hole | Garrey

OF ee eA 0.70 1.42 New York cott

68. Cyonoscion. oh 0.792 1.82 Woods Hole | Garrey

69. Come el. 5-302. 0.82 1.82 es a a

70. Anguilla. . 0.90 1.82 a ne

71. Pew cpa tere 0.635 1.91 North Sea Dakin

T2; SOURS. 225 ee esos 0.75 1.82 Woods Hole | Scott

73. Morone am........-. 0.735 1.82 iy apes Wap

74. Oncorhyncus......... 0.628 0.03 Fresh water | Green

75. Moroneam.......... 0.571 0.03 ot eet)

76. Anguilla. . 5 0.57 0.03 A x kin

77. Pleuronectes. ge ore 0.68 0.03 ia g ji

TOS a e TEE E A ak 0.507 0.03 Ka a Dekhuyzen

79. Rane P TE oe ok, 0.519 0.03 j gs

80. ie forio eS 0.567 0.03 g rs s

81. Abramis blicca....... 0.497 0.03 4s s

82. Cyprinus carpoo..... 0.527 0.03 X zi z

83. Tinca vulgaris....... 0.519 0.03 e n

84. Leuniscus eryth 0.495 0.03 = a z

85. Erythrinus. ......... 0.577 0.03 "i 7 ;

86. Abramis brama. ..... 0.51 0.03 s 5 Dakin

87. Cyprinus carpio 0.487 0.03 Hs

IX. Amphibia: i

88. Rana eect... 5 0.563 gee AE A E TE Bottazzi

89. Bufo viridis......... O761 eA ae eee ee ....| Bottazzi &

Ducceschi

90. Bufo vulgaris........ Oti” I oeoperoalaasr Marae ra es Bottazzi

X. Reptilia:

. Thallassochelys fog 0.61 2.29 Naples M

92. Emys europa........ 0.463 2.29 x er sie &

P aoe 0.440 2.29 : Bottazzi

XI. Aves:

94. Capon. . osc. ces O66 oats oe ee os hee D'Errico

U6. Tirkey... a 0S ee ee lea eet tees :

96. Gallus bank.........- 0.623 oe de 1ee Vian. dae eet Oe

XII. per l ie

9 A Delphinus phocena. i te 0. 4 YU Le ee ee soe ek eines sesse Pa

OB. Hore.. 1. OO eee Findlay

a Pe ee ey es BOS er aie iss eat as | Winter

646 THE AMERICAN NATURALIST [Vou. L

Species | ABlood | A Water Locality Observer

XII. Mammalia: | |

BO OY RA ET A | 0.601 AR E E er yah cee Findlay

ri rA AE Sea a ee cele E Pe 1, Sree NG PEE ee Ree Winter

PR PN a rd evo ols 0.625 Buics CoC pind Wy UH ee oats Findlay

Wl es ee CPO E ee BiG ae ie ae cere ee

E l EGGS BNR SAS aoe a ee er Wee. Soe eee Findlay

RE ee invite gis Wika Oc ida Se Se nter

106. Rabbit Pee ee oe bey ies wee aes Findlay

107. Pumas Tees ey a 0.57 A E T Mas N E AE inter

108. Tea A yews eas tee es Pe es ee er eo Bottazzi &

| Ducceschi

100. Shoop ss nnn a 0.55 BP piper es E EGE E Ae Winter

SIU Cag eae pues aes 0.615 TICES SE BARN Seer trae a aang eS Findlay

S345 Min oe a a Eo a rele eS ap ee ey

term think of the fluid circulating in the blood vessels of a

vertebrate. The term body fluid is also ambiguous. In

an invertebrate it has reference to that part which we call

the blood of a vertebrate. In the vertebrate we usually

think of the secretion of serous membranes as ‘‘body

fluid.’’ After all, the subject of discussion in this paper

is the fluid by which food is carried to tissues and wastes

carried away. Having thus defined the use of the terms,

let us examine the osmotic pressures of the blood of vari-

ous animals.

Table I, which follows, shows one hundred and eleven

determinations of the osmotic pressure of the blood of

representatives of nearly every animal phylum. Many

of these determinations are averages. Some of the forms

are wholly terrestrial, some live in fresh water, some in

either fresh or seawater, some live wholly in the sea.

Considerable variation in the osmotic pressure of the

blood is shown.

Of the marine forms given some are found in the Medi-

terranean, while others for the most part occur in the

ocean or in protected waters connected with it. There is

great variation in the osmotic pressure of the blood of

forms living exclusively in the Mediterranean. Great

variation is shown in the case of those living in the ocean.

In some cases in each environment, complete harmony

with or rather isotonicity with the environment is ap-

parent. In other cases this is not at all evident. For

No. 599] OSMOTIC PRESSURE OF THE BLOOD 647

example, the average A of twelve species of invertebrates

from the Mediterranean is 2.281°, while the average A of

the water in which they live is 2.29°. A simple case of

adaptation is thus evident. But the bony fishes, teleosts,

tell a different story.

It is worth while to contrast the osmotic pressure of the

blood with that of the external medium. To do this we

will break up all the forms into groups not according to

the environment alone, but also according to relationship.

If we should be guided by environment alone, the result

would be a confused tangle. Table II shows the average

A of these groups selected not only on the basis of re-

lationship but also taking into partial consideration the

environment.

‘*A,’? Blood, 12 Invertebrates, Mediterranean = 2.28° —Water

‘A,’’ Blood, 1 Cyclostome, Ocean, bay :

‘*A,’? Blood, 8 Elasmobranchs, Mediterranean = 2.346 —Water = 2.29

“tA,” Blood, 6 Elasmobranchs, Sen bays = 1.902 —Water = 1.85

‘‘A,’’? Blood, 4 Teleosts, Mediterranean = 1.054 —Water = 2.29

‘*A,?? Blood, 32 málaði Ocean, ete. = 0.744 —Water = 1.82

*fA,’’ Blood, 13 Teleosts, Fresh water = 0.545 —Water = 0.03

**A,’? Blood, 4 Amphibia, Fresh water = 0.551 —

FAST Blood, fees. sive trs coke REODE.

“*A,’? Blood, 3 Ave ise ewesy dds on O0T ee

**A,’’ Blood, 8 a E anly hs = 0.577 —

From this table it is evident that the blood of the marine

invertebrate is isotonic with the water in which it lives,

whether this be the Mediterranean or the ocean. As

stated above, it appears to be a simple case of adaptation.

But in the other cases the relation is not so simple. I

we compare the osmotic pressure of the marine teleosts,

fresh-water teleosts and the amphibia, etc., with the os-

motic pressure of the external medium great differences

are evident. And yet it can not be said but what all these

forms are adapted to their environment. But it is not

enough to make this statement, but to try to explain why

such a relationship becomes possible. The isotonicity

existing between the blood of marine invertebrates and

648 THE AMERICAN NATURALIST [ Vor. L

their environment has been discussed by Fredericq (’85-

04), Rodier (’99), Dakin (’08), Garrey (’05) and Bot-

tazzi (’97-’06). Now it is held that evolution of life

began in the sea. The single celled forms were completely

surrounded by the sea and it is easily understood why the

osmotic relations would remain primitive in case of these

forms. In gastrula type animals, such as celenterata,

practically all cells of the body are bathed directly by the

sea and as far as we know these forms also are in osmotic

equilibrium with sea water. Now with the appearance of

mesoderm and a body cavity much of the body is removed

from direct contact with the sea. But the complete equi-

librium remains. As Quinton (’00) says, the marine in-

vertebrate, though anatomically independent of the sea in

many of its organs, yet it is still physiologically open to

the sea which in an osmotic sense still ebbs and flows

throughout its body.

Protoplasm originating in the sea was built up with

certain relationships with sea water, which relationships

are still maintained throughout all marine invertebrates.

May not the sparsity of fresh-water porifera and ccelen-

terates and the comparative failure of fresh-water alge

be due to the difficulty of maintaining the integrity of

protoplasm when all cells of these forms are so freely

bathed by fresh water, the osmotic pressure of which is

nearly zero?

Next above the marine invertebrates is a single case of

a cyclostome which is in osmotic equilibrium with the sur-

rounding sea water. What the osmotic pressure of the

blood of a cyclostome in fresh water is, we have no record.

It should be noted here that cyclostomes are now regarded

as degenerate fishes and on that account any evidence

from these forms as to the higher course of evolution

must be treated with care. In the next place we find that

eight species of elasmobranchs from the Mediterranean

and six from the ocean possess blood: which is practically

isotonic with the sea water outside. Apparently they do

not differ from the marine invertebrates. But it is evi-

dent that the osmotic pressure of the blood is slightly

No. 599] OSMOTIC PRESSURE OF THE BLOOD 649

greater than that of the external medium. Furthermore,

analysis shows that the osmotic pressure of elasmobranch

blood is due to different substances from those which ac-

count for the osmotic pressure of the blood of marine

invertebrates. Therefore the elasmobranchs belong to a

second category. In the third group we will place the

marine teleosts. The osmotic pressure of their blood is

somewhat less than half that of the medium in which they

live. We have the case of four species from the Mediter-

ranean and thirty-two species from the ocean which show

this. The osmotic measurements show a decided differ-

ence between the blood and the surrounding medium. A

decided independence also. In the same group or pos-

sibly a fourth group we will place the fresh-water fishes

and with these the amphibians, reptiles, birds and mam-

mals. Thirteen species of fresh-water fishes possess blood

with an osmotic pressure less than that possessed by the

marine teleosts. Let us assume here that the fresh-

water fishes were derived from marine ancestors. In be-

coming acclimated to fresh water, the blood suffered a

decrease in its osmotic pressure. Whether this was in

direct response to the great decrease in the osmotic pres-

sure of the surrounding medium as compared with sea-

water is problematical, but appears probable. The am-

phibians were derived from the fresh-water teleosts.

Some of the amphibians still retain their aquatic habits

and structures. They in all probability possess the 05-

motic pressure of fresh-water fishes. Other amphibia

metamorphosed into terrestrial forms, taking with them

the osmotic pressures of the blood possessed by their fish-

like ancestors. Blood with the same osmotic pressure as

that of the fresh-water fishes flows on through the am-

phibia to the reptilia and on to the birds and mammals.

An examination of Table II shows the close similarity

between the osmotic pressures of fresh-water fishes, am-

phibians, reptiles, birds and mammals. According to the

above hypothesis, the order of evolution was I. Marine

invertebrates, II. Elasmobranchs, III. Marine teleosts,

650 THE AMERICAN NATURALIST [ Vou. L

IV. Fresh-water teleosts, amphibians, reptiles, birds and

mammals.

Let us examine each of these groups with regard to

their osmotic independence of the external medium.

That is, what is the effect of changes in the concentration

of the external medium on the osmotic pressure of the

blood of these groups.

First, the invertebrates. Let us recall Quinton’s state-

ment that marine invertebrates are still physiologically

open to the sea. For when the concentration of the ex-

ternal medium is changed, it is found that a change in the

osmotic pressure of the blood takes place. Fredericq

(785 and ’04) stated that the change in one was followed

by an equal change in the other. In a few hours the new

equilibrium is established. If the time of sojourn in the

modified sea water was small the equilibrium with it was

not completely attained. Moreover, all invertebrates did

not adapt themselves with the same rapidity to changes

in the external medium. On the whole, provided the ex-

ternal change was not too great, it was followed in time

by complete equilibrium between the osmotic pressure of

the blood and that of the modified sea water. This was

true in the case of sea water made dilute by addition of

fresh water and sea water made more concentrated by the

addition of salt. In other words, the organism possesses

no structures which render it independent of the changes

in the external medium. There are three structures con-

cerned in these changes. First the integument, second,

the intestinal wall and third the gill membranes. With

the appearance of gills, the body integument apparently

is the first structure to become impermeable. The in-

testinal wall is the first to show a selective action.

Second, the elasmobranchs. These had been placed by

investigators with the marine invertebrates not only be-

cause their blood possessed the same osmotic pressure as

the external medium, but it was thought that when the

external medium was changed, the same changes occurred

in the blood of the elasmobranch. I made extensive ex-

No. 599] OSMOTIC PRESSURE OF THE BLOOD 651

periments to test this (713) and found that when a change

was made in the external medium, though considerable

change took place in the blood of the dogfish, yet it was

considerably less than the external change. In fact it

appeared as though the change in the blood was roughly

proportional to the change in the external medium (p. 20,

Scott, 713). The condition was so marked as to show

clearly that the elasmobranch belonged in a category dif-

fering from that of the marine invertebrate.

Third, the marine teleost. Much emphasis has been

placed upon the claim that these forms are absolutely

independent of changes in the external medium. With

this claim, I must differ. The following evidence is the

basis of this difference of opinion. In the first place

Tables I and II show that the blood of teleosts from the

Mediterranean has a higher osmotic pressure than that of

blood of teleosts from the ocean. There is a correspond-

ing though greater difference in the osmotic pressure of

the water. Dakin ’08 in a trip from Kiel to Helgoland

found that the osmotic pressure of the sea water increased

74 per cent. and that the osmotic pressure of the blood of

the plaice showed an increase of 20 per cent. The cod did

not show so great a difference, being but 4 per cent.®

Garrey ’05 reported A of the blood of the tantog at Woods

Hole to be 0.86° while at the New York Aquarium, where

the harbor water is much more dilute than at Woods Hole,

I found the A of tautog blood to be about 0.70°. There-

fore it would appear that even blood of the marine teleost

is somewhat modified by changes in the external medium.

And yet practical independence has been achieved. This

is evident from the fact that the marine teleost lives in a

medium which has an osmotie pressure over twice as

great as that of the blood of the fish. ‘

Macallum (710) has explained the peculiar osmotic pres-

sure of the blood of marine teleosts as due to their origin

from fresh-water teleosts. This is based on morpholog-

3 On the other hand Dekhuyzen, 705, found a difference of 20 per cent,

in the osmotic pressure of cod blood according to the locality from which

the fish was taken.

652 THE AMERICAN NATURALIST [Vou. L

ical evidence of the evolution of the true teleosts from

ganoid ancestors from the elasmobranchs through forms

similar to the sturgeons and the bow-fins. I doubt very

much, however, whether ichthyologists would wish to con-

clude on this basis that all marine teleosts had their

origin from fresh-water forms. In fact certain paleon-

tologists trace the evolution of certain fresh-water tele-

osts from ancestral marine teleosts. The sea is the home

of the preponderating fish population. Here the class of

Pisces has found its greatest opportunities for range of

movements to escape enemies, in search of food or place

of breeding.

Facts concerning the osmotic pressure of the blood of

anadromous fishes throw light as to the possible if not

probable origin of fresh-water forms. Greene (’04) de-

termined the osmotic pressure of the chinook salmon in

Monterey Bay to be 0.76°. On the spawning grounds in

fresh water its blood had a A of 0.628°,a decrease of 17.6

per cent. Flatfish are known to be somewhat anadromous.

Dakin (’08) found the A of the flounder, Pleuronectes

flesus, in the North Sea to be 0.83°, while in the River

Elbe in fresh water its blood had a A of 0.68°, a decrease

of 18 per cent. The same author found that the blood of

the eel, Anguilla, in fresh water had a A of 0.57°, quite

similar to that of fresh water fishes. After a day in sea

water another specimen had blood with a A of 0.745°.

Eels taken from seawater had blood with a A of 0.634°.

Eels from seawater placed in fresh water for three days

possessed blood with aA of 0.582°, practically the same as-

for fresh-water forms. At Woods Hole, ignorant of this

work of Dakin’s, I made observations on the 4 of the

blood of the white perch, Morone americana. This form

can live equally well in salt or fresh water. Taken from

the slightly brackish waters of Tashmoo Pond, Marthas

Vineyard, Massachusetts, the blood showed a A of 0.635°.

The upper end of this pond is the source of drinking

water for Oak Bluffs. A number of perch were placed

in running tap water for a day, when the blood showed a

No. 599] OSMOTIC PRESSURE OF THE BLOOD 653

A of 0.571°, similar to the fresh-water fishes. Others of

this lot were placed in sea water for two days, when the 4

of their blood was 0.766°. Others taken directly from the

Eel Pond (sea water) showed a A of 0.735°. The result

is similar to Dakin’s. On the whole the conclusion seems

justified that’ anadromous fishes are able to adapt them-

selves to a degree to the great changes in the osmotic

pressures of the external medium, which they meet in

passing from salt to fresh water or vice versa by a slight

corresponding change in the osmotic pressure of the

blood.

It is commonly known that sturgeons are anadromous.

For some reason the elasmobranch has been shut out of

fresh water. There is but one elasmobranch known to

inhabit fresh water, Carcharias nicaraguensis of certain

lakes in Nicaragua. Although the integument of the

shark is impermeable, yet I have found the gills to be still

permeable to salts (Scott & Denis, ’13). The ganoids de-

rived from elasmobranchs ventured up fresh-water

streams. They returned to the sea. Rodier (’99) states

the A of the blood of Acipenser sturio to be 0.76°, which

places it in the same group as the marine teleosts. What

the A is in fresh water is not known. The modern

sturgeon is a long way from the modern shark. Never-

theless it is conceivable that the ancestral ganoids tried

fresh-water conditions. Is it not possible that these con-

ditions, fresh water and food found in fresh water had

some influence on the change in structure. During all

subsequent periods when evolutionary changes were tak-

ing place some forms went back and forth from sea to

fresh water. Some forms remained in fresh water. Dur-

ing this period of experimentation, impermeable mem-

branes were built up. In the meantime the blood had

become modified, due to the temporary sojourn in fresh

water. The osmotic pressure was reduced; the mem-

branes once made practically impermeable remained S0,

and when those forms returned to the sea and remained

there they retained almost unmodified the osmotic pres-

654 THE AMERICAN NATURALIST [ Vou. L

sures they had acquired during their fresh-water experi-

ence. We can thus speculate that in some such way the

present osmotic pressures of the blood of marine and

fresh-water teleosts were acquired. Whatever may be

the case with the marine and fresh-water teleosts, it is

more clearly indicated that the osmotic pressure of the

blood of terrestrial forms is derived from fishes which

lived in fresh water. The present day anadromous fishes

constitute all that remains of a movement which at one

time was far more general.

The chemical composition of the blood throws further

light on the question. The osmotic pressure is due to

. substances dissolved in the blood. These are chiefly salts.

Quinton (’00) states that sodium chloride represents from

85 per cent. to 90 per cent. of all the dissolved salts of the

blood. The sodium chloride content can be ascertained

from a study of the chlorides which are easily deter-

mined. Let us ascertain the changes in the sodium chlo-

ride content of the blood of the forms under discussion.

In the first place what is the total salt content of sea water.

According to Bottazzi (’97) the total salt content of water

from the Mediterranean is 3.78 per cent. The water of

the ocean contains about 3.22 per cent. salts. Of course

there is some variation. The percentage of salts in fresh

water is very small, 0.05 per cent. (Sumner, 705). What

is the percentage of salt of the blood of forms living in

the sea? Quinton (’00) made forty-nine determinations

of the sodium chloride content of the hemolymph of ten

species of marine invertebrates belonging to five different

groups and found that these averaged 3.24 per cent. He

made 26 determinations of the sodium chloride content of

the sea water in which these forms lived, and found that it

was about 3.31 per cent. According to these researches

of Quinton, the blood of the marine invertebrate contains

about the same percentage of salts as the water in which

they live. Moreover, it follows that the osmotic pressure

of the blood is determined almost wholly by the salts of

the blood and not by any organic solutes. It was because

No. 599] OSMOTIC PRESSURE OF THE BLOOD 655

of this relationship that Quinton felt justified in making

the statement that the marine invertebrate while anatom-

ically closed to the external medium, is yet physiologically

open to it. That functionally speaking the marine in-

vertebrate is still freely exposed to the sea without, which

still practically ebbs and flows through its body.

Macallum (’10) says:

In Limulus, the amount of total salts in the blood, 2.982 per cent.,

approaches that of the sea water,—which may be found along the Atlantic

coast. At St. Andrews, New Brunswick, the total salts of the seawater

collected in April were 2.417 per cent., but in sea water collected in

August, 3.165 per cent. In the blood of the lobster, the total salts as

ascertained were 2.852 per cent., which is between the two concentrations

given above for the salinity of the sea water at St. Andrews where the

lobsters from which the blood was taken were obtained. The blood of

Limulus is but slightly modified sea water. It would appear as-if the

sodium chloride of sea water passes freely into the blood of the lobster

till the sodium chloride concentration in both is approximately balanced.

This agrees entirely with the work of Quinton. For

some reason, the marine invertebrate has not been able to

' keep the sea out. One asks why the question of the per-

meability of membranes of fishes to salts is of such inter-

est to the comparative physiologist? One answer is that

impermeability represents independence of the sea the

osmotic pressure of which differs so much from that of

fish blood. And this independence is not to be found

among the marine invertebrates.

As shown above, elasmobranch blood possesses the

same osmotic pressure as that of the marine invertebrate

and that of the sea without. . But analysis shows that the

osmotic pressure of elasmobranch blood is due to entirely

different causes. For example, what is the salt content

of elasmobranch blood? It should contain about 3.22 per

cent. salts in order that its total osmotic pressure be due

to salts. But Rodier (’99) found that the blood of elas-

mobranchs did not contain over 1.7 per cent. sodium

chloride. Dakin (’08) found the blood of the dogfish to

contain but 1.45 per cent. sodium chloride. My analysis

of the blood of another species, Mustelus, at Woods Hole

656 THE AMERICAN NATURALIST [Von L

showed 1.424 per cent. sodium chloride. Fredericq (’04)

found the blood of Scyllium to contain but 1.71 per cent.

salts, while Macallum (710) found the blood of the dog-

fish, Acanthias vulgaris, contained 1.7739 per cent. sodium

chloride. In other words the sodium chloride content of

the blood of elasmobranchs will account for only about

half of its total osmotic pressure. Evidently a great

change has come about. ‘‘The difference between the A

of the serum and that due to salts of the serum depends,’’

as Macallum (’10) says,

“on urea and other organic solutes.” Urea is present in large quanti-

ties in the blood of elasmobranchs.

Staedeler and Frerichs (58) obtained as much as two ounces from the

the liver of a single shark. In ’90 von Schroeder found that Scyllium,

another dogfish, contained blood with 2.6 per cent. urea. Rodier (’99)

computed that one third the osmotic pressure of the blood of elasmo-

branchs was due to urea.

In 713, I found that Mustelus blood contained 1.55 per

cent. urea. Macallum (710) in Acanthias vulgaris found

an average of 2.026 per cent. urea. Due to dissociation,

the salts have twice the osmotic pressure, approximately,

as the urea, although the urea and salts are present in

about equal quantities. But the urea and salts are not

sufficient to account for the osmotic pressure of the blood.

The difference is due to the presence of ammonia salts, as

Macallum found. For example, he found 0.1727 per cent.

ammonia in the serum of the dogfish. This would fully

account for the remaining percentage of the depression

of the freezing point unaccounted for by the presence of

the salts and urea. So that we see, that while super-

ficially the elasmobranch resembles the marine inverte-

brate in the osmotic pressure of the blood, yet below the

surface a marked change has taken place. Several ob-

servers had noted that the osmotic pressure was slightly

greater than that of the sea water. This at least is

another indication that the equilibrium is not like that

existing between marine invertebrates and the sea. For

some reason the elasmobranch has lost in salts. Their

place has been taken by nitrogenous solutes. The con-

No.599] OSMOTIC PRESSURE OF THE BLOOD 657

dition is lacking in the marine invertebrate. Some one

has characterized the jellyfish as organized sea water.

According to Macallum the blood of Limulus is but

slightly modified sea water. The blood of the marine in-

vertebrate has remained at the same low level so far as

the presence of nitrogenous compounds is concerned. To

what may we ascribe this new condition? Is it due to

great proportion of nitrogenous food? To the partic-

ular kind of liver? To the great development of the mus-

cular system? To a peculiar function of the kidney?

Questions can at present be asked only. We lack infor-

mation as to certain aspects of elasmobranch physiology.

However much the elasmobranch may have experi-

mented in the matter of unique nitrogenous content of the

blood, it is certain that this condition is lacking in the

teleosts. And the lack there is carried over to the forms

which developed further. For the osmotic pressure of

the blood of teleosts is again determined almost wholly

by the salts present. The salt content of the blood of

marine teleosts is considerably less than that of elasmo-

branchs. Quinton (’00) found the blood of eight species

of marine teleosts to contain 1.076 per cent. salts. Rodier

(799) found that the blood of the ganoid, Acipenser sturio,

had a salt content varying from 0.643 per cent. to 0.979

per cent. The blood of Lophius, a strictly marine form,

contained 1.164 per cent. salts. eHamburger states that

teleost blood contains 0.936 per cent. salts, but whether

these are fresh-water or marine species is not stated.

Macallum (710) found that the blood of the cod, Gadus

callarias, contained 1.2823 per cent., while that of the pol-

lock, Pollachius virens, contained 1.2934 per cent. salts.

It is evident, therefore, that the percentage of salts in the

blood of the marine teleost has been decreased as com-

pared with the total saline content of elasmobranch blood.

Moreover, the osmotic pressure of the blood of the teleost

is due almost wholly to the salts present. Macallum (710)

proved this. He found that the 4 of the salts of cod blood

was 0.71°, while that of the entire blood was 0.765°.

658 THE AMERICAN NATURALIST [Vou. L

There is a difference of but 0.055°. The A of the salts of

the blood of the pollock was 0.737° while the A of the en-

tire blood was 0.825, showing a difference of but 0.088°.

In other words, almost the entire osmotic pressure of the

blood of the teleost is due to the salts. The urea, am-

monia or other organic solutes present must be very

small and are represented by the difference above men-

tioned, namely 0.055° in the case of the cod and 0.098 in

the case of the pollock. How different is this condition

from that found in the elasmobranch where in one case

noted by Macallum, and which is typical, the difference

between the A of the saline contents of the blood and the

entire blood was 0.961°, a difference as great as the aver-

age A of the marine teleost and as stated due to the

relatively enormous amount of urea and other organic

solutes in the blood of the dogfish. Again the question

arises: What brought about this change between the com-

position of elasmobranch blood and that of the teleost?

Was it due to the migrations to and from fresh water be-

fore certain species of teleosts took up their home per-

manently in the sea? And yet the marked difference be-

tween the two is not alone a difference in salt content. It

is far more the absence from the blood of urea, ammonia

and other organic solutes. Let us use Macallum’s data

as a basis for comparison. The blood of marine teleosts

contains about 30 per cent. less salts than the blood of

elasmobranchs but it contains 90 per cent. less organic

solutes. The distinct loss therefor is in organic solutes.

This therefore must have been a significant factor in the

evolution of the higher form. Now what is the most ap-

parent structural difference between the elasmobranchs

and teleosts? Itis of course that the skeleton of one con-

sists of cartilage and the skeleton of the other is bone?

It does not necessarily follow, however, that the power to

build a bony skeleton depends on the absence of organic

solutes from the blood, nor is there apparently any close

connection between them.

The fresh-water fishes in all probability agree with the

No.599] OSMOTIC PRESSURE OF THE BLOOD 659

marine teleosts in low percentage of organic solutes and

this characteristic is maintained by all the higher forms.

Dakin found that the blood of the plaice at Helgoland con-

tained 0.92 per cent. salts, while at Kiel in brackish water

it had a salt content of 0.85 per cent. Mosso (’90) stated

that marine teleost blood had a higher salt content than

that of fresh-water forms. Dakin (’08) found the blood

of the eel in sea water to contain 0.605 per cent. salts,

while in fresh water its saline content was 0.466 per cent.

Quinton (’00) found that the blood of fresh-water teleosts

contained 0.7 per cent. salts. Atwater (’91) found that

the flesh of fresh-water teleosts contains less salt (15 per

cent. less chlorine) than that of marine teleosts. Sumner

(705) obtained a similar result.

The anadromous fishes possess blood that is less saline

in fresh water than in sea water. It is also true that

strictly marine teleosts of the present day vary a little in

the saline content of their blood when the salinity of the

external medium changes. These facts indicate that the

decreased salinity of the blood of fresh-water teleosts was

brought about in response to the low saline content of the

external medium. During the migrations that took place

in the past when there were probably more anadromous

fishes, this diminution in salts took place. Those forms

that remained in fresh water retained the percentage of

salts they acquired by their sojourn in fresh water. At

the same time they built up membranes which maintain

an equilibrium in spite of the differences in the osmotic

pressure of the blood within and the fresh water without.

Similar membranes were formed in case of the marine

teleosts, which maintain an equilibrium with the sea water

in spite of the fact that the osmotic pressure of sea water

is over twice that of the teleost blood. The evidence at

hand indicates that the last membranes to become prac-

tically impermeable to salts were the gill membranes.

And yet though impermeable to salts they still are re-

quired to be permeable to gases. a

Now the blood of amphibia contains about 0.7 per cent.

660 THE AMERICAN NATURALIST [Von L

salts. This closely resembles that of fresh-water fishes.

The blood of mammals contains a slight increase in its

saline content. Bunge (’89) states that human blood

serum contains about 0.84 per cent. to 0.86 per cent. salts.

Macallum (710) calculating from Abderhalden’s analyses,

concluded the total saline content of the blood of the dog

amounted to 0.935 per cent., that of the cat to 0.933 per

cent. and that of the sheep to 0.905 per cent. To quote

from Macallum: ;

In mammals, according to Abderhalden’s analyses, there is an extraor-

dinary similarity in the inorganic composition of the serum of the num-

ber of the forms taken and the ratios of the sodium, potassium, calcium,

and magnesium are almost parallel with those in the Teleosts and

Elasmobranchs.

Macallum had an opportunity to analyze the blood of

‘‘the whale common in the Pacific off the coast of British

Columbia,’’ and the parallelism between the inorganic

constituents of its blood and that of the horse and pig was

remarkable, thus bringing the whales very close to the

Ungulates to which some anatomists relate them.

The above studies of the osmotic pressures of the blood,

the change in the permeability of the protecting mem-

branes and the inorganic and organic composition of the

blood are understood only by viewing them from the

standpoint of evolution. The increase in saline content

of mammalian blood as compared with amphibian and

fresh-water teleosts can be ascribed to the regulative ac-

tion of the kidney. Most investigators give the impres-

sion that the osmotic pressure of the blood of animals is

definite and fixed. This is not true. Findlay calls at-

tention to the variation in the osmotic pressure of human

blood at different times of day. For example, a distinct

though slight rise (0.03°) is noted after meals. This

question needs further study. My investigations showed

that Mustelus canis can pass with entire safety through a

range of 0.15° (+ and — ) in its osmotic pressure. The

range through which invertebrates can pass is much

greater. The observations of Dekhuyzen (705) and Dakin

(708) show that the range becomes limited in the case of

No.599] OSMOTIC PRESSURE OF THE BLOOD 661

marine teleosts. The range is very much more restricted

in fresh-water teleosts and higher forms. Protoplasm is

an ever-changing substance. There is a constant ebb and

flow. Protoplasm of the higher forms has evolved through

long ages to a condition wherein it is associated with the

same salts it was entirely surrounded by when it first

began to be. The amounts have changed, but the propor-

tions have remained quite constant and the kinds have

remained the same as those in the sea. And that is why

the surgeon must inject a 0.9 per cent. saline solution into

the veins of his patient suffering from hemorrhage. And

that is why human blood has a certain osmotic pressure.

Macallum ascribes the first great reduction in salts which

took place in the elasmobranch to be due to the kidney,

whose primary function was not the elimination of the

wastes of metabolism, but the regulation of the concentra-

tion and composition of the salts of the blood. The elas-

mobranch kidney is very inert and sluggish in the matter

of the elimination of the organic wastes. The teleosts

acquired the habit of still further keeping down the saline

content while at the same time they eliminated the urea

readily. However, I do not see that the process is neces-

sarily limited to the kidneys alone. A thorough study of

the elasmobranchs and teleosts is needed to throw light

on this puzzle. I can see why the migratory habits of

teleosts or teleost ancestors (ganoids) would account for

reduction in salt content of the blood, but this throws no:

light on the reduction of salts in elasmobranch blood as

compared with invertebrate blood. Nor does Macallum

indicate any use the large amount of urea might serve.

Balgioni (06) found that salt solution alone led to stop-

page of the elasmobranch ventricle in diastole. It in-

creased diastolic tonus, while urea increased systolic

tonus. The presence of the two in about equal amounts

mutually neutralized each other and made the continuous

rhythm of the heart possible. All we can say is that for

the kind of protoplasm of which the elasmobranch heart

is composed, the urea is a necessary constituent of the

blood. Furthermore it does not appear to be necessary —

662 THE AMERICAN NATURALIST [Vou L

for the teleost heart. At any rate we are aware that once

we begin to question further, the necessity of further

knowledge becomes evident. This paper can be brought

to a close in no better way than by quoting a statement

made by Claude Bernard (’65). We may accept it as one

of the laws of evolution and conclude that inquiries con-

cerning the osmotic pressures of the blood of animals

amply prove its truth.

Chez tous les étres vivants, le mileau intérieur, qui est un produit de

Yorganisme, conserve des rapports nécessaires d’éxchanges et d’équilibre

avec le mileau cosmique extérieur; mais, à mesure que |’organisme

devient plus parfait, le mileau organique se spécifie et s’isole en quelque

sorte de plus en plus du milieu ambiant.

BIBLIOGRAPHY.

Atwater, W. O. The Chemical Composition and Nutritive Value of Food

Fishes so tae Invertebrates. Report of U. 8. Fish and Fisheries,

p. 679,

Baglioni, s pa Kenntniss des N-stoffswechsels der Fische. Zentralblatt

Baglioni, 8. Beiträge z allegotnatis inen Physiologie des Herzens. Zeit-

schrift f. Allgemeine pieg VL LET

Bernard, C. Introduction a Vetude a la CER expérimentale, Paris,

1865.

Bottazzi, P. La pression osmotique du sang des animaux marins. Arch.

iol., XXVIII, s. 61, 1897.

Botsenal. F. Sulla teipulastone della pressione osmotica negli organismi

animali. Arch. Fisiol., III, p. 416, 1906.

Bottazzi, F. Koranyi und Richter. Physikalische Chemie und Medizin.

Erster Band, p. 475, Leipzig, 1907.

Bottazzi, F., and Ducceschi. Résistance des erythrocytes alcalinité du

plasma et pression osmotique du sang des différentes classes des verté-

brés. Archiv. ital. biol., XXVI, p. 161, 1897.

Bottazzi, F., and Enriques, P. Ueber die Bedingungen des diio

Gleichgewichts und des Gleichgewichtsmangels zwischen den organi-

Arch. für Anat. wnd Physiol., Sup. Band., p. 109, 1901.

Bunge. Lehrbuch der Physiologischen und Pathologischen Chemie, Leipzig,

p. 120, 1889.

Dakin, W. J. Osmotic Concentrations of Blood of Fishes Taken from Sea

Water of Naturally Varying Concentration. Bio-Chemical Journal,

Liverpool, III, p. 258, 1908.

Dakin, W. J. Variations in the Osmotic Concentration of the Blood and

Coelomie Fluids of Aquatic Animals Caused by Changes in the External

Medium. Bio-Chemical Journal, Liverpool, III, 10, p. 473, 1908.

No.599] OSMOTIC PRESSURE OF THE BLOOD 663

Dekhuyzen, M, C. Ergebnisse von Osmotischen Studien, Bergens Museum,

arborg,

Dekhuyzen, M. C. ton Akad. van Wetensch., Amsterdam, vol. 8, p. 587,

1905.

Dekhuyzen, M. C. Sur la pression osmotique dans le sang et dans 1’urine

es poissons. Arch. nier., X, p. ’ 5.

Findlay, A. Physical Chemistry and its Ppt pee in Medical and Bio-

logical Sciences. Longmans, pp and Co.,

Fredericq, L. Influence du mileau ambiant sur re composition du sang

des animaux aquatiques. eats de Zool. Experiméntale, 2d series,

III, Notes et Revue, p. 34, 1885.

Petric, L. Sur la concentration moleculaire du sang j ey tissus chez

les animaux aquatique. Archiv de Biol., XX, p. 709,

Garrey, W. E. The Osmotic Pressure of Sea Water and ae ye Blood of

Marine Animals. Biological Bulletin, VIII, 4, p. 257,

Greene, C. W. Physiological Studies of the Chinook Salmon. Bull, U. 8.

Bureau Fisheries, XXIV, p. 431, 1

Hamburger, H. J. Ueber die Salz und Robrzncker-Losangen hewirkten

Veränderungen der Blutkörperchen. Archiv fiir Anat. und Phy siol.,

Pp. 94, .

Macallum, A. B. The Inorganic Composition of the Blood in Vertebrates

and Invertebrates and its Origin. Proceedings of the Royal Society,

B. 82, p. 602, 1910.

Mosso, Z. Ueber verschiedene Resistenz der Blutkörperchen bei verschie-

denen Fischarten. Biologische Centralblatt, Bd. X, p. 570, 1890.

Quinton, M. R. Communication osmotique chez 1 invertébrés marins normal

entre le mileau intérieur 1’animal et le mileau extérieur. Compt. rend.,

131, p. 905, 1900.

Quinton, M. R. Permeabilité de la paroi extérieur de l’invertébré marin

non seulment a l’eau mais encore aux sels. Compt. rendus, 131, p.

952, 1900.

Rodier, E. Observations et expériences comparatives sur l’eau de mer le

sang et des liquides internes des animaux. Travaux des. labor. de la

stat zool. d’Arachon, p. 103, 1899.

Scott, G. G. A Physiological Study of the Changes in Mustelus canis

Produced by Modifications in the Molecular Concentration of the Ex-

ternal Medium. Annals New York Academy Sciences, Vol. XXIII, pp.

1-75, 1913.

Scott, G. G., and W. Denis. The Relation of Osmotic Pressure to Absorp-

tion Phenomena in the Dogfish. American Journal of Physiology,

XXXII, No. 1 1, 1913.

Staedeler salt Fre w Journal für Pract. Chem., Vol. 73, p. 48, 1858.

Sumner, F. B. Physiological Effects upon Fishes Changes in the

Density and Salinity of Water. Bull. U. S. Bureau of Fisheries, XXV,

p. 1905. 2 :

von Bakroodie, W. Ueber die Harnsstoffsbildung der Haifische. Zeitschr.

Winter, J. De la ETRA moleculaire des liquides de A Or

Arch. de Physiol., VIII, p. 114, 1896.

THE GENETIC BEHAVIOR OF MICE OF THE

COLOR VARIETIES ‘‘BLACK-AND-TAN”’

AND “RED”?

L. C. DUNN

Bussey INSTITUTION

Harty in 1914 there were received at the Bussey Insti-

tution certain stocks of mice obtained from fanciers in

England. Some preliminary studies of the mice were

made by Professor Castle and Dr. Little. A more inten-

sive study of one race, the black-eyed-white, was after-

ward made by Dr. Little and independently by Dr. Detlef-

sen. The remaining stocks were turned over to Mr. W. F.

Whittier, who carried on experiments with them partly

at the Bussey Institution, partly at the Massachusetts

Agricultural College, recording some 2,500 offspring.

After devising the color grading scale and the general

methods followed in the later experiments, he relinquished

the work to the present writer. Since that time about

2,000 young have been recorded, bringing the total to

4,500. All the work has been done under the advice and

direction of Professor Castle.

The principal varieties which have been used in these

experiments are known in England as ‘‘black-and-tans’”’

and ‘‘reds.’’ The genetic character of these mice was at

the outset quite unknown, and in this paper it is proposed

to give some account of their genetic behavior, and since

they have proved to be forms of yellow mice, to assign to

them and their derivatives places in a scheme of classifica-

tion of the yellow varieties.

The black-and-tan race has presented throughout the

more interesting problem. In appearance these mice are

of an intense shiny black dorsally, with a belly super-

ficially clear yellow. The belly hairs, however, are marked

by having dull black bases, hidden by the longer and over-

664

No. 599] THE GENETIC BEHAVIOR OF MICE 665

lying yellow areas of the hairs. Yellow-ticked hairs are

occasionally seen on flanks and head, encroaching on the

black pigmented parts. This peculiarity increases some-

what with age, but never to such an extent as to make the

body color predominantly yellow.

When bred inter se, they have been found invariably to

be heterozygous, no homozygous black-and-tan having

been discovered among a dozen individuals tested by suit-

able matings. As recessives they have given all-black

mice more intense than any we have seen derived from

other sources. Forty-two matings infer se of pure-bred

black-and-tan parents produced 148 young, an average of

3.52 to a litter. Of these young 93 have been black-and-

tan and 55 black, a ratio of 1.69:1. This approximates a

2:1 ratio more closely than the 3:1 ratio usually given by

Mendelian heterozygotes. The black recessives breed true,

and when mated to black-and-tans have produced equal

numbers of black-and-tan and black young (18:18). The

approximation of a 2:1 ratio in matings of black-and-tans

inter se shows their gametic similarity to yellow mice

whose unfixable nature was first shown by Cuénot (703).

Figures given by this author combined with those given

by Castle and Little (710), by Little (710 and ’11) and by

Miss Durham (’11) total 2,673 young produced by yellow

X yellow matings. Of these 1,783 were yellow and 890

non-yellow, a ratio of almost exactly 2:1.

The small average size of litters produced by black-and-

tan parents mated inter se gives added evidence of their

resemblance to yellow. Castle and Little (710), ın con-

firmation of Cuénot’s observations, showed that yellow X

yellow matings produced litters of smaller average Size

(4.71) than yellow X non-yellow (5.57), and following

Cuénot they attributed the difference to absence of homo-

zygous, yellow-yellow zygotes. The 2:1 ratio and the

small-sized litters serve also to relate the black-and-tans

with Castle’s “‘sooties”’. and Miss Durham’s t‘ sables,

both of which were shown to be heterozygous yellows

carrying black as a recessive.

666 THE AMERICAN NATURALIST [ Vou. L

The reds, by their appearance, gave promise of being

some form of yellow. The color, as the name implies, is

orange-red dorsally, the belly being a lighter shade. Up

to the age of three weeks the young mice are dusky yellow-

red, the red apparently being obscured by a darker pig-

ment. As they grow older they become progressively of a

brighter and more intense reddish hue.

Genetically these mice behave much like black-and-tans.

None has been found which has bred true, and the relation |

of reds to non-red recessives is in the same approximate

ratio of 2:1. The recessives in this case are ‘‘chocolate,’’

in color a deep, rich brown, showing an intensity com-

parable to that of the blacks derived from the black-and-

tans. Thirty-one matings of red with red have produced

a total of 136 young, of which 77 have been red, and 59

brown, a ratio of 1.30:1. The average size of these litters

was 4.40. Eleven matings of red with brown produced 34

red and 31 brown young (equality expected), the average

size of litters here being 5.90.

So far we have dealt only with the pure stocks, each of

which is fairly uniform, although small fluctuations in

density of pigmentation do occur. When, however, these

two sorts are crossed with each other, yellow mice of vari-

ous shades are obtained, which form two graded series,

roughly parallel, one bearing black pigment and produc-

ing black recessives; the other bearing brown pigment

and producing brown recessives. Classification in these

two series is complicated by the fact that juvenile colors

are not uniformly retained, but in some cases increase and

in other cases decrease in intensity when the fur is

moulted. All animals have therefore been assigned a

numerical color-grade at the age of three weeks, this age

having been determined as the time when the relation of

yellow to black or brown pigment is most definitely visi-

ble; and although many animals have been re-graded at

intervals throughout life, each has been designated by his

original grade.

The cross of black-and-tan with red produced in F, two

No. 599] THE GENETIC BEHAVIOR OF MICE 667

classes of young. (1) One of these may be described as a

black-and-tan in which the black pigmentation is lessened

in amount and intensity, this decrease being attended by

an increased development of yellow pigmentation. This

class closely resembles the variety known as sable. (2)

The other class of young consisted of blacks, which also

were less intensely pigmented than the recessives pro-

duced by pure-bred black-and-tans mated inter se. It

was found convenient in classifying the young of later

generations to recognize six arbitrary grades of black-

ness of which yellow (showing no black pigment) forms

grade 1, and black-and-tan grade 6. On this scale the

mean of the F, ‘‘sable’’ young was close to 3.5, the inter-

mediate point between yellow and black-and-tan. The

distribution can be plotted by translating the descriptive

terms in Mr. Whittier’s notes into terms of numerical

grades, as follows:

These descriptive notes were made before the grading

scale had been adopted, and it is quite probable that no

real discontinuity in the variation occurred as would be

suggested by entire absence of animals of grade 4. No

such discontinuity is found in the work done since the

grading scale was adopted.

The F, black young were mated inter se and back-

crossed with browns to test their gametic composition.

When mated inter se they gave 28 black and 11 brown

young, nearly a 3:1 ratio (29:10). Back-crossed with

browns they gave 37 blacks and 33 brown young, nearly a

1:1 ratio (35:35). F, blacks apparently, then, were sim-

ple heterozygotes, not differing from ordinary heterozy-

gotes produced by crossing homozygous black with homo-

zygous brown. :

Thirteen of the F, sables were tested by mating with

browns. One hundred and thirty-three young resulted, of

which 70 were yellows of various shades and 63 non-yel-

lows. Of this latter group 32 were black and 21 were

668 THE AMERICAN NATURALIST ` [Vou. L

brown, equality being expected. The yellows also may be

divided into two groups, in one of which the eyes and fur

contain black pigment, while in the other the correspond-

ing parts contain brown pigment. In both of these yellow

groups the amount of black or brown pigment varied.

Again translating Mr. Whittier’s descriptions into terms

of the numerical scale, we have the following distribution:

Grae s+ L2 3 4 oR ee, Total

(1) The black series—Frequency... 0, 4, 8, 3, 8, 2( ?), 25

(2) The brown series—Frequency.. 13, 1, 4, 16, 1, 2, 37

It was frequently found to be impossible to determine

by inspection alone whether ‘a particular yellow animal

belonged to the black or the brown series, because yellow

fur containing a small amount of black pigment closely

resembles that which contains a considerable amount of

brown pigment. Consequently these back-cross young

(produced by an F, sable mated with brown) had to be

tested themselves, either by inter se matings or by cross-

ing with browns. The classification of the back-cross

young in the above tables is based partially on breeding

tests and in the cases where these were lacking classifica-

tion is based on inspection at the age of three weeks. It

is uncertain whether any individuals were obtained from

the F, sable X brown cross which showed the full inten-

sity of pure-bred black-and-tans (grade 6), although two

animals are recorded in the notes as black-and-tan with-

out qualifying terms.

As a result of back-crossing with browns the F, sables

(out of black-and-tan X red) and testing the young pro-

duced by crossing them with browns, two graded series of

yellow mice may be recognized as follows.

Rlack Series Brown Series

Producing as

Recessives

Grade | Designation | eb saa as

| Recessives

Grade | Designation

scenes |

6 Black-and-tan

Black 6 Brown-and-tan Brown

3-5 | Black-sable lack 3-5 | Brown-sable Brown

2 ooty yellow Black 2 e rown

1 llo lack 1 Yellow Brown

No. 599] THE GENETIC BEHAVIOR OF MICE 669

The brown-and-tan and brown-sable varieties are new.

They resemble black-and-tan and black-sables, respect-

ively, in which all black pigment in the fur has been re-

placed by brown pigment. The parallelism between the

two series is strongest at top and bottom; red has no exact

counterpart in the black series, since its yellow is more

intense than that of sable. All members of the two series

when crossed inter se fluctuate about their parental mean

grade. The greatest fluctuations are noted among the

offspring of sables; the least among black-and-tans and

reds. We suspect also that a like gradation occurs in the

amount and intensity of black and brown pigments in the

black and the brown recessives of these series, though on

account of the self color of these varieties this point is

difficult of verification, except by breeding tests. From

some tests which have been made and others which are

under way, the evidence seems to show that blacks from

sables and yellows have less intense young when crossed

with agoutis, than do the blacks out of pure black-and-tan.

Tables and curves for this cross will be given at a later.

time.

It is significant now that sables and black-and-tans may

be synthesized by a cross of blacks out of the black-and-

tan race with reds, showing that the black recessives carry

the same differentiating element as do the black-and-tans.

Such a cross produced 45 young, 20 of which were black-

and-tan or sable, while 25 were black. The F, blacks were

heterozygous for brown, inter se matings producing 32

blacks and 13 browns.

When a black which was heterozygous for brown was

mated to a red, yellows falling in both the black and the

brown series were produced as follows:

Black Series

Grade e Sa L2 3 6 6 Bik sa

Frequency... s5- LES a, 2% 5

Brown Series

Grade <. eee 1 38.4 &. Br. tom

6 13

Frequency ... -crire 1, 2303 4

S70 THE AMERICAN NATURALIST [ Vou. L

The element added in this last cross is plainly the brown

gamete carried by the black. This brown gamete, how-

ever, has received something additional from the black-

and-tan race, so that when red unites with this changed

brown gamete the result is a darkening and intensification

of the brown pigments to produce a brown-and-tan or

brown-sable, a process quite parallel to that which pro-

duces black-and-tan and black-sable in the pure black X

red cross.

A few crosses were made between pure-bred black-and-

tan and brown, and although the numbers here are small,

the indication is that the result will be the same as in the

black X red cross. F, consisted of blacks and black-

sables; the sables when back-crossed to browns gave ap-

proximately equal numbers of blacks (26) and browns

(20) and also the two yellow series as follows:

Black sables (mean grade 3.5).. 11 Brown sables (mean grade 4)... 10

Yellows and sooties............ 3 Reds and yellowS.........--+++ 0

TOE RE DRS a aas- 14 CC RRS 2 pal gey Fe tr ee 20

This back-cross with the recessive brown gives a direct

index of the yellow gametes of the F, sables. That they

vary in darkness should be borne in mind during the dis-

cussion of the difference between black-and-tan and

yellow. |

The reds in suitable crosses showed the same tendency

to produce fluctuating blends. Mated with creams they

gave yellows of an intermediate shade (16) and recessive

non-yellows (10). These light reds were bred inter se

and tested by crossing with browns. The young (100 in

number) fluctuated in intensity about the shade of the

light red parent or parents. Full intensity was not recov-

ered except in back-crosses. Hence red is likewise a form

of yellow, differing from it by an added intensity which

blends in crosses.

The foregoing evidence has merely pointed to the yel-

low nature of black-and-tan and red; has classified them

and their derivatives among the yellows, and has hinted

No. 599] THE GENETIC BEHAVIOR OF MICE 671

at the possible difference between these forms and ordi-

nary yellows. It is time now to inquire as to the real

genetic nature of these mice, and to attempt a preliminary

explanation of their differences from yellow. By far the

largest number of mice have been bred and are being bred

toward this end.

Let us consider first the black-and-tan variation. By

its behavior it evidently forms two sorts of gametes, black-

and-tan (yellow) and black. Each of these has an added

something which makes the zygote into whose composition

it enters darker than ordinary yellow or black. We may

call this something ‘‘darkener’’—be it singular or plural

—and indicate the gametes by YD and BD. Red, similarly,

forms gametes red (yellow) and brown; and these also

show an addition which we may call ‘‘intensifier.’? The

gametes of red are then YI and bI. The sables produced

by red X black-and-tan can only be referable to a union of

YD and bI, or YI and BD since YDYI is non-viable, and

since YDbI and YIBD unions have been demonstrated in

the brown X black-and-tan and red X black crosses, re-

spectively, and have produced in both of these latter cases

similar sables. The presence of the darkener and the in-

tensifier in the same zygote weakens both and demon-

strates their physiological and genetic independence. —

The next point to be noted is that both darkener and in-

tensifier are variable. All gametes formed by zygotes

containing D or I are not equivalent in their D or I con-

tent. It is possible to demonstrate this for the darkener;

the variable intensification from crosses with red cannot

yet be as satisfactorily shown on account of the difficulties

of grading. For light on the action of the darkener we

may turn to the agouti crosses.

The ordinary wild house-mouse when bred pure, shows

the agouti pattern and gray color with great uniformity.

It possesses the black and yellow pigments of the black-

and-tan mouse as well as brown pigment, but contains no

factor to dilute or darken these pigments. - These facts — |

make it an ideal race with which to test for a suspected -

672 THE AMERICAN NATURALIST [ Vou. L

darkener which acts on the black pigment of a yellow

mouse.

Yellow, Cuénot showed, is an allelomorph of agouti and

non-agouti. Black-and- bi in turn, is not an alternative

form of agouti like the light-bellied gray mouse, but a

yellow, and hence should be allelomorphic to agouti.

And such it is as far as its yellow component is con-

cerned, but not as regards its darkener. The F, young

from a cross of black-and-tan by wild agouti vary in

darkness about a mode midway between black-and-tan

and agouti. Black-and-tan we may regard as full dark-

ness and assign to it an arbitrary grade of 6. Wild

agouti we may regard as entire absence of this darkness

and assign to it the grade 1. F, from a cross of these two

has consisted of two sorts of young. (1) The first sort

may be considered as the result of a union of the YD

gamete from black-and-tan with the agouti gamete. These

mice have been called sable agoutis; since they have the

general pattern of sables. The bellies are yellow; the

darkness of the dorsal hairs is variable through the same

range as that of the sables, while all hairs on flanks, head

and parts of the back are agouti ticked. These yellow F,

young graded on the sable scale show the following distri-

bution:

ceed a tec Sas be Cuelebe tas 2, 0, 4, o; Total

(2) The second sort of F, young may,be called non-

yellow and referred to a union of the BD gamete from

black-and-tan with the agouti gamete. These are simply

much-darkened agoutis. The belly is gray like the wild

agouti and the flanks are agouti ticked. The head and

middle of the back are covered by hairs which are black

for most of their length, a very narrow yellow band being

present near the tip, or in some eases lacking entirely in

an area of hairs down the center of the back. This type

is known as dark agouti and has also been graded accord-

ing to darkness on a scale parallel but not exactly equiva-

No. 599] THE GENETIC BEHAVIOR OF MICE 673

lent with the sable scale. Grade 1 was taken as ordinary

wild agouti; grade 6 was taken as a gray-bellied black in

which the agouti pattern had been lost and in which the

darkness was equivalent roughly to that of the black-and-

tan. In grade 2 the extent of black in each hair is in-

creased, and the wide yellow band diminished; in grade 3

the yellow is left only in a narrow band; in grade 4 the

yellow ticking is lost from hairs in a streak down the

center of the back and in grade 5 the area of all black hairs

is extended to cover the whole back, ticking being limited

to the sides of the body. On the basis of such a scale the

F, dark agoutis were distributed as follows:

Grade r. coe ei ieee eee 2, 3, 4, Total

Frequency -ta cece ho cee to raea 8, 23, 5,

These F, dark agoutis bred inter se have produced 155

young, of which 109 have been dark agouti and 46 black,

indicating that the F, dark agoutis were heterozygous for

black. The distribution of 58 of these F, dark agoutis 1s

as follows:

rie tes 1 BiB ae Be Tome Mean Grade

Frequency oui itap pies Wee a ae ae AE, 2.2

By using as parents the darkest of these agoutis, re-

gardless of generation, dark agoutis were obtained, which

when three weeks old approximated the grade 6. They

resemble all-black mice with gray bellies except for occa-

sional ticked hairs on their flanks.

It will be remembered that all darkness in these dark

agoutis was acquired originally from the BD gamete of a

black-and-tan mouse, since the range of darkness in the

wild agouti used has never been above grade 1. Careful

grading of the young from matings among dark agoutis

should then furnish information as to the variation 1n the

“‘darkener.’? If the ‘‘darkener”’ is a multiple thing such

matings should afford it opportunity to segregate or Men-

delize. A tabulation of matings among all classes of dark

674 THE AMERICAN NATURALIST [Vou. L

agoutis of the young born since the introduction of the

grading scale follows:

DISTRIBUTION BY GRADE OF YOUNG PRODUCED By DARK AGOUTI PARENTS

OF VARIOUS GRADES

Grade Distribution of Young

Parents | |

ETETA Be 8 T ee! Grade | ToM

LT 32 | 82 | 1.00 1

2K22 11 | 21 i 8S BB 1

4x 2 WOFI (bt 2 | 34 | 2.00 1

4X4 6| 7 | 13 | 3.46 | 0

5X4 LIE Bios 4 4 46 | 4.41 | 14

5X5 IR +15 | 11 9 4 | 61 | 4.84 8

6x4 tf 2136148 4 9 | 46 | 4.28 2

6x5 tti ii i | 7.) SI rae 12

6 x 6 OTF ee fat .29 11

AKP (8 ><3)*)) 17 | 80-1289 2 | B8 2,20 | 282

6X1? MV ee ee | 1 | 38 | 2.20 0

It can be seen from a glance at this table that the varia-

tion in amount of darkness is a continuous one, from a

gray-bellied agouti dorsally all-black, through every pos-

sible gradation to a wild-type segregate. The continuous

nature of the variation was noted throughout the grading

of the dark agoutis in the ever-present tendency to create

more classes for the young by adding half and even quar-

ter grades, a temptation which was yielded to only in the

grade 5.5, this grade being given to dark agoutis which

showed many agouti hairs on shoulders and legs. There

is nowhere any evidence of a simple unit-difference be-

tween wild agouti and the darkener derived from black-

and-tan. The only segregation is that seen in grade 1 ani-

mals which bred true and gave no evidence of possessing

the darkener.

The above statements are not intended to be final. To

date the evidence indicates the presence of a continuously

variable and non-Mendelian character which gives the ap-

pearance of a blend in cross-bred young, and which in the

pure black-and-tan race has been added to yellows carry-

1 From the cross black-and-tan X wild. |

2 Highest grade dark agouti X wild.

No. 599] THE GENETIC BEHAVIOR OF MICE 675

ing black and has darkened and increased their black pig-

ment to the greatest possible extent. The evidence has

failed to exclude wholly the possibility of interpreting the

darkness of black-and-tan by multiple factors, but the con-

tinous nature of the variation in hybrid young would call

for the postulation of such a large number of modifiers

that this view could be neither proved nor disproved. The

nature of the darkener is still to be determined, but its

action on both agouti and non-agouti young seems to be to.

increase the total amount of black or brown pigment pro-

uced.

The experiments with the red race do not admit of as.

definite conclusions as were reached concerning black-and-

tan, because of the difficulties of grading for intensity. It

is safe to say, however, that red is a race in which yellow

has been greatly intensified by a process similar to, though

distinct from, that which has produced the darkness of

the black-and-tan.

STATISTICAL WEIGHTING FOR AGE OF

ADVANCED REGISTRY COWS

C. W. HOLDAWAY

VIRGINIA AGRICULTURAL EXPERIMENT STATION, BLACKSBURG, VA.

Any study of milk production that is made from a

statistical standpoint must necessarily be complicated, for

the reason that advancing age in a cow up to the time she

is mature enables her to produce more milk and butter

fat. A further difficulty lies in the fact that after ma-

turity the effect of age on production has not been de-

termined with any degree of certainty. Whether or not

the increase in capacity is directly in proportion to the

advance in age; at what age is the maximum of produc-

tion reached; what relation is there between age and per

cent. of fat in milk, and at what age is a cow past the

power of full productiveness, are all questions that need

investigation in a broad way.

Necessarily, the various breed associations must have

made some comprehensive investigations to enable them

to fix standards for milk and fat production, and since the

only extensive authenticated records that we have are

records of these associations, this study was made for

the purpose of determining if their records were con-

sistent with their standards, and if these standards could

be used as a basis for weighting cows of different ages.

METHOD or CoLLEcTING DATA

Seven-day records only were used, these being secured

from the Holstein-Friesian Blue Book, Vol. 24. For the

purpose of future investigation all the animals in two

direct lines of descent were tabulated, one from a female.

the other from a male, both animals being noted ones in

the breed. The names, herd book numbers, ages at time

676

No. 599] ADVANCED REGISTRY COWS 677

of record, pounds of milk, pounds of fat, and per cent. of

fat were all tabulated. Each animal was given an arbi-

trary number which denoted its position in the genera-

tion, and the position of all its direct ancestors in their

respective generations back to the primary ancestor of

the population. All advanced registry males were tabu-

lated also and numbered.

RECORDS OBTAINED

From the female, Aaggie Grace, No. 2618, H.H.B., only

456 advanced registry records were obtained in 10 genera-

tions. In order to secure these records about twice as

many animals were tabulated, the others consisting of

the A.R.O. sires and their daughters that had not them.

selves made A.R.O. records but had two or more A.R.O

daughters.

The male, Paul De Kol, No. 14634, H.F.H.B., in 9 gen-

erations produced 9,639 female progeny with A.R.O.

records. About twice this number of animals were tabu-

lated to secure these records.

TABULATION OF DATA

Necessarily, before this large accumulation of data

could be studied systematically, it was necessary to tabu-

late it in concise form, and for this purpose correlation

tables were made for each population, each table involv-

ing a pair of variables. Thus age was compared to

pounds of milk, age to pounds of fat, and age to percent-

age of fat; three tables to each population. From these

tables the means of the characters in classes, class aver-

age deviations, population means, average and standard

deviations and correlation coefficients were worked out.

Then from these data, curves were drawn to illustrate

its trend graphically.

RESULTS

The correlation tables I and II, compiled for the pur-

pose of studying the frequencies and distributions of the

population originating in the male, Paul De Kol, are not

678 THE AMERICAN NATURALIST [ Vou. L

shown here. The one and one and one half year class

and the classes over ten years of age were small. For

this reason unbalanced and irregular results would be

expected for these classes, and by referring to the curves

it will be seen that the premise was justified. The two

and three year classes were represented by 1,690 and

1,346 individuals, respectively.

Table III gives the average deviations, mean pounds of

milk, standard deviations, correlation coefficients and re-

gression coefficients of the population with respect to age

and pounds of milk and pounds of fat. Although the

mean age is four years, the three and one half year class

actually reached the mean pounds of milk of the popula-

tion, as can be seen from Table IV. Correlation probably

amounting to causation is shown in the tables up to six

years of age, and after that age is reached the correla-

tion is practically zero.

TABLE III

Correlation of Pounds of Correlation of Pounds

Milk to Age ot Fat to Age

Average deviation <.20.4:5...00% 69.8 .

Standard deviation .............. 92.4 +0.449 3.65 + 0.018

Moan ponda serso cueir viv isis 395.5 =+ 0.635 14.00 + 0.025

ene He eR ee ere ee 4.0 + 0.013 4.0 + 0.009

Correlation coefficient ........:... 0.604 = 0.004 0.57 = 0.005

Regression weight to age ........ 29.84 + 0.0006 1.11 + 0.00003

Regression age to weight ......... 0.012

; 0.29

Coefficient of variability C ....... 23.4 + 0.001 26.0 + 0.001

Table IV. This table gives the means, average devia-

tions, and plus deviations of the different age classes for

both milk and fat production. From these tables the

curves for milk and fat production were plotted. They

formed also the basis for calculating the curve which is

used as a comparison with the Holstein-Friesian curve of

fat and milk requirement. These tables also afford an

interesting study from the standpoint of capacity of

cows for milk production at different ages.

Considering first the curves for milk production (Fig.

1) it will be noted that curve 1, which represents the

No. 599] ADVANCED REGISTRY COWS 679

pounds of milk required by the Holstein-Friesian Asso-

ciation, must be calculated from the pounds of fat re-

quired. This was done by taking the average per cent.

of the whole population and calculating the number of

TABLE IV

Milk Production Fat Production

Age, EE Si RON cao as VS BO iS ea

14 268 34 302 170 8-4 he | 10.2 4.76

2 308 44 352 220 107 4 is 12.44 7.00

24 326 43 369 237 113° | 176 13.06 7.62

g 372 56 428 296 130 1 > 2.14 14 9.70

34 396 5 451 319 12. 2.37 15.17 9.7.

4 428 t 482 350 15.2 .48 .68 | 12,24

43 446 57 503 371 15.5 oO 80 | 12.36

l; 458 517 385 af 36 76 | 138,30

5} 461 517 385 16.4 oO 4 13.24

(i 474 ( 536 404 0 .55 19.55 | 14.09

64 474 € 534 402 16.8 .43 "29 | 18.71

476 € 536 404 16.7 3 .00

$ 466 64 530 398 16.8 .69 .49 4.03

475 € 540 408 17.0 54 54 4,08

5 476 58 534 402 16.6 46 06 3.60

476 i 527 395 17.1 43 .53 4,07

$ 475 60 535 403 | 17.0 48 48 | 14.03

1 449 76 525 393 16.6 7 .30 3.85

104 477 68 545 412 | 16.7 1 80 | 14.33

1 461 35 496 3 16.0 0 .00 2,53

ue 470 36 506 374 15.7 43 | 13 1.66

pounds of milk, having the average per cent. that would

be necessary to make the required number of pounds of

fat. The reason for using the average per cent. of fat

of the whole population as a basis for calculating the

Holstein-Friesian Association requirement curve was

that since the correlation coefficient between age and per

cent. of fat was so small in a table shown subsequently

for another population, and since the popular concept is

that per cent. of fat is not influenced by age, we felt justi-

fied in using it. Attention is called to Table V, which

does not bear out this assumption entirely. For milk and

fat requirement, however, there is a strong correlation to

age, so the classes were considered separately, each class

having its own mean and deviation. Curves 2, 3, and 4

were based on these class means and deviations. Curve

680 THE AMERICAN NATURALIST [Vou. L

No. 2 is the mean of the population. Curve No. 3 is the

plus deviation from the mean. Curve No. 4 is a curve

Curves of Milk Progvetion -fig.I

-+—++— b

—— =

Pom RAF TIA SSE BE U A ARERR 50

TH Ro P FO f

a A O

5 8 Se ee ee ee a oa ma n UNS SEn aun a

nmn e snn S [8 Et S Gee Gee ee ee ee eee ee 500

ERE

ua PN

Soe? 2a coe N

Hf zE $

f 5'4 APG A DA SA PPP ts

7 eae

= H so X

£.

Ai 200

leo

/ 2 3 4 Lee i Z o 40 y

-AGE 17 Years -

which was plotted to show what the requirements ought

to be if the means, deviations and varying capacity of the

different classes are taken into account. In plotting this

curve it was necessary to consider the basis upon whic

the minimum requirements of this population ought to

be placed.

The minus deviation point can not show what ought to

be required of the class as a minimum, for such point

would weight individuals inversely in proportion to their

capacity, A greater deviation from the mean of the class

No. 599] ` ADVANCED REGISTRY COWS 681

indicates here greater capacity for production of that

class, and as the capacity for production of the class in-

creases, so should the requirements increase. Therefore,

the curve of minimum requirement should be represented

as following the curve of plus deviation in character and

should be in a minus direction from the mean.

In order to conform to these conditions some basis must

be established for calculating the minus points of the

curve, or, in other words, the minimum requirements for

each class. The average deviation of the whole popula-

tion seems to be the logical basis upon which the minimum

requirement should be based, for by its use the whole

curve may be lowered an amount corresponding to the

average deviation of the whole population below the mean

of the population. The average deviation from the mean

of the whole population is 69.8 pounds of milk. Tf all

classes are to be given the benefit of the average devia-

tion the calculation should start from the point at which

the means are at the maximum, which is about the six-

year class. Hence the six-year class is allowed as the

minimum requirement, the 69.8 pounds below the mean of

the class and the requirements of the other classes are

worked out from this point to conform, as said before,

to the maximum deviation curve.

An inspection of these curves brings out the following

points:

That the official requirements weight animals of an age

from 18 to 21 months too heavily. The curve indicates

that they are entitled to a reduction as great as for any

other age. For the purpose of discouraging such early

breeding, however, the requirements for this particular

class should be prohibitive and they are.

That the production increases up to at least six years

of age instead of five, which the Holstein-Friesian Asso-

ciation requirements set as the maximum age production.

That for this reason the 5- to 6-year-old: animals and

possibly the 7- to 8-year classes have an adyantage over

all other classes. ee aoe

That a comparatively small number of animals made-

682 THE AMERICAN NATURALIST — [Vou. L

the requirement after 9 years of age, hence by selection,

only the best animals were retained, thus drawing the

curve down almost to a straight line. The tendency of

the curve, however, is to recede, showing that the animals

of these ages should not be weighted as heavily as younger

animals. A study of a number of repr tatives of the

whole breed would be necessary to determine this point.

One of the most striking points shown by these data

and one which substantiates the opinion of practical

breeders of Holsteins, also brought out in the practical

investigations of Eccles,! is the difference in production

and capacity between 2- and 24-year-old and 3-year-old

cows. The difference in the means of the production

between 2 and 24 years was 18 pounds only, while be-

tween 24 and 3 years it was 46 pounds, or a total of 64

pounds between the 2- and 3-year classes. Between the

3- and 4-year classes the difference is almost as great,

being 56 pounds, but the deviation of the latter class is

not quite as great as the former. This seems to indicate

that the 3-year animal is still at a disadvantage by reason

of its immaturity in growth and body development. That

the average deviation of 24-year class was 43 pounds

while the 3-year class deviated 56 pounds is significant

also and leads to the conclusion that at 24 years of age

the Holstein is still growing, and this, combined with the

great strain of milk production, limits the capacity of

the class.

It may be said by some that few 3- and 4-year-old

animals are tested for advanced registry in comparison

to two year olds and aged animals, and in consequence

of this, only the best of the class make the requirements.

This is not borne out by the data, the number in the 3-

year-old class being second largest of all animals.

Curves oF Far PRODUCTION

A study of the curves based upon the actual fat pro-

duction of this population (Fig. 2) brings out a number

1 Bul. No. 135, Missouri Agricultural Exp. Station.

No. 599] ADVANCED REGISTRY COWS 683

of points, many of them corroborating those brought out

in the discussion of the milk-production curves.

Owing to the variation of the weight classes in per

cent. fat, the curves of milk production and fat produc-

tion agree very well when compared with the Holstein-

Friesian Association requirement curve.

The requirement curve in fat production (No. 4)

crosses the Holstein-Friesian Association curve at a

Curves of fat Production -Fig Z

7 a ae +

> APTN A

a aes euag p |

| A win! ee

AHHH ++ Nal S

Ly ey eels «

A 2

=m" ee en

eN NY X

yN NN à

Z) E AR

/ 4 9

‘a Soi aes GSE Ge Sl ee Ge ete 3

ob br e 9

: abd |

S pad pp ee tatan ie

mae Se Gees Gear Chi Sd LES ee oe Ge ae

Gis Od Be A es OO oe a A a aa Ss

4 LELEL i

LELLA

HHH

“= ee eS . e a

ge jp Yeors—

greater age than that worked out for milk production.

This would indicate that the classes up to 34 years pro-

duced milk containing a lower per cent. fat than the mean

of the whole population. This is correct, as can be found

684 THE AMERICAN NATURALIST [ Vou. L

from the means of the classes. (See average per cents.

of class means, Table V.) A similar condition obtains

with the age classes after ten years. It would appear

from this that mature cows give milk slightly richer than

immature cows, or than old cows past 10 years of age.

A rather peculiar condition with reference to the fat

production curve is shown in the mean results of the

half-year ages up to the 64-year class. Hach half-year

class advances but slightly, if at all, from its preceding

year class, then there is a sudden drop to the next full-

year class. The milk production curves indicate the same

condition, though to a lesser extent, and as previously

noted, the frequencies in these half-year classes are not

more than 60 per cent. of the full-year classes. No good

explanation is offered for this. It might be inferred that

a cow freshening at 24 years is not much better able to

withstand the strain of milk production than a 2-year-old,

and that this condition continues. However, in many re-

spects this theory does not appear sound.

Attention is called again to the points of curve 4 for

fat production given in Table IV. This curve is plotted

for the purpose of showing what the requirements ought

to be according to the performance of cows that have

made records. The animals involved in this curve repre-

sent 45 per cent. of all the A.R.O. records that had been

made up to the time of publication of Vol. 24, hence the

TABLE V

AVERAGE PER CENTS. FAT OF THE CLASSES

Age, Years | Per Cent. Age, Years | Per Cent.

1} | 3.28 7 | 3.51

r T E SIGS 3.27 73 | 3.61

ga 3.47 1 eas aa | 3.58

3 3.4 8} 3.49

3} | 3.24 9 | 59

eee 3.55 } | 3.58

4} | 3.48 10 | 3.69

5 | 3.58 10} | 3.50

e ie agents ea 3.56 No. | 3.47.

6 3.58 11} 3.34

ea 3.54 |

No. 599] ADVANCED REGISTRY COWS 685

numbers are ample. First, the means of the classes of

this population were plotted. Then their ability to devi-

ate in a plus direction, or, in other words, to produce more

fat as individual classes was taken into account. The

class that had the maximum production and deviation

ability was allowed, as a basis for its minimum require-

ment, the full average deviation of the population in a

minus direction from the mean, and finally the other

classes that could not produce as much and had not the

ability to deviate as much as this maxinium class, were

allowed the full minus deviation of the population plus

the difference in deviation ability between their particular

class and the maximum class which forms the apex of the

curve.

If these fundamental allowances are fair, impartial and

accurate, the curve is accurate, and the only question that

remains is whether or not it should alter the requirements

of the Holstein-Friesian Association. If curve 4 touches

the Holstein-Friesian Association curve at any point and

does not coincide with it throughout, then the latter should

be changed. It does touch it at both beginning and end,

showing that all classes after the 24 years and up to 114

years have an advantage over the others. This advan-

tage is greatest for the classes between 53 and 11 years

of age. a

The next consideration in connection with curve 4 is its

application, and, when dealing with this, two things

should be kept in mind; first, the practical, and secondly,

the more concise and mathematical application. The

practical application finds its expression in the endeavor

of the Holstein-Friesian Association to make a uniform

advance per day in the fat requirement for the seven-day

test up to the age at which it was considered the maximum

production was reached. Table VI compares the increase

in the amount of fat required each year over that required

in the previous year from two up to six years, with the

increase in amount of fat that the year classes are able

to produce as calculated from curve 4.

686 THE AMERICAN NATURALIST [Vou. L

TABLE VI

H. F. A. Requirements | Curve 4 Requirements

AEO a Fat Increase, | Fat Increase, | Fat Increase, | Fat Increase,

Yearly Daily Yearly Daily

2 to 3 1.6 lbs. 0.00488 | 2.70 lbs. 0.00740

8 to 4 HE 0.00438 2,04! 0.00696

4tod 65°" | 0.00438 1.06 :* 0.00290

5 to 6 i AEE 0.79 1 0.00216

The table shows plainly that the daily increased re-

quirement from 2 to 3 years should be 0.0074 instead of

0.00438, or 1.7 times as much. From 3 to 4 years should

be 0.00696 instead of 0.00438, or 14 times as much. From

4 to 5, 0.0029 instead of 0.00438, of nearly 4, and from 5 to

6 years, 0.00216 instead of no increase.

Poruraton No. 2

The second population tabulated is that which began

with Aaggie Grace No. 2618, H.H.B., as the primary an-

cestress, and consists of only 456 animals. Correlation

tables 7 and 8 are omitted, but 9 and 10 are given, and

show all the data necessary for comparison with the

previous population. Of course, it must be borne in mind

that the comparison can not be too exacting, for this pop- —

ulation is altogether too few in numbers to secure smooth

results especially when comparing classes. In fact, the

class means and deviations, Table IX, included only the

classes up to 9 years because of the low frequencies after

that age. If Tables III and IV are compared with 9 and

10, a remarkable agreement is noticed throughout, es-

pecially in the essential points which have been discussed.

The correlation table for age to per cent. of fat is not

shown, but the coefficients of this table may be seen in

Table X. The correlation coefficient is so small that it

may seem negligible, but Table V shows that even with a

low correlation, important points might be brought out if

the data are sufficient.

No endeavor will be made in this paper to enlarge on

the exact mathematical application of these data. This

No. 599] ADVANCED REGISTRY COWS ` 687

will be taken up later in connection with a further study

of the two populations.

TABLE IX

CLASS MEANS AND DEVIATIONS OF POPULATION 2

| Age to Pounds Milk | Age to Pounds Fat

Age, Years |

Mean Pounds Milk | Average Deviation | Mean Pounds Fat | Average Deviation `

| 286 47.3 10.35 .65

2$ 310 48.6 il, .63

3 392 52.5 13.16 87

È 403 53.5 14.22 97

4 413 60.0 14.30 >, 20

4% 470 57.8 L28

5 469 56.6 16.42 30

54 484 41.4 15

6 504 66.4 17.23 26

63 406 93.9 17:75 3.37

7 480 59.2 16.45 .56

74 471 47.0 17.22 2.14

8 468 60.7 16.27 2.48

83 500 40.0 18.20 3.92

447 47.6 15.00 2.29

TABLE X

PoPULATION COEFFICIENTS; POPULATION 2 i

Age a Age yrs pica Mg ed

DD scsi ss 4 +2.8 | 13.89 +0.113 | 3.455+-0.014

Standard Deviations 90.39 +2. ‘59310.08 | 0.436-+-0.009

Correlation Coefficients N .........-... 0.592+0.02 | 0.581+0.02 | 0.08 0.031

Coefficient of Variability C........... 0.223-+-0.005| 0,258-+0.006 | 0.1260.

Regression Weights to Age ........... 26.51 +0.042) 1.057+0.002

Regression Age to. Weight | 0.02 0.546

ACKNOWLEDGMENT

_ The writer wishes to thank Prof. H. L. Price, of the

Virginia Polytechnic Institute, for the valuable help and

Suggestions given by him in this work.

SHORTER ARTICLES AND DISCUSSION

G@NOTHERA NEO-LAMARCKIANA, HYBRID OF

0O. FRANCISCANA BARTLETT X 0. BIENNIS

LINNÆUS

Œnothera neo-Lamarckiana is a name which I propose for a

synthetic hybrid that so closely resembles O. Lamarckiana De

Vries that I do not believe systematic botanists could separate it

from. the latter by characters which would enter into a specific

description. This does not mean that the hybrid is the exact

counterpart of any particular line of Lamarckiana carried for-

ward by the geneticists who are working with this form for it

must be remembered that there are numerous biotypes of this

species differing from one another in matters of greater or less

detail, and that workers with œnotheras know that Œnothera

Lamarckiana of systematic literature is a collective or poly-

morphic species, various forms of which can be isolated as bio-

types in the experimental garden. In my studies of the La-

marckiana-like hybrids I am selecting towards the type known to

us through the work of De Vries and through the seeds dis-

tributed by him.

. The parents of my hybrids are O. biennis from the sand dunes

of Holland and O. franciscana from California. I have given in

a recent paper’ the contrasting characters of these species to-

gether with descriptions of hybrids in the first and second gen-

erations. These parents were chosen after several years of

search among the cenotheras for wild species that might be

erossed with the hope of obtaining Lamarckiana-like types. In

this connection my attention was first called to franciscana by

Prof. Bartlett. In biennis and franciscana together are suggested

all of the essential taxonomic characters of Lamarckiana and

there seemed good reason to expect that among hybrids of the

second and later generations would be found forms with com-

binations of characters approaching very closely to the peculiari-

ties of Lamarckiana. In this respect my cultures now in the

fourth generation have yielded results quite as satisfactory as I

have hoped.

1 Davis, B. M., ‘‘ Hybrids of Œnothera biennis and Œnothera degrees

in the First and Second Generations,’’ Genetics, I, 197-251, 1916

688

No.599] SHORTER ARTICLES AND DISCUSSION 689

The results will, I think, show that Lamarckiana-like forms of

@Œnothera may be synthesized by simple crosses between wild

species provided the parent species are selected with care. I

believe that as the isolation of @nothera types proceeds a num-

ber of different crosses will be found to give similar results, but

this is the first successful combination that I have been able to

study experimentally. My earlier work? with Gnothera grandi-

flora Solander and certain American wild types was planned at

a time when grandiflora on historical grounds seemed to be a

more important type in relation to the problem of the origin of

nothera Lamarckiana than it does at present. That work was

not so successful as the later in producing Lamarckiane-like hy-

brids for the reason that the parent species did not have as

favorable characters for the end in view.

In spite of the recent paper of De Vries,’ to which I have re-

plied‘ in brief, my conviction is unshaken that Lamarek’s plant,

grown in the botanical gardens of Paris about 1796, was a form

of @nothera grandiflora Solander and can not be identified with

the Lamarckiana of De Vries’s cultures. On this view @nothera

Lamarckiana Seringe must pass into the synonymy of Gnothera

grandiflora Solander. Neither am I convinced that other speci-

mens in the collections of the Muséum d’Histoire Naturelle in

Paris, particularly sheets of André Michaux and Abbé Pourret,

may be referred to @nothera Lamarckiana. I believe that the

plant with which we are concerned in the experimental garden

had a later origin and must bear the name of De Vries as its

sponsor. At present our first certain date of the progenitors of

Cnothera Lamarckiana De Vries appears to be about 1860, when

the seed firm of Carter and Company in London introduced the

plant to the trade.

Both De Vries and Gates have accepted my suggestion that

Carter and Company obtained their material of Lamarckiana

from some English station and not from Texas, as they state.

We have no evidence that Lamarckiana ever grew in Texas and

to me there is no evidence that it was ever native to America.

2 Davis, B. M., “$S Hybrids of Gnothera biennis and O. grandiflora

7 ’ ome Hy a Sie, XLV, 193-233, 1911.

; ble

— De Vries, Hugo, ‘‘The Probable Origin

r.,’? Bot. Gaz., LVII, 345-361, 1914. s

+ Davis, B. M., ‘‘ Professor De Vries on the Probable Origin of Œnothera

Lamarckiana,’? Amer. Nart., XLIX, £9-64, 1915.

690 THE AMERICAN NATURALIST [ Von. L

On the other hand, various races of Lamarckiana are at present

growing wild in a number of English localities, the best known

stations being on sand hills of Lancashire near Liverpool. A

conspicuous @nothera flora was present in this region as early

as the beginning of the nineteenth century, as shown by an ac-

count in Smith’s ‘‘English Botany,’’ 1806. There seems to be

no reason why @nothera Lamarckiana might not have arisen in

such a locality as a hybrid of species introduced into England

possibly through Liverpool as a port of entry. Thus we are

dealing with dates of introduction or origin that are reasonably

close to present times; attempts to associate Lamarckiana with

very early introductions into Europe appear no longer to have

important support.

It is necessary to bear in mind this historical setting, since it

may seem to my readers very improbable that @nothera La-

marckiana should have arisen as a hybrid between franciscana,

a species of western America, and biennis of Holland, England

and other European countries. There is, however, nothing im-

probable in the possible meeting at Liverpool, with its world-

wide commerce, of species of Ginothera from far corners of the

earth. Furthermore, I should be the last to suggest that the

particular races or species which give my neo-Lamarckiana have

been the actual parents of the strains of Lamarckiana cultivated

by De Vries. To strike the identical parental lines of such an

assumed hybrid would in the case of the cenotheras be a most

extraordinary piece of luck. It is remarkable that my results

have proved so satisfactory; I have no doubt that other species

crosses may sometime be made which will give hybrids as close

or even closer to Lamarckiana.

The line of neo-Lamarckiana, which I now have in the F,

generation from the original cross, was derived from a single.

selfed plant in the F, (14.53c), which fell well within the range

of variation given by De Vries for @nothera Lamarckiana. A

description will later be published of this plant together with an

account of its progeny through successive generations when these

have been carried along somewhat further. The F, generation

gave very few neo-Lamarckiana types, but these were closer to

the large-flowered forms of De Vries’s cultures. This F, gen-

eration was grown from earth-sown seeds and incomplete germi-

nation may have been responsible for the small proportions of

neo-Lamarckiana, 7 in a total of 291 plants. The F, generation

No.599] SHORTER ARTICLES AND DISCUSSION 691

was grown. this summer from what seemed to be the most promis-

ing plant of the F, (15.53a). This culture was from seed ger-

minated in Petri dishes and was complete, since the residue of

ungerminated seeds were empty of contents. From 764 seed-

like structures 668 seedlings appeared, but there was at once a

large mortality among weaklings most of which were unable to

free their cotyledons from the seed coats. Only 558 seedlings

lived to be potted and a further mortality reduced the number

that was set out in the garden to 549. Of these plants 198 as

rosettes presented characters of Lamarckiana while 351 developed

rosettes for the most part with narrower leaves suggestive of

franciscana. All of the shoots from the 198 Lamarckiana-like

rosettes have shown Lamarckiana characters of foliage, inflo-

rescence, and flowers but about one fourth of the plants seem

likely to persist this summer as rosettes. The group of neo-

Lamarckiana in the F, generation is therefore large constituting

about 36 per cent. of the total number of plants in the culture.

In the group of neo-Lamarckiana there is some variation, but

the best plants are so close to the Lamarckiana of De Vries that

I can only distinguish them by small plus or minus expressions

of a few characters. Thus the central shoot is not so strongly

developed proportionally to the side branches. The leaves are

a little broader. Sepal tips do not spread so widely. Buds

may not be quite so stout. The pubescence is somewhat heavier

over certain portions of the plants. Time will tell whether even

these small differences can be eliminated by judicious selection

It is of course not enough for critical bearing on De Vries’s in-

terpretation of the behavior of Lamarckiana that a hybrid should

be synthesized taxonomically similar to it. Such a hybrid must

also show a behavior parallel to Lamarckiana in its essential fea-

tures. The two striking peculiarities in the breeding habits of

Lamarckiana are (1) its ability to produce two types (twin

hybrids) in the F, when mated to certain other species, and (2)

its peculiarity of throwing through successive generations the

same types of ‘‘mutants’”’ in small, fairly constant proportions.

Late in the season of 1915 reciprocal crosses were made between

neo-Lamarckiana (15.53a) and plants of biennis and bienms

(Chicago), forms which De Vries has used in his studies on twin

hybrids from Lamarckiana. The conditions were not favorable

for the technique of crossing and I am repeating the experi-

692 THE AMERICAN NATURALIST [Vou. L

ments this year. However, I obtained from the cross biennis X

neo-Lamarckiana two distinct classes of plants, (1) a narrow-

leaved, smaller-flowered type with heavy pubescence and red

papille (109 plants), and (2) broad-leaved forms, some larger-

flowered, with a much lighter pubescence and few or no red

papille (11 plants). Also, the cross neo-Lamarckiana X biennis

(Chicago) gave two clearly defined classes distinguished at a

glance by their size and foliage, (1) tall and narrow-leaved (64

plants), and (2) shorter and broad-leaved (11 plants). These

crosses appear to have given twin hybrids and it should be said

that the two groups were recognized and separated when the

plants were in the rosette stage and that they consistently: pre-

sented differences throughout all stages of their development.

I shall from time to time make further studies of this behavior

with different generations of neo-Lamarckiana. If biennis and

biennis (Chicago) are pure species (a matter not yet established)

this behavior would indicate that neo-Lamarckiana develops at

least two classes of fertile gametes for both pollen and ovules.

It thus seems probable that the behavior of neo-Lamarckiana

when crossed to other species of @nothera will parallel that of

De Vries’s Lamarckiana and thus support the view of several

critics of the mutation theory that Lamarckiana, because it gives

twin progeny in the F, of certain species crosses, must be itself

a hybrid, producing different classes of gametes.

With respect to the ability of neo-Lamarckiana to throw

‘‘mutants’’ a most interesting situation is presented by its be-

havior this summer in the fourth generation. We have noted

that a sowing of 764 seed-like structures gave 668 seedlings of

which 198 developed as rosettes or mature plants into neo-

Lamarckiana. Of the remaining 470 seedlings (668-198) only

351 lived to produce rosettes, a much larger group, however,

than that containing the parent type, neo-Lamarckiana. We

have then in the fourth generation neo-Lamarckiana, an impure

or hybrid species, reproducing itself from at least 26 per cent. of

its seeds. The exact percentage can not be told, for we do not

know whether any plants of neo-Lamarckiana were among the

119 seedlings that died. In throwing a large progeny of a type

very different from the parent F, plant, neo-Lamarckiana in the

F, exhibited a behavior with strong resemblance to what Bartlett

has described as ‘‘mass mutation.’’ The types included a num-

ber of dwarf forms, but most of the plants resembled franciscana,

No.599] SHORTER ARTICLES AND DISCUSSION 693

although generally stronger, larger-leaved, and with considerable

variation in flower size. It should be noted that no forms similar

to the parent biennis were present; this type of segregate seem-

ingly is either not produced or appears but rarely.

The conditions of sterility in neo-Lamarckiana are likely to bear

directly on the peculiarities of its behavior in comparison with

that of De Vries’s plant. My hybrids agree with Lamarckiana

in having pollen about one half sterile, but the F, parent plant of

this year’s cultures showed seeds 87 per cent. fertile while the

seed fertility of Lamarckiana is much lower, being reported by

De Vries in extensive experiments as from 34.546 per cent. and

for two lines of mine running in tests 26-30 and 32-36 per cent.,

respectively. The variation noted by De Vries is believed by

him to depend upon whether or not the plants are heavily

manured. The question at once arises may not the mass varia-

tion of neo-Lamarckiana in the F, be correlated with its very

much higher seed fertility? What would happen if neo-Lamarck-

iana should develop a greater degree of seed sterility or if some

lines should be segregated with seed sterility approaching that

of De Vries’s Lamarckiana? Would the plants eliminated come

from among the neo-Lamarckianas or would they come from the

assemblage of variants from this parent type? Should they

come from the variants, as seems to me probable since neo-

Lamarckiana is a sturdy plant, then a condition might be reached

where the variants would appear rarely or in small proportions

and this would parallel exactly the present behavior of De Vries’s

Lamarckiana in throwing its ‘‘mutants.’’ I shall watch intently

for indications in my cultures of increased seed sterility and

among my plants of neo-Lamarckiana select steadily towards the

higher degree exhibited by Lamarckiana. It is interesting that

the last stages in the experimental synthesis of a Lamarckiana-

like hybrid should be concerned chiefly with selection towards a

definite degree of seed sterility.

@nothera neo-Lamarckiana illustrates clearly my concept of

an impure species of Œ@nothera.5 It is a plant that breeds true

in a proportion of its offspring but is heterozygous since it de-

velops varied types of gametes as proved by the assemblage of

offspring which differ sharply from the parent plant, and further

indicated by its behavior in producing twin hybrids. The facts

of a high degree of pollen sterility (about 50 per cent. ) together |

5 Davis, B. M., ‘‘The Test of a Pure Species of @nothera,’’ Proc. Amer.

Phil. Kis LIV, 226-245, 1915.

694 THE AMERICAN NATURALIST [Vou L

with a certain amount of seed sterility indicate the probability

that other classes of gametes are eliminated or fail to func-

tion and that possibly certain types of zygotes may be formed

which are unable to live. Mnothera neo-Lamarckiana therefore

shows itself to be impure or heterozygous because it develops

different types of gametes even though the plant when selfed

reproduces itself in a fairly large proportion of its progeny.

This behavior seems to me quite the same in principle as that of

De Vries’s Lamarckiana, the only difference being that the total

number of individual variants thrown by Lamarckiana is much

smaller than those at present thrown by neo-Lamarckiana. How-

ever, as has been noted, the seed sterility of Lamarckiana is very

much higher than that of neo-Lamarckiana in the F, generation,

and it is my working hypothesis that this fact is at least partly

responsible for the smaller numbers of variants produced by the

former.

(nothera Lamarckiana of De Vries’s cultures seems to me

best interpreted as an impure species producing regularly be-

cause of its heterozygous or hybrid nature a number of classes

of gametes relatively few of which, because of the extensive

sterility, both gametic and zygotic, are able to form viable seeds

different from those that reproduce the species. M#nothera

Lamarckiana breeds true in its high degree because only the

gametic combinations that reproduce Lamarckiana survive the

mortality visited on most of the gametes and zygotes. By this

view Lamarckiana is very much the reverse of a representative

pure species which De Vries has assumed it to be and its ‘‘mutat-

ing habit’’ is the result of its hybrid origin and heterozygous

nature rather than a spontaneous expression of homozygous germ

plasm. The fact that the same types of ‘‘mutants’’ from La-

marckiana are produced by successive generations in fairly stable

proportions indicates that their differentiation lies in the mechan-

ism of segregation in heterozygous germ plasm rather than in a

sporting tendency (mutation) which would be expected to ex-

press itself in ever-varying ways and degrees,

It is worth noting how different is the conception, here ex4

pressed, of the constitution of a pure species from the view

formerly and probably now very generally held. Formerly a

Species was considered pure if it bred true. Now we believe that

a species may be impure and still breed very largely or even

wholly true if a degree of sterility is present sufficient to render

No.599] SHORTER ARTICLES AND DISCUSSION 695

abortive or infertile all types of gametes or zygotes that may be

produced except the ones which carry forward the heterozygous

line. The test of a pure species is then not that it should breed

true (that is a corollary), but that it should produce gametes

uniform except as they may differ with respect to the factors

for sex characters.

In laying such great stress on the phenomena of gametic and

zygotic sterility so very extensively present in the genus @nothera

it must not be supposed that we have as yet established the de-

grees to which sterility may be genetic in its character or to what

extent it may be of a physiological nature. Only the sterility

that has as its cause the failure of the reduction divisions to

produce fertile gametes or the failure of the gametes to conjugate

freely can properly be of a genetic nature. There is probably

also a type of sterility due to physiological causes, as perhaps

malnutrition, and this might affect gametes and zygotes which

under favorable conditions would be fertile. We are very far

from an understanding of the causes of sterility in @nothera,

to what extent cytological or to what degree physiological, and

it would at present be most unsafe to carry lines of speculation

very far in this field as regards the material under consideration.

Professor De Vries has expressed strongly a belief in the

futility of my attempts to synthesize a Lamarckiana-like hybrid,

taking the stand that unless the parent stock is known to be

stable mutability might be inherited from one or both of the

parent species, or that variants, the result of a cross from impure

stock, might be mistaken for mutations. As a matter of fact

Œnothera biennis is known to be unstable, producing a small

series of ‘‘mutants,’’ while 0. franciscana has not been tested for

its purity. Apparently Professor De Vries and I are working

from assumptions that are far apart. The inheritance of a

mutating habit such as that claimed for Lamarckiana would

mean to me the inheriting of a heterozygous germ plasm running

back to some hybrid origin. To me phenomena such as is ex-

hibited by Lamarckiana in throwing its ‘‘mutants’’ indicates in

itself the probability of heterozygous germ plasm. If this be-

havior is to be presented as evidence of mutation the purity of

Lamarckiana must be established beyond all reasonable doubt

and this in my opinion has not been shown. The tests of cross

breeding, when twin hybrids result, and the very high degrees

of gametie and zygotic sterility strongly indicate genetic im-

696 THE AMERICAN NATURALIST [ Von. L

purity. And back of this is an obscure history for the material

with no evidence it seems to me that Lamarckiana was ever

present as a native species of any flora. The chief value which

the study of my Lamarckiana-like hybrid may have for the prob-

lem of the origin and status of @nothera Lamarckiana is likely

to be a clearer understanding of how an obviously impure

species, neo-Lamarckiana, may arise, a species which seems likely

to present a breeding behavior parallel to that of Lamarckiana,

and most important of all the significance of sterility in the

working out of these results. It appears to me a matter of no

vital importance to the status of a hybrid whether its parents are

pure or impure. If markedly impure the problem of analysis

for future generations merely becomes the greater. Since no

species of Œnothera has as yet passed the tests for a pure species,

we are at present in all of the @nothera work talking of an ab-

straction when this concept is considered.

BraDLEY Moore Davis

UNIVERSITY OF PENNSYLVANIA,

August, 1916

STATISTICAL STUDIES OF THE NUMBER OF NIPPLES

IN THE MAMMALS

Ir is perhaps not unnatural that a subject of such fundamental

interest as that of the nourishment of the young in the mammals

should have attracted the attention of observers from the time of

the Greek philosophers. It is only within the last few years that

attempts have been made to solve various problems by the appli-

eation of the statistical method to series of quantitatively re-

corded data.

The materials may be divided for convenience of review.

TYPE, VARIATION AND CORRELATION In NuMBER oF MAMMÆ

The statement made by Parker and Bullard,’ on the basis of

their splendid series of data for swine, that the standard devia-

tion of the number of nipples is 0.6906 in the males and 0.7905

in the females at once arouses the suspicion of a biometrician.

The constants actually are:

1 Parker, G. H., and C. Bullard, ‘‘On the Size of Litters and the Number

of Nipples in Swine,’’ Proc. Amer. Acad. Arts and Sci., 49: 399-426, 1913.

No.599] SHORTER ARTICLES AND DISCUSSION 697

For Males For Females

E STUE E E +ew i ein Ba 12.4365 + .0182 11.9077 + .0159

Coefficient of Variation ............. 11.901 + .105 10.752 = .096

Thus instead of the females being ‘‘over 14 per cent. more

variable than the males’’ they are in absolute terms actually

1997 = .0175, or over 13 per cent., less variable. Relative vari-

ability as measured by the coefficient of variation is 1.149 + .142

per cent. lower in the female that it is in the male. This lower

variability of the female is also quite in evidence if the ma-

terials be split up into groups with regular and irregular ar-

rangement of the nipples. Thus:

For ‘‘Rrecuiar’’ CLASS

PO A CES o = 1.485 + .017, C. V. = 12.03 + .14

ORME, es ila ak o== 1,315 = 021, C.: ¥.== 11.16 + 18

ceed udu en, Ot ee as 0.170 + .027, 0.87 + .19

For ‘‘TRREGULAR’’? CLASS

e ae o ey to ear ge o = 1.461 + .020, OC. Vse 11,61 +16

TO eo ee o= 1.210 + .016, C. V = 10.02 + .14

MERON eects a3 0.251 + .026, -1.59 + 21

However measured, the variability of the number of nipples in

the female is always significantly less, not greater, than in the

male.

Furthermore a rather noteworthy sexual differentiation seems

so far to have escaped notice. The mean number of nipples for

male pigs is in all cases higher than that for female pigs. Thus:

ALL Pres

Male isi SS 12.4365 + .0182

Vengo 11.9077 + .0159

Diference ... 5545. 0.5288 + .0242

CLASSIFIED AS REGULAR CLASSIFIED AS IRREGULAR

Bo 12.3425 + .0233 Malos 35. ae 12.5833 + .0234

Reale 11.7849 + .0214 Fansa . css: 12.0777 + .0232

Difference .... 0.5576 + 0316 Difference .... 0.5056 + .0330

In all cases the males have on the average more nipples than

the females. The regularity of the differentiation is brought out

698 THE AMERICAN NATURALIST [ Vou. L

by the accompanying table in which actual values have been re-

duced to per mille frequencies. Pigs with 12 nipples or fewer

are preponderantly females; pigs with 13 nipples or more are

preponderantly males.

Number of Nipples Male Female | Difference

8 0 3 +. 3

9 6 LY + 1.1

10 0.6 143.7 + 53.1

ll 162.7 217.9 + 55.2

12 332.0 370.0 + 38.0

13 167.6 154.5 — Ta

14 163.4 82.8 — 80.6

15 49.3 0.0 — 29.3

16 29.8 ge — 22.7

17 3.0 1.7 | — 13

18 1.0 ae | — 7

1000.0 1000.0 |

The correlation between the number of nipples on the two

sides are:

MANE oe oe a, .6359 + .0073

Fanos.. 63 cs .5419 + .0088

Difforante oo en.n. .0940 + .0114

AH Pige 2. sae, .6063 + .0055

The correlations are fairly high. Those for males seem to be

slightly larger than those for females.

CORRELATION BETWEEN THE NUMBER OF THE YOUNG IN THE

LITTER AND THE NuMBER OF MAMM2 IN THE DAM

The relationship | th ber of young per litter and the -

number of mamms in the female has at various times aroused

considerable interest. As Pearl? has pointed out, two kinds of

correlation are to be recognized. First, interracial correlation,

that between the mean size of the litters and the mean number

of mamme in the females of a series of races or species. Second,

intraracial correlation, that between the number of mamme in an

individual mother and the number of young that she bears.

It is the rather obvious interracial correlation that has given

rise to such statements as that of Gegenbaur: ‘‘Die Zahl der

Zitzen steht in inniger Beziehung zur Menge der Jungen.’’ It

2 Pearl, R., ‘‘On the Correlation between the Number of Mamme of the

Dam and Size of Litter in Mammals. I. Interracial Correlation,’’ Proc.

Soc. Exp. Biol. Med., 11: 27-30, 1913.

No.599] SHORTER ARTICLES AND DISCUSSION 699

was the problem of intraracial correlation with which Alexander

Graham Bell? was dealing when he studied the fertility of the

multi-nippled race of sheep at Beinn Bhreagh.

Notwithstanding the simplicity of the biological problem a cer-

tain amount of confusion seems to have arisen. Thus Parker

and Bullard (loc. cit.) state: 2

It is the chief object of our paper to discuss the relation of the size

of litters to the number of nipples in the domesticated swine, Sus

scrofa Linn.

But instead of determining the correlation between the num-

ber of teats of the sow and the number of her young they have

actually calculated the relationship between the number of sib-

lings in the litter in which a pig was born and the number of

nipples which she herself possesses! Surely it should not require

specialization in animal behavior to convince one that the teats

which are of real service to a young pig are not its own, but those

of its mother!

Pearl* has quite correctly determined the correlation between

the number of nipples in the individual mothers and the number

of young in their litters. This he finds to be very low,> r = 0.195

=+ .086. i

It is rather difficult to agree with Pearl in his statement that

It would seem, a priori, that natural selection should have operated

to bring about a high correlation, both intra- and inter-racial between

these two variables, size of litter and number of mamma in the dam.

There seems no reason whatever to suppose that natural selec-

tion would tend to produce a correlation between the number of

mamme in the mother and the size of her litters within a race,

providing it has produced an average number of nipples suffi-

3 Bell, Alexander Graham, Science, N. S., 9: 637-639, pl. 5, 1899; loc. cit.,

19: 767-768, 1904; loe. cit., 36: 378-384, ‘i912.

t Pearl, R.‘ ‘On the Correlation between Number of Mamme of the Dam

and Size of Litter in Mammals. II. Intraracial Correlation in Swine,’’

Proc. Soc. Exp. Biol. Med., 11: 31-32, 1913.

5 Wentworth (Jour. Agr. Res., 5: 1148, 1916) records another very low

coefficient on unpublished data, but does not state specifically whether it is

between the number of mamme of the mother and the number of her young

as in Pearl’s series, or between the number in a litter (weighted with their —

own number) and number of nipples in the individual pigs, as in the series

of Parker and Bullard. :

700 THE AMERICAN NATURALIST [Vou. L

ciently large to maintain the race. On the contrary, any theory

of ontogeny or phylogeny which demands the existence of a

mechanism to provide an embryo pig with the particular num-

ber of nipples which would agree closely with the number of

young she may be destined to bear as an adult would seem to be

not merely cumbersome, but unnecessarily teleological. Since

male pigs have more mamme than females, the cost to the organ-

ism is apparently not prohibitive! What one should expect as

the result of the action of natural selection would, therefore, not

be the development of a regulative mechanism to provide the

mother with a number of nipples in close agreement with the size

of her future brood, but the development of a number of nipples

sufficiently large for the needs of the race.

Pearl’s own data show only 7 out of 57 ‘‘disadvantageous’”’

combinations, and the table as it stands takes no account of early

deaths.’ Furthermore, his series is small, only 57 individuals,

and apparently hardly typical of swine as a class. Parker and

Bullard on the basis of a thousand litters show that the (em-

pirical) modal number of nipples is twice the modal number of

young, and that the average number of nipples is much more

nearly twice the number of young than in Pearl’s short series.

Thus the data of both Pearl and Parker and Bullard indicate in

the words of the latter authors that ‘‘disadvantageous combina-

tions in which the number of young pigs outrun the provision for

€ Natural selection can not be expected to accomplish more for the de-

velopment of any character than to bring it to and maintain it at a stage

of development necessary for the survival of the species in competition with

others. That correlation between the number of the young and the number

of nipples is not necessary under conditions of domestication is shown by

the classic observations of Minot on the guinea pig (Jour. Phys., 12: 103,

1891) in which he pointed out that in his studies 143 litters showed a

variation of from 1 to 8 in the number per litter, with a model frequency on

2 and an average of 2.5, although the number of developed mamme is two.

That the number of young born may regularly exceed the number of

nipples in a species persisting under natural conditions is shown by the

recent studies of Hill and O’Donaghue on the marsupial Dasyurus viver-

rinus (Quart. Jour. Mier. Sci, N. S., 59: 133-173, 1914) in which they

have shown that a remarkable number of eggs are discharged from the

ovary at each ovulation and that as a rule more young are borne than can

possibly survive because of the limited accommodation of the pouch.

7 Unfortunately trustworthy figures showing directly the mortality of

new-born or recently born pigs seem not to be available. That such mor-

tality is considerable is indicated by certain of the figures given for another

purpose by Evvard.

No.599] SHORTER ARTICLES AND DISCUSSION 701

milk, cannot be of frequent occurrence.’’ The development of

just such a ‘‘ factor of safety’’ and not the origination of an intra-

racial correlation is, as emphasized above, just what one would

expect of natural selection.

Natural selection, if operative, should, however, bring about an

interracial correlation, and this is exactly what observant biol-

ogists have always noted and Pearl has expressed statistically by

the value r—=.594 + .046, with non-linear regression—a value

distinctively higher than that for the intraracial relationship.

hus, as far as they go, these observations instead of evidencing

against natural selection, actually show the very conditions to

exist which might be expected as the result of the action of this

factor of organic evolution.

INHERITANCE OF NUMBER AND ARRANGEMENT OF NIPPLES IN

SWINE

Attempts at the Mendelian analysis of inheritance of number

and arrangement of mamme in swine have been made by Went-

worth, who has suggested that the presence of rudimentary

nipples is a sex-limited,® sex-linked,’® or sex-limited** character,

His final stand is that the pair of rudimentaries posterior to the

inguinal pair behave as a Mendelian unit character in heredity,

but that somatically it develops in males, which are RR or Rr,

but in the females only when they are RR, where E indicates the

presence and r the absence of the factor for rudimentaries.

It is interesting to return to the sexual dimorphism with re-

spect to number of mammæ demonstrated above on the basis of

Parker’s and Bullard’s splendid series of data and to consider

it in connection with the hypothesis advanced by Wentworth.

Pearson many years ago showed”? that with continued random

mating the distribution in any generation subsequent to an orig-

inal random pairing of RR and rr individuals is

4RR + 4Rr + łrr.

8 Wentworth, E. N., ‘‘Inheritance of Number of Mamme in Swine,’’ Rep.

Am. Breed. Ass., 8, 1912.

9 Wentworth, E.. N., ‘‘Another Sex-limited Character,’’ Science, N. S.,

35: 986, 1912. sie

10 Wentworth, E. N., ‘‘Sex-linked Factors in the Inheritance of Rudi-

mentary Mammæ in Swine,’’ Proc. Iowa Acad. Sci., 21: 265-268, 1914.

11 Wentworth, E. N., ‘‘Rudimentary Mamme in Swine a Sex-limited Char-

acter,’’ Science, N. S., 43: 648, 1916.

12 Pearson, K., Phil. Trans. Roy. Soc. Lond., A, 203: 59-60, 1904.

102 THE AMERICAN NATURALIST [Vou.L

Both Pearl! and Jennings have followed him in this point. If

the thousand litters studied by Parker and Bullard come from a

population homozygous and heterozygous with respect of a pair

of rudimentary nipples in the 1:2:1 proportion and mating at

random,** then three out of four males as compared with one out

of four females should, if Wentworth’s hypothesis be correct,

show the pair of rudimentaries. Thus the average number of

mamme in the males should be 1 higher than in the females. As

a matter of fact it is .529 + .024 higher.

Further discussion on the basis of the present data would of

course be idle.

In his largest paper Wentworth?® has presented data which

indicate sensible parental and grandparental correlations for

number of mammez. In view of the irregularity of the frequency

distributions due to the modes on the even numbers and the

smallness of the series, as well as the fact that the number of

boars was very limited, little weight is to be given to the exact

numerical values of his coefficients.

A more detailed analysis of the extensive series of data col-

lected by Parker and Bullard may throw considerable light upon

the problem of inheritance. The results must be expressed in

terms of fraternal or sororal correlation. Those who are so ob-

sessed with Mendelian theory that they are unwilling to learn

anything about a series of data for which their method fails,

should discontinue the reading of this review at this point.

Correlation between the number of nipples in siblings may be

very readily found by means of intra-class correlation formule*’

involving first and second moments for the individual classes

(litters).

Let £m be the number of nipples in a male, x; the number of

nipples in a female pig, nm the number of males and ny the num-

ber of females in a litter of nm -+ npn individuals. Let 3

denote summation within the litter and 8 a summation for litters.

For any litter the moments are therefore S (am), Z(tm*)y S(O)

18 Pearl, R., AMER. NAT., 47: 606-609, 1913.

14 Jennings, H. S., Genetics, 1: 64, 1916.

15 Random mating of course applies only to the particular character in

question, which is one which would hardly be consciously selected by any

breeder.

18 Wentworth, E. N., ‘Inheritance of Mamme in Duroc J ersey Swine,’’

Amer. NAT., 47; 257-278, 1913.

17 Harris, J. Arthur, Biometrika, 9: 446-472, 1913.

No.599] SHORTER ARTICLES AND DISCUSSION 703

3 (x). Since in a symmetrical intra-class correlation table the

variates are weighted in an (~—1)-fold manner the fraternal

correlation for males is given at once by direct summation from

the data table of Parker and Bullard by the formula, written for

simplicity in an entirely unreduced form,

SIE (En)? — SE (En) (ies — 1) @m)] \

p = —Slttm(tm = 1) _S[7m(%m — 1)]

Sm — 1)2(@m*)] _ (Be ip!

S[1tim(%m — 1)] S[11m(%™m — 1)]

or substituting actual values

Tr tee = O20 Æ 019.

Apparently complex, the formula is really on closer inspection

very simple indeed.

One altogether similar for the females gives the sororal corre-

lation

373 + .018.

Tent =

Thus the correlation for the females is .050 + .026 higher than

that for the males.

For the cross correlations, that between number of nipples

borne by male and female pigs of the same litter, the constant

is given by

SIZ (Em) E (27)] _ S[nsZ(%m)] x S[rm=(xs)]

ee Se as

Sinn (x 2)] - (Sezen) [SIN E(Em)] _ (e ?

S(Mmns) S(Mmns) S(nynm) S(nstm)

Taz, = .287 + .020,

a value apparently distinctly lower than that for either males

or females alon

If the ee a between the siblings be determined irre-

spective of sex the value is

SIZ (a) P Siin — 1)2@)] Y

_Sin(n — 1)] -( S[n(n — 1)]

= Sim — DERA] - (4¢ — 1)2(z)]\’’

704 THE AMERICAN NATURALIST [ Vou. L

where the n and x without subscripts denote number of individ-

uals per litter and number of nipples per individual, without

reference to sex. Numerically

Tri = 005 + .019.

This is lower than the relationship for either of the sexes in-

dividually considered, just as one might have predicted on a

priori grounds from the low value of the cross correlation and

from the differentiation in the number of mammez in male ang

female pigs.1®

The correlation coefficients here given show that there is a

very material degree of resemblance with respect of nipple num-

ber in pigs from the same litter.1° Indeed the correlation is

about one third of the maximum value. Such correlation can

be due only to differences in intra-uterine environment or to a

strong inheritance of nipple number. The latter seems by far

the more probable explanation.

J. ARTHUR HARRIS

18 Harris, J. Arthur, ‘‘On Spurious Values of Intra-class Correlasias

Coefficients Arising from Disorderly Differentiation within the Classes,’’

Biometrika, 10: 412-416, 1914.

19 These values of the fraternal correlation will be but slightly influenced

by the weighting of the individuals in the determination of the correlations,

since nipple number is but slightly correlated with number in the litter.

eos :

s : ii

oted to the Advancement of the Biologi

_ Special Reference to the Factors of.

The American Naturalist

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

sent to the Ene of THE AMERICAN NATURALIST, Garrison-on-Hudson, New Yor

ho! rticles containing summaries of research work bear ring on the

Sperra of Tabea evolution are especially welcome, and will be given preference

in publica’

kandjad reprints of cpi bana are supplied to authors frec ef charge.

üne

Further reprints will be supplied at c

Su tions and adve

subscription ae is four dollars a

Aroeri should be sent to the tag agian The

Foreign postage is fifty cents and

Canadian postage twenty-five cents additional. The rge for single copies is

forty cents. The advertising rates are Four Dollars for a page.

THE SCIENCE PRESS

Lancaster, Pa.

Garrison, N. Y.

NEW YORK: Sub-Station 84

Entered as second-class matter, April 2, 1908, at the Post Office at Lancaster, Pa., under the Act of

Congress of March 3, 1879.

= NATURAL HISTORY OF MAN

The a of ———— and Vicinity

ES S. BOYER, A M., F.R.M.8. |

A descri n of the diatomaceous flora o epee org ah

in nook form re ee the J. Fag palapah company. 143 piei

Se OLDMAN, 77 Brixton Hil, London, England

a To Lepidopterists

2 w. ı F. H. ROSENBERG

57, Haverstock Hill, London,

More than 700 a illustrations, enlar, a ae

complete inde Bound in Art

Sent pustage .

dress Charles S. Boyer, Pa

1649 Columbia Philadelphia, Pa.

SCIENCE

A WEEKLY JOURNAL ne TO THE

ADVANCE ENT OF SCIE.

FRIDAY, NOVEMBER 24, 1916

Bs eee Working-day Rational? Professor Frederic —

| "Phe Care of Pamphlet Collections

: Tracy I.

The Brain Calor the U. S. National prasi Dr. Ales f

| The Yale - of Sigma Xi 2

Thal Endowment of a Medical School at the Univonity

Chicago.

ientific Notes and News. a ate fe

Cs xert a | ; upon Itself ? Professor £ r

eiat ‘Vision and Orientation: g G. Daban.

The Number of Bacteria : ri Poe wk

‘in Living Cells: : ae

ao oo = = n

THE

AMERICAN NATURALIST

Vou. L. December, 1916 No. 600

EXPERIMENTAL INTERSEXUALITY AND THE

SEX—PROBLEM?!

RICHARD GOLDSCHMIDT

KAISER WILHELM-INSTITUT FUR BIOLOGIE, BERLIN

Untm a comparatively recent time the sex-problem

was one of the white spots on the map of biology. There

is hardly another problem which has been such a play-

ground of dilettantism, and if we look through the older

literature on the sex-problem we find almost as many

philosophers and economists inventing sex-theories as

there are biologists. But there is scarcely another prob-

lem in biology which in the brief period of a decade and

a half has emerged from a state of absolute ignorance

into one of most hopeful knowledge. The all-important

step which has been made during this time is the com-

plete insight into the elementary mechanism which pro-

1 Evening lecture delivered at the Woods Hole Marine Biological Labora-

which appeared following a preliminary report in Sitzungsber. Ges. Morphol.

Physiol. München, 1911, namely: Goldschmidt, R., ‘‘Erblichkeitsstudien an

Schmetterlingen,’’ I, Ztschr. induct. Abstammungsl., 7, 1912; Goldschmidt,

R., u. Poppelbaum, H., ‘‘ Erblichkeitsstudien an Beekie apin" II, Ibid.,

Vol. 12, 1914; PRR SEER H., ‘‘ Beitraege zur Kenntnis, etc.,’’ Ibid.;

Goldschmidt, R ‘¢A Preliminary Report, ete.,’’ Proc. Nat. Ac. Sc., I, 1915;

Goldschmidt, R., ‘‘Genetic Factors and Enzyme Reaction,’’ ReneS Vol. 43,

1916. See farther the related chapters in Goldschmidt, R., ‘‘ Einführung

in die Vererbungswissenschaft.,’’ 2. ed., Leipzig, 1913, and in Correns, C.,

and Goldschmidt, R., ‘‘ Vererbung ii Bestimmung des Geschlechts,”

Berlin, 1913. A complete áccount of the entire work, with the n

illustrations, is in preparation. I am greatly indebted to Dr. R. re Spaeth

for revising this MS.

705

706 THE AMERICAN NATURALIST [Vou. L

duces the normal distribution of the two sexes. To-day it

is well known to every one how this insight was reached

practically simultaneously by the study of Mendelian in-

heritance and the cytological investigation of the chromo-

somes and how the two solutions are in most. wonderful

harmony. The outstanding facts which I regard as one

of the corner-stones of modern biology are familiar even

to the beginner in biology. Cytologically one of the

sexes, the heterogametic sex, contains only one X-

chromosome, the other, the homogametic sex, two of

them. Every one knows how the maturation division

separates entire chromosomes and therefore the hetero-

gametic sex produces two kinds of sex-cells, i. e., with

and without X-chromosomes, but the homogametic only

one kind, all with X-chromosomes; and how chance-

fertilization produces again the two parental combina-

tions, that is, the two sexes. You are furthermore famil-

iar with the fact that the Mendelian experiments yield

practically the same result. If we substitute the term

‘‘sex-factor’’ for ‘‘X-chromosome’’ and ‘‘hétero-homo-

zygous’’ for ‘‘hetero-homogametic,’’ the facts are iden-

tical. One sex is heterozygous for the sex-factors, say

Ff; the other homozygous, say FF ; the first one produces

two kinds of gametes, the second only one, and chance-

fertilization results in equal numbers of the parental com-

binations. The results of experiments on sex-limited -

inheritance have shown, finally, that both sets of facts are

the same thing, only expressed in different language, in

other words, that the X-chromosomes are the vehicles for

the distribution of the sex-factors. If we state further

that there are animals in which the heterozygous Sex is

the female, and others where it is the male, we know the

elementary facts from which any further study of the

sex-problem has to start.

Now that we know the elementary mechanism of sex-

distribution, can we regard the sex-problem as solved?

I do not think so. What are the sex-factors and how do

they determine sex? Are the two sexes clean-cut alter-

_ natives and is it therefore impossible to transform one

No. 600] EXPERIMENTAL INTERSEXU ALITY 707

into the other, or are they nothing but limiting points of

a series, which might approach each other or even become

interchanged? These and many other questions can only

be approached by experimental modifications of the

normal sex conditions. There is one road towards this

goal, the application of the well-known influence of the

internal secretion of the sex-glands upon the sex-char-

acters in castration- and transplantation-experiments,

as well as the study of analogous experiments by nature.

Another approach is rather an unexpected one; this has

been followed in my experiments, an account of which I

now have the pleasure of giving you.

Insect-breeders have long known that in crosses of

species as well as of geographic varieties a comparatively

high percentage of sexual abnormalities are produced. |

Furthermore, any collector of moths knows that similar

abnormalities, usually called gynandromorphs or hermaph-

rodites, appear occasionally in nature and one of the

moths in which these occur is the gipsy-moth. Since this

moth has a very wide range of geographic distribution

through very different climates a considerable differ-

ence of geographic varieties is to be expected. Fur-

thermore, as is well known, the sexes are extremely

dimorphic. Therefore, seven years ago I began hy-

bridization experiments of European and J apanese

gipsies. I was rather fortunate in striking the right

forms which gave, from the outset, the most interesting

results. )

The first result was that crosses of Japanese females

with European males yielded normal offspring, whereas

in the offspring of the reciprocal cross, European female

X Japanese male, all males were again normal, but all

females showed in all parts of their bodies admixtures

of male characters. I first called these animals gynan-

dromorphs. But as this term is usually applied to

animals with bilateral or antero-posterior or similar

mosaic of the two sex-characters, it seems advisable to

use another term for these forms, which in general repre-

708 THE AMERICAN NATURALIST [Vou. L

sent a definite step between the two sexes.2 The phe-

nomenon shall therefore be called intersexuality. Further

experiments now proved that intersexuality segregates,

F, giving normal and intersexual animals. It was

further shown that in some experiments the females re-

mained normal and the males became intersexual. When

the experiments were repeated, however, different results

appeared. And as the material used came from different

strains, suspicion arose that there are many different

races of gipsies, differing in regard to those things which

are responsible for the intersexuality. This suspicion

was strengthened when it became probable that the pecu-

har thing responsible for intersexuality could be influenced

by external conditions. If this is the case, conditions in

Japan ought to be very favorable for an origin of such

racial variation, as the Japanese islands show climatic

conditions varying from an almost arctic to an almost

tropical climate. These suppositions turned out to be

correct, as was proved by the further study of these ques-

tions in Japan in 1914 and with the material brought

from there to this country.

In order to make the results clearer and their bearing

on the sex-problem more evident, I want to apply just

the opposite method from that used in the research work,

and to give first the general interpretation of the results,

which differs only in a few points from that used since

1911, and then the experimental facts from which it has

been derived. What we have to explain is that two

nearly related forms, both normal in regard to sex-in-

heritance, produce, if crossed in one direction, normal

offspring, in the other direction, normal males and inter-

sexual females; (2) that, as we may now say, the degree

of intersexuality is definite in a given cross, but different

in different crosses; (3) that intersexuality shows Men-

delian segregation; (4) that males may become inter-

sexual too in certain crosses.

The explanation for these and the other facts later to

2 The special meaning of this will be discussed in another paper in con-

nection with very interesting new facts. |

No. 600] EXPERIMENTAL INTERSEXUALITY 709

be reported is the following: Both sexes contain the an-

lagen for either sex. In both sexes, irrespective of the

zygotic constitution, both anlagen might become patent.

Which one is to appear depends entirely upon the quanti-

tative relation of both.? If we apply the usual symbols

and keep in mind that the female sex is here the heter-

ozygous one we have the following formule:

|[FF|Mm=9, [FF|MM=¥.

(It will soon be explained why the female set FF is put

in a square.) The female set as well as the male set act

independently and with a definite quantitative strength.

In order to have a convenient term we call this quantita-

tive value of the sex-factors their potency or valency.

3 The idea that sex-determination is a quantitative rather than a qualita-

tive process is of course not new. Practically all writers about the cytology

of sex-determination have developed such ideas, beside the well-known old

and new metabolistic theories. Such a quantitative view was proposed by

myself in 1904 (‘‘Der Chromidialapparat, ete.,’’ Zool. Jahrb. (An.), 21)

Zellf., 6). Other such deena were proposed by R. sig 1905-07

r Sexuellen Differenzierung, ete.,’’ Verhdlg.

deutsche Zool. Ges., 1908, 1906, 1907), based on his views about the nuclear-

plasmic relation. Again others of a strictly quantitative character were

discussed by E. B. Wilson (‘‘Studies on Chromosomes,’’ III, Journal Ezp.

in rmato-

-A

—

genesis, etc.,’’ Trans. Amer. Phil. Soc., 21, 1906), and by Th. Boveri

(‘* Ueber ieee pe ae zur Aendhioehalsetinrn mung Sitzber.

physikal, medizin. Ges. Wiirzhburg,’’ 1908-09). A full discussion of the

relative values of hace i and ne views in regard to sex-de-

termination is given by Th. H. Morgan (‘‘A Biological and Cytological

Study, ete.,’? Jour. Exp. Zool., 7, 1900). ae decides for the latter. How-

rmer views. The first attempt to prove a quantitative theory of sex-

determination experimentally has been made by R. Hertwig in his well-

own experiments with frogs (since 1905). Quite another type of theory,

practically the same as is to be used in this paper, combining the quantita-

tive view with the Mendelian and cytological results, has been developed on

an experimental basis in my papers from 1911 and 1912 (l. c.) and has

been adopted by many writers (i. e., Doncaster and Harrison, Standfuss,

Witschi). A third view has since been developed by Riddle in a series of

preliminary papers (Carnegie Year Book, 1913, Science, Vol. 39, 1914,

Natur., Vol. 50, 1916, etc.). His theory is based partly on chemical

Studies of pigeon eggs, partly on hybridization experiments with doves. It

is not impossible that in the latter something like intersexuality is pro-

duced, if I understand the short accounts thus far coer,

710 THE AMERICAN NATURALIST [Vou L

If you like to form a definite idea you might assume that

the potency means a certain concentration of enzyme

which acts according to well-known laws.* In order to

make the situation clear we assume that we are able to

measure their potency. And we find that the female

factorial set [FF] is 80 units strong and every male factor

M = 60 units. Then in the female formula [FF] Mm the

female set overpowers the one M present by 20 units,

whereas in the male formula MM the two Ms with

the value of 120 are 40 units stronger than [FF] =

_ Now we face two possibilities. Either the slightest pre-

ponderance of one set over the other, say by only one

unit, is sufficient to determine the male or female sex,

or there is a definite minimum of preponderance neces-

sary—we call it the epistatic minimum—beyond which

one or the other sex appears. Let us now suppose this

minimum to be 20 units. Then of course 40 units are

left between the two extremes male-female preponder-

ance. If we call the difference value between the male

and female factorial set e, then we have a female, when

[FF] — M = > 20 and a male, when MM — [FF] > 20, or

in other words the limiting values for e for the two sexes

are + 20 and — 20. We can now express this conception

graphically in the following diagram, where the values

Oe g Ti í EF i eat, Bo

-20 -10 o

"Nia. 1

of e are arranged on a straight line. Individuals to the

right of + 20 are females, to the left of — 20 are males.

But what of the intermediate points? These are the

intersexual animals; if they are heterozygous for M they

are intersexual females and if they are homozygous for M

they are intersexual males. How does this diagram now

. 4 For proof that this conception comes near the truth see Goldschmidt, R.,

Genetic Factors and Enzyme Reaction,’’ Science, N. S., 43, No. 1099,

1916. Very important new facts will be published later, which will prob-

ably enable us to replace the symbolistic — language, used here,

by more definite physico- chemical conceptions.

No. 600] EXPERIMENTAL INTERSEXUALITY 711

explain the fundamental experiment? Suppose we have

two races both normal in regard to the quantitative be-

havior of their sex-factors, but with different absolute

values of the potencies. The values are supposed to be

the following:

Weak European Race Strong Japanese Race

2 [FF| Mm Mm

80 60 100

3 FFM FF| MM

80 60 60 100 80 80

It is evident that both races, if bred true, behave nor-

mally. Here we must now add that the female factors

or anlagen are inherited exclusively maternally, without

any paternal influence. Therefore the square. But the

M’s are typical Mendelian sex-factors. Let us cross now

a Japanese 2 with an European ¢. F, is then

F, 2 |FFi Mn, F, ¢ |FF| MM.

100 60 100 80 60

The value e is then + and — 40, the offspring is normal.

The reciprocal cross European female with Japanese

male gives

F,? |FF|Mm, F, 3 [FF] MM.

a s 80 80 60

Now we see that in the female e or FF—M=—0. The

animal is intersexual, exactly half-way between male and

female.

Instead of deriving further theoretical expectations

we shall now see how the experimental facts fit these

general conceptions of a quantitative nature of sex-de-

termination. The first point is of course the question of

the different absolute and relative potencies. There are

primarily two ways open for testing it. One is to influ-

ence these potencies experimentally. Only a few pre-

liminary steps in this direction have thus far been made,

712 THE AMERICAN -NATURALIST [Von. L

which may be omitted here. Another way would be to

find many races which differ constantly in regard to the

potency of these factors, which could be shown by the

results of cross-breeding. The expectations are, then,

that with increasing value of. M, in crosses, where these

races are used as males in combination with ‘‘weak’’

females, a corresponding type of female intersexuality

must appear, giving a complete series from femaleness to

maleness. And if we can find a race with so high a

potency of M, that the combination lies beyond the

epistatic minimum for maleness, then all the would-be

females will be transformed into males. I now have at

hand such races from Europe and Japan and can pro-

duce at will and in 100 per cent. of the offspring all grades

between the two sexes. Thus we have one Japanese race,

the race G, medium strong in regard to the potency of the

factor M. If we cross these males with females of the

Japanese race K, which shows comparatively low potency

of the female factorial set, all would-be females in F,

are slightly intersexual. We might put them at the point

+ 15 in our diagram (Fig. 1). Their antenne are

feathered, but less than in males, a portion of the wings

assumes the brown color of the male, there are not as

many eggs as in a normal female, but the mating in-

stincts as well as the copulatory organs are still female

and the eggs may be normally fertilized. Then there is a

European race F and a Japanese race H, both with a

still lower potency of the female factorial set. If we

cross these with the same Japanese males G we get some-

what slighter female intersexuality. All secondary sex-

characters are more male-like; the instincts are still

female and the animals attract the males and mate. Then

one of the characteristic hairy egg sponges is laid, but it

contains no eggs, only hairs. The copulatory organs are

already changed in the direction of the male and no suc-

cessful mating and egg deposition is possible, although

the abdomen is filled with ripe eggs. Then there 1s

another European race F with very low potency of the

female factors. If we mate these with the same males G,

No. 600] EXPERIMENTAL INTERSEXUALITY 713

intersexual females appear which are more than half-way

etween males and females. The secondary sex-char-

acters are almost male. The instincts and behavior are

about intermediate between the sexes. Males are

searcely attracted or not at all and no mating occurs.

The copulatory organs show the strangest combinations

of the male and female types, but there are still typical

but rudimentary ovaries left. There is now another

Japanese race in my possession, the race X. This one

exhibits a still higher potency of the male factor M. If

we cross this one with the European race F, a still higher

degree of intersexuality appears. Now we have animals

which externally are almost indistinguishable from true

males. But certain characters, especially 1 in the copula-

tory organs, still show their female origin. The instincts

are entirely male and they try—always unsuccessfully—

to mate with females. But the most interesting feature

is the sex gland. This is a body looking externally like

a testis, but showing in sections every single step between

an ovary with nothing but immature eggs through a

mixture of ovarial and testis tissue to a real testis. This

is of course the highest grade of intersexuality which can

be reached. The next step would be the complete trans-

formation of the would-be females into males. And this

can be obtained too. I have two Japanese races O and

A which show such a high potency of M that crossed with

any European females and the Japanese females H they

produce nothing but males, all would-be females being

converted into males.®

5 It might be added here that two different lines of female intersexuality

can be distinguished in regard to the most conspicuous feature, the color of

the wings. In one line intersexuality begins with white wings, then dark

cunei appear on the wings; they grow larger and larger, forming streaks

along the veins until finally only fine white spots are left on a dark wing.

The second series shows even in the first grades of intersexuality male color

all over the wings without any streak formation. Which type appears

depends on both races involved, and is due to physiological conditions in

regard to pigment formation which are not yet entirely clear. The inter-

sexual males always exhibit the first type. Color photographs of an almost

complete series of the second type are given in my papers from 1912 and

1914, Photographs of the first type in the female series are not yet pub-

lished. A complete series of the transformation of the female copulatory

714 THE AMERICAN NATURALIST [ Vou. L

I think these facts alone would be sufficient to prove

that the above given quantitative conception of sex-de-

termination is the right one. But we can quote further-

more a real experimentum crucis. We have seen that the

very weak European race F crossed with the medium

strong Japanese males G give a fairly high grade of

intersexual females. In our diagram of the values of e

we might put them at the point — 10 (Fig. 2). The same

< =

x iF;

€- as pode ees Q

: -20 19 o +10 +20

inii .

5 z

FIG. 2

weak females crossed with the very strongest males A

gave nothing but males, the would-be females standing

now on the point, say —25. Now the strongest of the

weak races in regard to FF, i. e., the Japanese race K,

gave with these same males G slight female intersexual-

ism, say at point +/15. If our conception is right, the

same females, crossed with the very strong males A,

ought to give medium intersexual females, to be put at

the point 0. And this is the actual result.’

Now we have reached a point where we want to know

how the complexes [FF] and M are inherited. Anybody

familiar with Mendelian analysis knows which results are

to be expected if F and M are independent Mendelian

factors. These tests, so far as fertility allows them to be

organs based on our material is pictured in the paper of Poppelbaum

(l. ¢.). S of the transition from an ovary to a testis are

given in my pa of 1914. In regard to the occasional appearance of a

single normal Tenkai in cultures where all females are transformed into

males, see our publication in the Proc. Nat. Ac. It was supposed that they

originated from a case of ‘‘non-disjunction.’’ Their behavior in breeding

does not agree with this supposition but suggests another explanation, which

at the same time explains the occasional appearance of intersexual males

the Massachusetts race, which behaves about like the Japanese race K.

No. 600] EXPERIMENTAL INTERSEXUALITY | 1715

made, prove, of course according to expectation, that the

factor M is Mendelian and carried in the X chr

All results of F, and back crosses agree in this ERA

But what is FF|? Originally I believed that I could

prove its Mendelian character too. This was a mistake,

produced by the interference of another phenomenon,

connected with the wing-color inheritance. The experi-

ments made since my first communications prove thus far

that the complex is inherited only through the

mother, maternal grandmother, etc., 7. e., in the proto-

plasm of the egg.

We are now prepared to consider the production of the

intersexual males, thus far omitted in our discussion. A

series of them has been produced extending almost to

femaleness. They are easily distinguishable from inter-

sexual females (a point which is of special interest; one

explanation being that factors contained in the Y chromo-

some are responsible for it). The intersexual males

always exhibit the first of the above-mentioned (foot-note

page 713) types in regard to wing coloration, white streaks

appearing on the dark wings. While the wings assume

more and more the female shape, the dark color becomes

confined gradually to the wing venation, and in the ex-

tremest types thus far bred only a few brown spots ap-

pear upon some veins. All the other characters, like size

of abdomen and copulatory organs, change hand in hand

with this. But the behavior of the sex glands is not yet

clear. In low grades of male intersexuality the testis

always contains some ovarial tissue.” But the highest

grade intersexual males of almost female exterior con-

tained a paired’! sex gland of somewhat testis character,

filled with giant bundles of apyrene spermatozoa, and

containing no eggs.

If we now deduce tiom our general interpretation how

these interesting males might be produced, we realize that

7See picture in Poppelbaum’s paper. The lower grades of male inter

Sexuality are photographed in our papers from 1912 and 1914. Pictures

of the higher grades are not yet published.

8 The ripe testis is an unpaired organ.

716 THE AMERICAN NATURALIST [Von. L

they are expected to appear when [FF] is comparatively

high in potency and MM comparatively low. There are

two possibilities for this event. (1) If we revert to our

original example of a cross between weak European and

strong Japanese races, we have in one direction:

PUF xo dan. g

[FF| Mm x [FF| MM.

80 60

kas

100 80 80

As [FF] is inherited maternally, any generation of this

cross will have the weak set [FF] — 80. There is no com-

bination of two M’s possible which is not at least 20 units

higher than FF, therefore no intersexual males can oc-

cur. Now take the reciprocal cross:

Jap X Buig

[FF|Mm x FF MM.

100 80 60 60

The F, males are [FF] MM and therefore normal, as

100 80 60

e=—— 40. F, from this cross has again the maternal

and grandmaternal set |FF —100. The Mendelian

factors MM are recombined and we get the combinations:

MM and MM.

80 60 60 60

The latter males are therefore:

[FF| MM.

100 60 60

This means that e has just the limiting value —20. It

follows, that if we have two races, in which the relative

values differ only slightly from this example to the dis-

advantage of the weaker M, say M = 59 instead of 60,

male intersexuality is to be expected in the F, generation

of a cross, where the mother belongs to the stronger race.

This is indeed one of the actual facts.

According to the above derived formule these inter-

sexual males ought to number exactly one half of the

No. 600] EXPERIMENTAL INTERSEXUALITY Ti

male individuals. But this has so far never been the

case. In order to understand it we have to point now to

a fact which we omitted in our previous discussions. We

have located the intersexual animals, males and females,

at a certain point between the two sexes, as represented

in the two diagrams. But in fact these points are only

the mean of a certain range of variation in regard to the

grade of their intersexuality. Whether this means that

the potencies of the factors are variable or only their

ultimate effect shows variation produced through influ-

ences during the development, is a question not to be

discussed here. The fact is, however it might be caused,

that the value of e as the measure of intersexuality, is

variable around a mean. If this is the case, the following

expectations are to occur (Fig. 3): If the values of e

LT ee fof ay

e = = Pi i å a

Aor -10 o +10 st ọ

Fie. 3

approach the epistatic minimum (20 and — 20 in our dia-

gram), then a point must be reached where the vari-

ability-curve stretches beyond this limit. That means—

curve 1—that in cases of low grade intersexuality some

+ or — individuals might overlap into the normal. Of

course any position of these curves and a corresponding

numerical relation of normals and very slightly inter-

sexual animals might be met with, as the curves 2 and 3,

Fig. 3, indicate. And indeed such cases have been found

as well on the female as on the male side. Types of this

kind, realized in the F, crosses with male intersexuality,

are those with which we have dealt in the previous para-

graph. Thus far all facts agree with this conception, in

one culture up to F,. Further checks are in progress.

But there is another possibility for the production of

intersexual males. Let us assume that we could find two

races of the following constitution:

Strong race 9—[FF| Mm :g=|FF| MM,

100 60 100 60 60

718 THE AMERICAN NATURALIST [Vou. L

Weak race 9=|FF| Mm g= MM.

80 50 80 50 50

The F, cross of these, using the strong race as mother,

would give:

P, 9 [FF] Mm = normal female,

100 50

Fog MM = intersexual male, as e = — 10.

100 60 50

Every single male ought to be intersexual. - Through the

kindness of Dr. Machida in Tokyo, who first has reached

this strange result, I could bring back from Japan two

such races and they breed exactly according to ex-

pectation.

This is about the skeleton of our work around which

of course many interesting experimental, morphological,

embryological, cytological and physiological facts are

grouped. Iam rather optimistic in regard to the general

conclusions, which might be drawn from these facts as

well as regards the sex-problem as on some fundamental

questions of heredity. Combining these facts with the

work on hormone action as related to sex we can, I think,

form a pretty clear idea about sex-differentiation and

determination.?

If we put them in line with the facts of experimental

embryology concerning the determination problem we

see the outlines of a promising theory of heredity. But

this can be discussed only when a full account of the work

has been given.

9 The relation of gametic or zygotic intersexuality to the hormonic inter-

sexuality produced in castration and transplantation-experiments in Crus-

tacea, birds and mammals (especially the work of G. Smith and Steinach)

to which now, after F. Lillie’s discoveries, the case of the free-martin has

to be added, has been worked out in the chapter on Sex in my treatise on

geneties (l. ¢.), 2d edition. A detailed comparison between the moth-work

and the facts known about human intersexuality, containing, too, a discus-

sion of the relations of zygotic and hormonic intersexuality, was sent to

Germany for publication about half a year ago, but it is not known whether

it has been published there or adorns the waste-paper basket of His Britannic

Majesty’s censor,

PIEBALD RATS AND MULTIPLE FACTORS

E. C. MacDOWELL

STATION FOR EXPERIMENTAL EvoLuTion, Corp Spring Harsor, L. I.

INTRODUCTION

THE experiments of Castle and Phillips (:14) with pie-

bald rats afford the largest mass of recorded data on the

influence of selection in mammals. For 17 generations,

the area of pigmentation on their hooded rats has been in-

creasingly modified. In one line (the plus race) the pig-

mentation has been extended; in the other line (the minus

race) it has been reduced. When rats from the plus or

minus race are crossed with fully pigmented rats, such as

the normal wild, or the Irish variety, the hooded pattern

behaves as a simple Mendelian recessive, disappearing in

the first generation and reappearing in one fourth of the

offspring in the second generation. These results lead to

the conclusion that hoodedness appears when a certain

germinal unit, or factor, is in a zygote in a homozygous

condition. Besides this, Castle concludes that the factor

determining hoodedness fluctuates, and, in accord with its

fluctuations, the amount of hooding varies. It follows —

that the selection of extreme grades of hoodedness results

in the simultaneous selection of extreme variations of the

factor. Moreover, Castle (:16, p. 722) finally concludes

that the selection of these extreme grades of hoodedness

influences the direction in which the factor for hoodedness

varies.

These conclusions-bear on one of the most generally in-

teresting and vital questions before biologists. If, be-

sides deciding which individuals shall mature and repro-

duce, selection can influence the direction in which the

units of inheritance, or factors, vary, there can be no

question but that

719

720: THE AMERICAN NATURALIST [ Von. L

selection, as an agency in evolution, must then be restored to the im-

portant place which it held in Darwin’s estimation, an agency capable

of producing continuous and progressive racial changes (Castle, 15b, .

COT),

Castle’s experiments have justly become famous. For

eight years they have been continuously in progress; they

have involved large amounts of arduous labor; they have

been conducted with unflagging zeal and high ideals of

scientific attainment. The conclusions drawn from such

an important investigation should receive painstaking

consideration.

The writer has been conducting selection experiments

` which have led him to conclusions different from those

reached by Castle. Although these experiments have not

involved the expenditure of so much labor and time as did

Castle’s work, they include three times as many genera-

tions and four times as many individuals as are reported

by Castle. One investigation was on rats, the other on

flies, yet there are so many similarities in the results that

the writer was led to make a careful analysis of Castle’s

papers in an attempt to discover the basis for the conflict-

ing conclusions. The final result of this study was to

make the writer feel that the following statements in re-

gard to the hooded rats are too positive.

All the evidence we have thus far obtained indicates that outside

modifiers will not account for the changes observed in the hooded pat-

tern, T a clear Mendelian unit (Castle, :15b, p. 722).

Aare e ean be no doubt that only a single genetic factor is’ here

navsived (Castle, :16, p. 95).

It is precisely this last named eategory of cases [a single RER

basis undergoing quantitative variation] which alone can explain our rat

results (Castle, :15b, p. 725).

Energetic attacks have been made on the interpretation

Castle has given to his results, and certain unwarranted

criticisms have been duly answered. That the theory of

multiple factors may be applied to the results as pub-

lished in 1914 in the Carnegie Institution Publication No.

195 was indicated therein by Castle, and further empha-

sized by Muller (:14). Most of the criticisms of the ex-

periments with the hooded rats have been based on the

No. 600] PIEBALD RATS AND MULTIPLE FACTORS T21

generalizations that have been made, and not directly on

the data. In this paper the writer has used the original

data, making verifications where possible, and recalcula-

tions of many of the constants. A few inconsistencies

and arithmetical errors were found.

That there may be no misunderstanding as to the na-

ture of the multiple factor interpretation, the following

scheme is suggested. In the absence of any factor that

determines uniform color—in other words, in the pres-

ence of two doses of the factor for hoodedness—the

amount of pigment on hooded rats may be influenced by

several factors. Some of these increase, others reduce,

the pigmented area. The factors that increase the pig-

mented areas (plus factors) form Mendelian pairs (alle-

lomorphs) with the factors that decrease the pigmented

areas (minus factors). Dominance is lacking; if a factor

is contributed to the zygote by both parents, that factor

has more power than if it had come from only one parent.

Furthermore, environment, or other conditions which are

not inherited, being outside the germ plasm, have such a

modifying influence on the pigmented areas that the po-

tential differences between individuals determined by dif-

ferent combinations of factors in the germ plasm, are fre-

quently concealed. It is not pretended that this is the

only application of the multiple factor hypothesis that

can be made, but it is hoped that the following arguments

may become more significant with this suggested applica-

tion in mind.

The writer herein undertakes to show that the concep-

tion of multiple factors may still be applied to Castle’s

data. The points that favor the multiple factor interpre-

tation of the rat experiments, as well as certain objections

that are said to definitely disprove this theory, are brought

together in the following paragraphs.

Points FAVORING THE MULTIPLE FACTOR INTERPRETATION

1. The gradual divergence of the plus and the minus

races may be brought about by the sorting out of groups

722 THE AMERICAN NATURALIST [Von. L

of different factors. It has been generally recognized

that this is a possible conception. The following authors

have considered this point: Castle and Phillips, :14, Mul-

ler, :14, Hagedoorn, :14, MacDowell, :15.

2. The hybrid ancestry of the original parents affords

a source for a large amount of heterozygosis. It is the

reduction of this heterozygosis that selection is supposed

to accomplish, in separating the two races, plus and

minus.

3. Such a reduction of heterozygosity would be hin-

dered by the large number of matings made between rats

less closely related than brother and sister. This point

has been discussed by Muller.

4. By breaking the correlation between the soma and

the germ plasm, environment has probably played a large

part in hindering the reduction of heterozygosity. Ap-

parently it has been assumed that there is a close rela-

tionship between the germ plasm and the soma, that the

smooth curve of the averages.in successive generations

proves that the germinal variations, to which the rise in

the curve is due, are small and constantly occurring.

But, since the rôle of environment is not known, the

gradual advance in the averages can not prove anything

as to the size of the germinal variations. The presence

of regression makes it clear that environmental, or extra-

germinal, influences are active in producing variability

in the hooded pattern. Regression is really a measure

of the degree of independence of the soma and germ

plasm. Regression expresses the inverse relationship

between the actually tested breeding possibilities and the

appearance of the parents. There can be no question as

to the activity of environmental influences; of their power

and nature nothing seems to be known. The immediate

environment of the undifferentiated blastomeres is prob-

ably as important a factor in the final appearance of a

character as the germ plasm itself.‘ The factors in the

germ plasm are like chemicals that will react in a definite

way in connection with certain other chemicals; when

No. 600] PIEBALD RATS AND MULTIPLE FACTORS 723

different ones are combined with the first ones, the results

may be reversed. Now to study one variable (germ

plasm) through a measure (soma) influenced by a second

variable (environment) will seldom give correct results

if the effect of the second variable is not clearly recog-

nized and discount made for its influence. In the present

case, it appears that the curve of the averages can only

show the degree to which the variations due to environ-

ment and the germinal variations tend to go in the same

direction. That there is a rise in the curve shows that,

on the average, they are a little more likely to agree in

direction than to contradict each other. On the other

hand, since the environmental variations can not be ac-

counted for and eliminated, the curve gives no informa-

tion as to the actual or relative potencies of either set of

variables. That there are no fluctuations in the curves

may have been assumed to prove that environment is con-

stant and therefore does not demand consideration. But

this conclusion can not be safely drawn from the facts.

The curves probably do mean that, when generation is

compared with generation, the variations of the environ-

ment are cancelled out; they mean that these environ-

mental, or extra-germinal, variations occur within a gen-

eration, and probably within a family or within the gonads

of the parents. Environment might well be ignored were

the ultimate question to be answered, ‘‘How much can

selection change the average grade of hooded rats?’’

But this is not the main question. The question to be an-

swered is, ‘‘ What is the nature of the changes in the germ

plasm ?’?

In view of all this, one can find slight justification for

assuming that the germinal variations were small and

constantly occurring. It seems entirely possible that the

environmental, or extra-germinal influences were strong,

perhaps even more effective than the germinal constitu-

tion. In this case, there would be no need to assume a

very large number of factors to find a multiple factor ex-

planation for the slow advance wrought by selection.

724 THE AMERICAN NATURALIST [Vou. L

Such strong environmental influences would, for the most

part, effectively confuse the various combinations of ger-

minal factors, and selection would continue to produce

slight advances for a long time.

5. Castle has explained (Castle and Phillips, :14, p.

24) the significance of the ‘‘. . . observed reduction of

variability” for the multiple factor interpretation; he

stated at the same time that ‘‘. . . extensive modifica-

tion through selection is possible without any marked

falling off in variability.” Since the observed reduction

Neo"

re es. We Sie AER - wW Fie Ts t4 eT

Generations

G. 1. Standard deviations of the plus race (solid line) and minus race

(broken line) in the various generations of selecting. Ordinates represent

standard deviations in terms of grades of hooding; abscisse represent genera-

tions of selecting. Calculations made from the data as given by Castle and

Phillips, 14, in Tables 1-13 and 16-28.

in variability is not considered to be marked, Fig. 1 is pre-

sented to show the facts graphically. The standard devi-

ations plotted in this graph have all been calculated di-

rectly from the data, and in several cases they differ

slightly from those given by Castle. The decrease in

variability that is shown by this figure is the expected re-

sult of reduced heterozygosity accompanying continued

selection.

6. The question of the rate of advance has been an-

No. 600] PIEBALD RATS AND MULTIPLE FACTORS 725

swered by the statement that ‘‘no slowing up is observ-

able in the rate of change of the racial character under

selection either plus or minus ”’ (Castle, :16, p. 96). This

is assuredly a very vital point in the contention that mul-

tiple factors will not explain the results. For, if the rate

of advance has not fallen off, and if, during seventeen

generations, each selection has been as effective as the

preceding one, it certainly would look as though this prog-

ress were due to constantly varying germ plasm, and not

to the sorting out of certain groups of factors. Were a

sorting out of factors going on, each advance would re-

strict the possibilities for further advances, so that in a

series of selections the rate of advance would decline.

In Castle’s ‘‘Heredity,’’ page 122, Fig. 41 are shown

the curves of the averages of the first eight generations

of the plus and minus races. These curves begin with

the average of the offspring that appeared after the first

selection. From this point on, the advance shown by the

curves is gradual. But should not the advance resulting

from the first selection be recorded? The average of the

first selected generation was not the point of departure.

To show the advance resulting from the first selection,

the first point of the curve must give the average of the

hooded race before the first selection. Unquestionably

the difference between the average of the unselected race

and the first selected generation was an advance due to

selection, yet this advance is apparently ignored in the

statement quoted above, as well as in the figure cited.

The first selection resulted in a very much greater ad-

vance than any other single selection in the whole series.

It took the ten subsequent selections to separate the means

of the two races as far as the first selection separated

them. If each selection had produced a like advance, the

eleventh generation of selection should find the averages

of the two races eleven times as far apart as they were

after the first selection instead of twice as far apart.

Failure to consider the advance due to the first selection

has concealed one of the most striking features of the

726 THE AMERICAN NATURALIST [Vou. L

whole series of experiments, namely, that the first selec-

tion brought about an immediate and abrupt establish-

ment of two races with means 3.05 grades apart. The

greatest divergence between the two races due to a single

selection in all the following generations was 0.64 grade.

This followed the third selection. In the second genera-

tion there was a reduction of the average of the plus race.

Castle explains this as follows:

. To obtain larger numbers of offspring, several new pairs were added

to the experiment in this generation which did not appear in Table I

either as offspring or parents, but which were derived from the same

general stock as the parents of generation one (Castle and Phillips, :14,

p. 9).

3.00

2.00

t! 2 2 4 3 6 7 8g bloi BDU I

G enerations

Fic. 2. Rate of divergence of the means of the two selected races in the

various selected PEA ns. The point for the second generation has been

nasa gee the curve; see explanation in text. Ordinates represent grades of

Saci a represent generations of selecting.

No. 600] PIEBALD RATS AND MULTIPLE FACTORS 727

After the third generation there is, in general, a gradual

decline in the effectiveness of selection, till in the fifteenth

generation the advance is 0.12. In support of this state-

ment, which stands in direct disagreement with the quo-

tation at the head of this section, Fig. 2 is offered. In

this figure the ordinates represent the increases in the

differences between the two racial means in successive

generations. Since the decline in the second generation,

as explained, seems to have no immediate significance,

this point has been omitted from the curve. The advance

in the third generation has been calculated from the dif-

ference between the two races in the first generation,

which of course gives a slightly smaller advance than if

the difference in the second generation had been used.

The greater part of the falling off occurs in the plus race,

but both races show the same general tendencies, namely,

a sudden advance as the result of the first selection, with

much reduced advances following subsequent selections.

It may be supposed that there was a greater degree of

heterozygosity in the parents selected to start the minus

race than in those selected to start the plus race. This

might explain the smaller initial advance in the minus

as well as the more prolonged and slower subsequent ad-

vances.

7. Regression, as Castle uses the term, is the difference

between the averages of the selected parents and their

offspring. It is due to the imperfect correlation of two

variables, intra-germinal and extra-germinal differences,

and so, as stated, it forms a gauge of this relationship.

Its amount will be reduced by a reduction in the amount

of variability of either variable. In the later generations

the regression is reduced. We see no reason to suppose

that environment as a whole acts any differently in dif-

ferent generations. Therefore this reduction in the

amount of regression becomes further evidence in sup-

port of the supposition that the germ plasm is more uni-

form, more homozygous in the later generations. |

728 THE AMERICAN NATURALIST [Vou. L

8. As Muller has correctly reasoned, successful return

selections would be expected on the multiple factor view,

supposing the races were still ‘‘heterozygous even after

generations of selection’’ (Muller, :14, p. 571). It may

be added (MacDowell, :15, p. 95) that the failure of a

return selection to reduce the average, as long as the ad-

vance selection was progressing, would be strong evi-

dence against the multiple factor interpretation. As

long as there remained any heterozygosity in the race,

both advance and return selection should succeed in mov-

ing the averages.

9. In no case should return selection lower the averages

at a rate faster than advance selection was raising them

at the same time. That is, a certain degree of heterozy-

gosity will permit a certain rate of advance or decline of

the averages. The return selections that were started

after several advance selections did not show a decline

that could be compared with the sudden advance that oc-

curred in the first selected generation. The plus selec-

tions had reduced the heterozygosity and had thus set

closer limits on the effectiveness of return selection, as

well as of further advance selection. So return selections

from later generations should be less effective than re-

turn selections from the earlier generations. In the plus

race, generation 7, there is a difference of .84 between the

average of the offspring of rats selected to continue the

plus race and the average of the offspring of rats chosen

for a return selection. In generation 12 the correspond-

ing difference is .60. In the minus race the difference

between the averages of offspring from high and low

grade parents in 7, 8 and 9 average .50. In generation 12

the difference is .28. The numbers of rats are very small

in most cases, but it is interesting to note that as far as

they go they seem to show that return selection is less

effective in the later generations than in the earlier ones.

10. Castle has shown that the increase in variability

in the first generation of a cross between the plus and

minus races may be considered an indication of segrega-

No. 600] PIEBALD RATS AND MULTIPLE FACTORS 729

tion, therefore, of heterozygosity in the races (Castle and

Phillips, :14, p.30). Muller ( :14, p. 573) repeats Castle’s

suggestion that the further increase in the variability in

the second generation of this cross argues for the sup-

position that the two races differ in regard to several fac-

tors. This is a logical interpretation uncontested by the

facts, though of course it is not proved. On the other

hand, it has been proved, by the crosses between the two

races, that they are not distinguished by just one Men-

delian unit. Now if the increased variability in the F,

of the cross between the plus and the minus races be due

to heterozygosis in these races, and if selection is reduc-

ing this heterozygosis, crosses made after the races had

been selected for a longer time should give less variabil-

ity in F,. The figures show this to be the case. If there

is less variability in F,, in later generations there should -

correspondingly be less variability in the F,. We find:

F, from cross plus by minus, generations 5 and 6 = S. D. 0.71; generation

F, from cross plus by minus, generations 5 and 6 = 8. D. 1.01; generation

wao

11. The reductions in the averages of both the plus and

minus races after crossing with wild or Irish, first led

Castle to consider a factorial interpretation (Castle and

Phillips, :14, p. 25). Muller (:14, p. 574) has fully re-

stated the bearing of this on the multiple factor theory.

The cross has apparently undone selection to some ex-

tent by restoring some of the factors that had been se-

lected out in forming the two races; the cross has in-

creased the heterozygosity of the extracted hooded rats,

returning plus factors to the minus race and minus fac-

tors to the plus race.

In the light of the above interpretation, the conver-

sion of the minus race into the plus race by means of a

cross is significant. Selection for increase in pigmenta-

tion was started from extracted hooded rats from a cross

of minus with wild. The first generation of this selec-

tion made as sudden an advance as the first generation of

selected plus rats did at the beginnitig of the experiment.

730 THE AMERICAN NATURALIST [Vou. L

It is to be observed that a cross makes a profound dif-

ference in the effectiveness of return selection. Crossing

has so modified the germ plasm that rats from the minus

race immediately, without any gradual return to the ‘‘0’’

grade, repeated the history of the plus race. Further,

plus selection was carried on in this new race. Castle

(Castle and Phillips, :14, p. 21) emphasizes the fact that

this race is free from the objection urged against the

main experiment, namely that the closest inbreeding was

not carried out. Further interest in this closely inbred

race lies in the fact that, although it starts out with a

eurve almost identical with the first generations of the

plus race, the rate of advance falls off faster than it does

in the main plus race. One may suppose that the cross

produced an F, in which some rats had a degree of het-

erozygosity similar to that which existed in the original

unselected stock; a closer inbreeding reduced the hetero-

zygosity more rapidly.

12. The earlier generations of the plus race when

crossed with wild are only slightly reduced in pigmenta-

tion. In Table 43, Castle and Phillips (:14, p. 48) show,

among other things, the averages of hooded grandchil-

dren extracted from a cross with Irish. In comparison

with these are placed averages specified to be of offspring

from the same grade parents and the same generation of

the uncrossed selected race. References to the proper

tables of the uncrossed selected races show practical

agreement with the averages as quoted in this Table.

In Table 42, which gives corresponding results of crosses

with wild, three of the averages of the uncrossed races

are taken from the same generation as the parents

crossed, and three seem to be taken from the follow-

ing generation. It is a matter of importance to have

correct standards for judging the modifications due to

crossing. There might be a question whether one should

use the average of the generation from which the

parents came, or the following one; but in either case

the use should be constant. Although the averages of

No. 600] PIEBALD RATS AND MULTIPLE FACTORS 731

the generations from which the parents came have been

used for comparison in eight of eleven crosses, it ap-

pears to be a more fair procedure to compare the aver-

ages of the offspring produced by parents of the same

grades and generations as used in the crosses. Suppose

the hooded parent crossed was grade 2, from the fourth

generation, then the average of the offspring from par-

ents of grade 2 from generation 4 should be compared

with the hooded offspring in F,. In other words, the

average to be used for comparison would be found in

generation 5. On this basis the comparisons shown in

Table I have been made. Returning now to the state-

ment at the head of this section, that when crossed with

wild, the earlier generations of the plus race are only

slightly reduced in pigmentation, this table shows that,

when third generation parents were crossed, the extracted

hoodeds were lowered .04; when fifth and sixth genera-

tion parents were crossed, the lowering of their hooded

grandchildren was greater, .17; when the hooded parent

came from the tenth generation, the average of the hooded

grandchildren was lowered .76.

TABLE I

CALCULATION MADE FROM DATA FROM CASTLE AND PHILLIPS, :14, TO SHOW

THE EFFECTS OF CROSSING ON THE AVERAGES AND STANDARD DEVIA-

TI F THE EXTRACTED HOODEDS

Numbers in brackets are those given by Castle and Phillips.

A. Comparisons of the Averages of Extracted Hooded Rats with the Aver-

ages of the Offspring of Hooded Rats of the Same Grade and Genera-

tion as the Hooded Rats Used as Parents in the Various Crosses.

Wild by Minus Race

6 1

Generation from which hooded parent came 216 0

Average grade of F, hoodeds ............-- + 31 + .25 + .25

Average grade of uncrossed hoodeds ....... — 1.18 —1.72 — 2.12

Average grade of uncrossed hoodeds as

pabhibad o... ie eren (— 1.20) (— 1.59) (— 2.05)

Raised by orome oo. aaan 1.49 1.97 2.37

: Wild by Plus Race

Generation from which hooded parent came 3 5+6 10

Average grade of F, hoodeds .............- + 2.56 +2.97 +3.15

Average grade of uncrossed hoodeds ....... +2.60 +314 +3291

Average grade of uncrossed hoodeds as

PObUBOR -r r iaia Sees (+ 2.60) (+ 3.14) (+ 3.84)

Lowered by mosi er. o n: .04 :17 .76

ion THE AMERICAN NATURALIST [ Vou. L

Irish by Minus Race

Generation from which hooded parent came b 4

Average grade of F, hoodeds .............. — 62 — Ta — 04

Average grade of uncrossed hoodeds ....... — 1.28 — 1.64 — 1.83

Average grade of uncrossed hoodeds as

PMO e E O T R ee ee ok es (— 1381) (1.18) (—1.75)

Men OY CINE ee ei eee .66 21 .89

Irish by Plus Race

Generation from which hooded parent came 2 3

verage grade of F, hoodeds ............ + 1.27 + .95

Average grade of uncrossed hoodeds ....... +2.10 + 2.60

Average grade of uncrossed hoodeds as

WUBI a ee gs se kee Cale cree ets (+ 1.80) (+ 2.50)

Lowood by eee a A eee as 83 1.65

B. Comparisons of the Standard Deviations of Extracted Hooded Rats with

the Standard Deviations of the Offspring of Hooded Rats of the Same

Grade and Generation as the Hooded Rats used as Parents in the

Various Crosses.

Wild by Minus Race

6 10

Generation from which hooded a came 21%

B: Drot Pi hopdods ins ices eis Fee. creek 1.03 90 1.18

8S. D: of uncrossed hoodeds salsos oein 56 33 ol

S. D. of uncrossed hoodeds as published . (.49) (.44) (.24)

Enereased by rosé i. oe ees. AT .57 87

Wild by Plus Race

Generation from which hooded parent came 3 5+ 6 10

By D. of F; hooded eb ess cee 50 52 45

S. D. of uncrossed hoodeds ............... AT AT .29

S. D. of uncrossed hoodeds as published .... (.53) (.49) (.36)

Ihcronsod DY Grob Borst i ec lee es eels 03 .05 16

Irish by Minus Race

Generation from which hooded parent came 31%

D: D of Fy Rites o eanan aa .64 30 .84

8. D. of uncrossed hoodeds ..............- 53 34 26

S. D. of peat hoodeds as published . (.48) (.46) (.35)

Iperoasod by crosa ik ee te 11 -2 .58

Irish by Plus Race

Generation from which hooded parent came 2 3

oh. of F, koodedi o.. se .90 87

8. D. of unerosred hoodede .2 4.4.2). 0. 44, 38 AT

S. D. of uncrossed hoodeds as published... . (.75) (.53)

Increased by CONN... occu ao ee es 52 40

If the difference between these selected generations lies

in the changed position of the mode of continuous ger-

No. 600] PIEBALD RATS AND MULTIPLE FACTORS 733

minal fluctuations, one would have difficulty in accounting

for the above facts. If these various selected generations

differ in the number of multiple factors they bear, one

can easily understand that the reason that practically no

modification is apparent when the third generation is

crossed, is that the number of plus factors in this genera-

tion and in the wild are not very different; in the fifth

and sixth generations there may be a few more plus fac-

tors than in the wild, and in the tenth generation there

are several more. `

13. The early generations of the plus race, although

only very slightly lowered by crosses with wild, are strik-

ingly lowered by crosses with Irish. In a cross in which

the hooded parent came from the second generation, the

lowering was .83; when the hooded parent crossed came

from the third generation, the lowering was 1.65. Now

how may this fact be interpreted? If the change in the

means following a cross be assumed to be due to the ac-

tion of different numbers of factors in the races crossed,

it is clear that this particular wild is more like the plus

race in regard to its factors than is the particular Irish

race. In other words the wild race seems to have more

plus factors than the Irish race. When early generations

of the plus race are crossed with wild there is hardly any

change in the averages of the F, hoodeds, because there

are about the same plus factors in the wild as in these

early generations of the plus race. When these same

generations of the plus race are crossed with Irish there

is a considerable decrease in the averages because there

are fewer plus factors in the Irish than in the early gen-

erations of the plus race. Now if the germ plasms of the

wild and Irish differ in regard to the number of accessory

factors, and if the germ plasms of the plus and minus

races differ in this same regard, comparisons of all the

crosses between these races should show the following

results: crosses between wild and minus should give

greater modifications in F, than crosses between wild and

plus; crosses between Irish and minus should modify the

734 THE AMERICAN NATURALIST [Vor L

F, hoodeds less than crosses between Irish and plus. More

directly, the plus race should be more modified by the

Trish, the minus race more modified by the wild. Ob-

servation of Table I will show that these results are

realized.

As already noted in the case of crosses between the

plus and wild races, this table shows that in other crosses

the different generations of the selected races are differ-

ently modified. After long selection there is more modi-

fication as the result of crossing. This generalization is

supported by all the averages and all the standard devia-

tions in crosses involving the wild race; it is supported

by all but one average and by all but one standard devia-

tion in crosses involving the Irish race. If selection is

sorting out different groups of factors in the plus and

minus races, crosses made after many selections bring

together groups of factors more diverse than when crosses

are made after only a few selections. The greater the

diversity in the numbers of plus or minus factors in the

animals crossed, the more extensive will be the segrega-

tion in the second generation. More extended segrega-

tion may be expressed by increased variability and by

more pronounced modification of the averages of the F,

hoodeds.

14. The behavior of the ‘‘mutant’’ in crosses with the

plus and minus races gives clear support to the multiple

factor hypothesis. Castle (Castle and Phillips, :14, p.

29) has clearly demonstrated this point. The ‘‘mutant’’

is a suddenly appearing, quantitatively increased stage

of the hooded character, that is controlled by a Men-

delian factor. Crossed with the race from which it

sprang, the extracted individuals show no change from

the uncrossed race, either as to averages or variability;

crossed with the other race, modifications were found,

equalling those obtained when the two races were crossed

together. The newly discovered factor acts independ-

ently of the other factors, is not modified by them, and

does not modify them. . Being the one difference between

No. 600] PIEBALD RATS AND MULTIPLE FACTORS 130

the mutant and the plus race at the time the mutant ap-

peared, this factor affords a critical test for the interpre-

tation of the modifications that result from crosses.

OBJECTION TO THE MULTIPLE FACTOR INTERPRETATION

One new point since 1914 has been urged against the

application of the multiple factor hypothesis to the re-

sults. By the strength of this evidence the authors of

the rat publication are ‘‘forced to conclude that this unit

(hoodedness) itself changes under repeated selection in

the direction of selection ?. (Castle, 45b, p. 722). The

point follows:

The changes effected by selection show permanency under crosses with

wild rats. They change no more nor less than an unselected hooded

race does. A first cross of the selected race seems to show a partial -

undoing of the changes produced by selection, but a second cross made

on a still larger scale, involving over 1,000 second-generation individu-

als, showed no further change of this sort, but instead a return to about

what the selected race would have been had no crossing at all occurred

(Castle, :16, p. 96

If the grade of hooding of the plus race is reduced in

crosses with wild by the replacement of factors selected

out of the plus race, repeated crossing of the modified rats

should produce further reduction. On the basis of the

above claim that crosses do not produce such modifications

in the hooded pattern all the evidence formerly admitted to

favor the multiple factor interpretation has been swept

aside. No one would claim that a single strongly sup-

ported experiment may not upset large amounts of con-

trary evidence, but in such cases it is of utmost impor-

tance to have the validity of the crucial experiment fully

supported. Is the claim that crosses do not change the

selected races fully supported? The following are all

the data we are given on this point:

Extracted hoodeds from

hooded X wild .......75 rats, average 2.89; regression on grandparents .56

eted

Extracted hoodeds from extra :

hooded X wild ........ 263 rats, average 3.33; advance on grandparents .32

736 THE AMERICAN NATURALIST [ Von. L

Averaging the 75 hoodeds may first be criticized. These

include all the extracted hoodeds that came from crosses

between the wild and the plus races. The third, fifth,

sixth and tenth generations of the plus race are involved.

It has been shown that the early generations of the plus

race are not lowered very much by the crosses in com-

parison with the tenth generation, which was consider-

ably modified. Therefore among these 75 extracted

hooded rats are some that were lowered by the crosses,

but more that were practically unmodified. Moreover,

the 263 twice extracted hooded rats came from ancestors

that had been selected for at least ten generations. Only

16 of the 75 once extracted rats had ancestors that had

_ been selected for ten generations; the others, having an-

cestors selected for a shorter time, would be expected to

give lower averages. In testing for further lowering in

this second cross it would seem unjustified to use an aver-

age including rats not lowered by the first cross or rats

that had not been selected for an equal number of gen-

erations before the crosses. Modified by the above con-

siderations the comparisons stand as follows:

Extracted hoodeds from

tonih pon, ERR X Va eerie yo as ee cae oud pee bees average 3.15

Extracted hoodeds from

extracted hoodeds X wild ....... ite E wid A as be Side td average 3.33

Uncrossed, same generation and

grade as hooded grandparent ................... Sais? vaginal ek average 3.84

The conclusion has been quoted that the cross of the

extracted hoodeds with wild has not carried on a further

reduction, but it has shown a return, ‘‘to about what the

selected race would have been had no crossing at all oc-

curred.” Will the above figures support this conclusion?

The cross of the extracted hoodeds with wild does indeed

give a higher F, average than the cross of the tenth gen-

eration, but the difference is only slight (.18). These

two averages are based on very different numbers. It

is entirely possible that a larger number of rats extracted

from the first cross would have had a higher average than

that of the rats extracted from the second cross; in such

No. 600] PIEBALD RATS AND MULTIPLE FACTORS 737

a case the second cross would be said to show further re-

duction.

Whether this advance in the second cross returns the

hooded grade to about what the uncrossed race would

have been is a matter of what average is used to repre-

sent the uncrossed race. The original hooded parents

were the last parents to be selected in this series of

crosses. It seems clear then, as above reasoned in an-

other connection, that the average to be used in compari-

son with the two groups of F, hoodeds is that of the off-

spring of uncrossed parents of the same grade and gen-

eration as the original hooded parents used in the crosses.

If this average be accepted (3.84), it is plain that even

after the second cross there remains a considerable dif-

ference between the averages of the uncrossed and the

twice extracted hooded rats. There is reason to believe

that the changes produced by selection are modified by

crossing and that it has not been finally disproved that

further crossing does not cause further modification. So,

as far as can be judged from the data at hand, this crucial

test does not seem to offer a final blow to the applicability

of the hypothesis of multiple factors.

On the other hand, that modification actually does re-

sult from crosses is strikingly proved by the conversion

of the minus race into the plus by means of a cross.

This experiment has been referred to on page 729. Six

successive return selections did not return the average

of the minus race to the ‘‘0’’ grade. But after minus

race rats were crossed with wild, a single selection of the

plus varieties raised the average 2 grades above ‘‘0.”’

SuMMARY

By way of recapitulation, the points referred to are

summarized as follows:

A. Seventeen generations of selection need not have en-

tirely eliminated modifiers, because,

1. Matings less close than brother and sister have

tended to continue heterozygosity ;

738 THE AMERICAN NATURALIST [Vou. L

2.. Environmental influences may possibly act in such

a way that only occasionally does a selected in-

dividual carry germ plasm more homozygous

then the average.

B. The implied claim that the facts do not support the

supposition that selection has decreased the number

of modifiers, or has reduced the heterozygosity in

the two races of rats, has been answered by the fol-

lowing points:

. Selection reduces the variability.

. The rate of advance declines as selection is con-

tinued.

Parental regression is lowered by selection.

Return selections argue that heterozygosis Is still

present; they indicate that there is less hetero-

zygosis after longer selection, since selection re-

duces the effectiveness of return selections.

Crosses between the plus and minus races strongly

suggest that heterozygosity is still present by the

increase in variability in F,; they also appear to

show that there is less heterozygosity in a later

generation, since the increase in F, is less in a

cross after longer selection.

Crosses between the selected races and the wild or

the Irish race show that more modification ap-

pears in the F, hoodeds when crosses are made

after longer selecting.

| m

saa g

The reader is now in a position to judge whether the

writer is justified in concluding that there is still a ‘‘pos-

sibility that other as yet undiscovered factors might be

responsible for the apparent changes observed’’ (Castle,

:15, p. 722) and that the claim that ‘‘all the evidence we

have thus far obtained indicates that outside modifiers

will not account for the changes observed”’ is too sweep-

ing.

No. 600] PIEBALD RATS AND MULTIPLE FACTORS 739

Discussion!

A great difficulty has been placed on the discussions of

this subject by the different terminology used by those

holding different opinions. Calling the visible character

the Mendelian unit is a striking example of this difficulty.

There is a vital difference between a unit character and

a factor, which must be constantly recognized if this dis-

cussion is to progress.

It is unfortunate that the word selection has come to

have the significance of a slogan. For the nature of the

actual power of selection itself is not in question. What

selection is, can be easily defined and agreed upon. If the

nature of the changes in the germ plasm could be de-

termined, there would be little disagreement as to what

selection could accomplish. Even those who are not con-

sidered to be selectionists believe that natural selection is

very important in evolution. So the epithets, selectionist

and pure-lineist, fail to indicate the difference between

the two groups to which they have been applied. It

would be quite impossible to divide biologists into two

distinct schools on the basis of a subject upon which there

are many different shades of opinion. Any such classi-

fication would be inaccurate, even if the most precise

definitions of the classes were generally accepted. When

there are no accepted definitions, and those most clearly

cut are offered by individuals in the opposite group (each

one realizing the diverse ideas within his own group and

wishing to crystallize an opposing view in order to attack

it) such classification of opinion is far from scientific. In

the present instance, the classification into selectionists

and pure-lineists has tended to magnify the differences

between investigators. With a desire to try to overcome

1 Since the writing of this paper, there have appeared papers by Pearl

(‘*Fecundity in The Domestic Fowl and The Selection Problem,’’ AMER.

Nar., 1916, p. 89) and Castle (‘‘Can Selection cause Genetic Change?’’

AMER. NAT., 1916, p. 248) which have a close relationship to the present

discussion. It has been considered wiser to leave this paper as written

than to enter the controversy by including discussions of the two papers

mentioned.

740 THE AMERICAN NATURALIST [Vou. L

the exaggerated differences which seem to exist, the fol-

lowing discussion is offered. It is written with no wish

to codify or defend the opposing positions, but rather as

an attempt to formulate the issue a little more clearly by

presenting two views, which appear to have advocates, of

the nature of the changes in the germ plasm.

The view to be called the ‘‘first’’ is as follows: The

changes in the germ plasm are in the nature of fluctua-

tions, now larger, now smaller, but continuously appear-

ing; they lead in all directions. This is true of all in-

heritance, whether or not it be factorial (Mendelian) in

basis. If it refers to Mendelian inheritance the potential

grade of the factor in question, as found in any zygote,

acts as a mode about which the fluctuations in potentiality

occurring in the next generation are grounded. In other

words, although a zygote may include the strongest po-

tential grade of a factor that has appeared, the inevitable

fluctuations in this factor that are found in the different

gametes formed by this zygote will include, together with

those like and weaker than the parent, some with stronger

potentialities than the parent.

The view to be called the ‘‘second’’ is as follows: The

changes in the germ plasm are discontinuous; they ap-

pear fortuitously. They may strike out in almost any

direction, as a projectile may be aimed in ‘‘any direc-

tion,’’ in contrast to the ‘‘all directions” taken by the

waves of sound when the projectile explodes.

According to the first view, selection would result in

modification in any direction the breeder might desire.

irrespective of variational tendencies shown by the

animal. To maintain conformity to type would require

as constant selecting as would be required to. obtain di-

vergence. According to the second view, selection could

progress only in certain directions, depending on how the

germ plasm happened to change; the variational tenden-

cies of the animals would probably suggest these direc-

_ tions. Conformity to type would be considered to be a

fundamental phenomenon due to the conservative tend-

No. 600] PIEBALD RATS AND MULTIPLE FACTORS 741

ency of the germ plasm to maintain the status quo. On

the basis of the first view, the external influence (selec-

tion) would have major importance in defining the

course of evolution; on the second, the internal influence

(the inherent nature of the germ plasm itself) would

have major importance. In both cases, the nature of the

progeny would depend on the nature of the germ plasm

of the parents. In both cases selection would be able to

modify the race. But in neither case is the origin of the

changes in the germ plasm explained. The fundamental

causes of evolution are as much a mystery as ever.

Grant a certain hypothesis of germinal changes, and

selection becomes a more important factor in evolution

than when another hypothesis is granted. But even such

an increased importance of selection does not give it the

value of a fundamental creative cause of evolution.

There has appeared a theory that would give selection

still greater importance by saying that selection has the

power to build up unit-factors and induce mutation.

Unit-characters may arise gradually as the result of repeated selec-

tion in a particular direction (Castle, :12b, p. 28

In yellow animals, as in blacks, individuals of iok BE occur

the darkest known as reds, the lightest as creams. A complete series

of intermediates can be obtained if so desired. If we select any two

widely separated stages in this series fairly stable in their breeding

capacity and cross these, they Mendelize, i.e., they behave as if they were

a single unit-character difference between them. ... That difference

might equally well be half as great as it is, or a quarter as great, or a

thousandth part as great. A monohybrid ratio would result equally in

each case, upon crossing the two quantitatively different stages (Castle,

:12a, p. 358). /

Now this may be true for yellow guinea pigs, but the

rats clearly demonstrate that it is not true in all cases.

The two quantitatively different stages of the hooded

pattern represented by the plus and minus races do not

result in a monohybrid ratio when they are crossed.

However there has appeared a ‘‘unit-character’’ dif-

ference in one of these races of hooded rats. It appeared

suddenly, and it Mendelizes when crossed with other

742 THE AMERICAN NATURALIST , [Vou L

hooded rats. The occurrence of this ‘‘mutant’’ is

claimed to have been induced by selection.

It seems to us quite improbable that this plus mutation could have

arisen in the minus selection series. We believe that the repeated se-

lection which was practiced had something to do with inducing this

change in the plus direction (Castle and Phillips, :14, p. 31).

No reason for such a supposition is given. On the

other hand there is clear reason for supposing that such a

mutation would be far more easily detected in the plus

series if it occurred there. The same mutation occurring

in the minus race would perhaps have the same relation.

to that race as it had to the plus race when it occurred

there; since it would lack the extension factors of the

plus race, it would have a very different appearance and

would probably have a grade not far from ‘‘0.’’ It seems

that very few rats of this grade were bred or tested. Had

this mutation occurred in the minus race and been iso-

lated, it would have been possible to obtain it as it ap-

peared in the plus race, by proper crossing.

LITERATURE CITED.

Castle, W. E.

:11. Heredity. Appletons.

:12a. The Inconstaney of Unit-Characters. AMER. Nat., Vol. 46, p.

352.

:12b. Some Peron Principles of Animal Breeding. Amer. Breed.

Mag., Vol. 3, p. 2

15a. Mr. Muller on The Constancy of Mendelian Factors. AMER. NAT.,

Vol. 29,

:15b. Some perinei in Mass Selection. AMER. Nat., Vol. 49,

p. 713.

:16. Is Selection the More Important Agency in Evolution? Sci. E

VOL 2, p. 9L

Oe te W. E. and Phillips, J. C.

Piebald Rats and Selection Pub. Carn. Inst. Wash. No. 195.

eck A. L. and A. C., :1

:14. Studies on Variation ae Selection. Zt. f. ind. Abs. u. Vererbungs-

lehre, Bd. II, Heft 3, pp. 145-183.

MacDowell, E. C.

:15. Bristle Inheritance in Drosophila. Journ. Exp. Zool, Vol. 19,

p. 61.

Muller, H. J.

:14, The Bearing of the Selection Experiments of Castle and Phillips

on the Variability of Genes. Amer. Nart., Vol. 48, p. 567.

SOME FEATURES OF ORNAMENTATION IN THE

KILLIFISHES OR TOOTHED MINNOWS

HENRY W. FOWLER

THE ACADEMY OF NATURAL SCIENCES OF PHILADELPHIA

Tue killifishes, so named by the early Dutch settlers

about New York from their habit of living in the channels

or kills, embrace an interesting family of fishes. They

are known by other names, as top-minnows, cyprinodonts,

toothed minnows, millions fish, ete. Some of these names

are, however, more limited in scope and pertain to sec-

tions or genera. Top-minnow was applied from the habit

of many living at the surface, and cyprinodont, meaning

toothed carp, arose as some greatly resemble very small

carps or true minnows (Cyprinide), though were found

to differ in the presence of teeth in their jaws. Besides

this character are a number of others, in which they agree

with several related families to form the order of pike-

like fishes (Haplomi). Such are all internal and largely

have reference to the bony skeleton. In the abdominal

ventral fins (Procatopus excepted), and without true

spines in the dorsal and anal fins, the order resembles the

herring-like fishes (Isospondyli), but differs in the ab-

sence of a mesacoracoid bone. This latter character is in

agreement with the host of spiny-rayed fishes (Acanthop-

teri), but they usually have the ventral fins well anterior.

Though six families are included in the order of pike-

like fishes, only the mud-minnows (Umbride) and the

pikes (Esocide) occur in the Middle Atlantic States. The

killifishes differ from both in the extremely protractile

premaxillary bones, a condition very easily demonstrated

by examining the upper jaw and prodding its edge for-

ward. In form the body is oblong from elongate and

slender to deep and nearly orbicular. The head is usually

large and robust, often quite chunky. The mouth is small,

with short gape, though wide and terminal. The teeth are

extremely diverse, from broadly incisor-like to finely

743

744 THE AMERICAN NATURALIST [ Vou. L

villiform, and usually occur only in the jaws. The pharyn-

geal bones, unlike those of the true minnows or cy-

prinoids, often have fine teeth, rarely molar, and never

modified or in even numbers as in cyprinoids. The scales

are mostly large, cycloid, adherent, regular and without a

perfected lateral line. The dorsal and anal fins are single,

inserted usually behind the middle of the body, but no adi-

pose fin developed. The caudal is broad and, though

sometimes pointed, not forked. The paired fins are placed

low, and the ventrals abdominal.

Many genera and species, about sixty belonging to the

first and over three hundred to the last, have been de-

scribed. Of these about ten genera and fifty species occur

in the United States. The family reaches its greatest di-

versity in tropical America, and in the Old World the

largest number of forms occur in African fresh waters.

Killifishes live in fresh waters in nearly all situations, in

lakes of great elevation, or in sandy desert streams, pud-

dles and ponds. Others live in tidal waters, or along the

shores of sea-beaches, and all near or close to the surface.

The great changes with age, sex and season render many

of the species difficult of determination. All are of small

size, less than a foot in length.

In nearly all killifishes the sexual differences are well

marked, at least during the spawning or breeding season.

Often the males have enlarged fins, smaller in the females,

as in the may-fish (Fundulus majalis) and the zebra-

fish (Fundulus zebrinus). Still other characters occur in

some species which have been entirely overlooked or

searcely noticed by most writers. These are the minute

spines, or spinules, adorning the scales and fin-rays of

certain species during the spawning-season. Garman, in

his celebrated monograph of the killifishes,1 simply says,

‘fa minor sexual character is that of small spines appear-

ing on the fins of males in several genera in the breeding

time.’’ I have been unable to find any detailed account

of these structures, except casual reference to a few in

descriptions of species. These are usually quite short and

1 Mem. Mus. Comp. Zool., XIX, 1895, p. 11.

No. 600] ORNAMENTATION IN THE KILLIFISHES 745

of but slight value. So far as I have been able to examine

material, these little spines occur only in certain species

of the true killifishes, the pursy-minnows and the four-

eyed fishes, or the Funduline, Cyprinodontine and the

Anablepine, respectively. I have never seen any in the

top-minnows. It is interesting to note that the four-eyed

fishes, creatures with remarkable and extreme modifica-

tions of structure, should be the only group of viviparous

forms in which the spinules have so far been found to oc-

cur. These spinules are different in several ways from

the nuptial tubercles of cyprinoids, in that they are more

permanent, though very minute and inconspicuous. They

may easily be overlooked in preserved examples, owing

to the mucus exuded and covering the scales and fins.

This should be carefully cleaned away, before they can

be detected, and even then only with a good lens. Hach

spinule is found to arise on or close to the edge of the

scale, and not on its exposed surface, as the more dis-

tinctly straight conic tubercles of the cyprinoids. The

spinules are not always perfectly firm and rigid, but may

be flexible or delicate. Those on the anal fin rays are

generally curved slightly and are also often close together,

though not perfectly regular. Their arrangement or de-

sign is usually more or less complete in each species. At

least in one species their development occurs in the young,

as in the ornatus stage of the common mummichog. Prob-

_ably the spinules in most species are not permanent, but dis-

appear after the spawning-season. However, if the spawn-

ing-season for a certain species is protracted, males with

spinules may be found for a period of several months.

Preserved specimens of killifishes do not show scars or

pits like cyprinoids, and it may be that the spinules wear

away as well as drop off. I have not found any examples

with spinules in cold weather, or when spawning was ap-

parently over. In no case have the inner edges of the pec-

toral rays been found with spinules, like the tubercles of

certain cyprinoids. Doubtless such developments are to

be correlated with the spawning habits, as none of the

b ae

k ee an a iad. All

746 THE AMERICAN NATURALIST [Von. L

EXPLANATION OF FIGURES

the figures are drawn to the scale of millimeters and the accompanying

numbers signify such, so that the number of times the line is contained in the

lengthwise diameter of the figure will give its dimensions.

PLAT 1

Fundulus nisorius Cope. Fundulus floripinnis (Cope).

Fundulus zebrinus Jordan and Gilbert. Fundulus stellifer (Jordan).

PLATE 2

Fundulus heteroclitus macrolepidotus Fundulus diaphanus (Le Sueur).

(Walbaum). i Lucania parva (Baird).

Cyprinodon bovinus Baird and Girard.

No. 600] ORNAMENTATION IN THE KILLIFISHES — , T47

PLATE 3

Anableps anableps (Linné). Upper figure female, lower male, and enlarged

scale to left.

killifishes have been seen to clasp the female as Reighard

describes the creek chub (Semotilus atromaculatus). In

the spawning behavior of the mummichog I could not de-

termine if the male in any way secured or held fast to the

female by means of his anal spinules, though possibly

they may be of some such use. Killifishes greatly para-

sitized with sporozoa or myxosporidia have been found,

the adult spawning-males sometimes greatly distorted,

though with the development of the spinules more or less

perfected. Among species of killifishes represented by

spawning-males without spinescent ornamentation which

I have examined are Fundulus punctatus, F. similis, F.

majalis, F. lucie, F. nottii and F. notatus.

In the common killifish or mummichog (Fundulus hete-

roclitus macrolepidotus) of the tidal waters of our Atlan-

tic coast, the male is furnished with little spinules on the

anal rays. They are better developed on the outer or ter-

minal branches of the rays. They are also often irregu-

larly placed, though usually a pair may be found on each

segment, or as a spinule projecting out on each side of the

fin. None of the scales or other fins with spinules.

Spawning-males 76 to 82 mm. long. The female has a

well-developed anal tube extending along the front of the

anal fin for at least half the length of the depressed fin.

748 THE AMERICAN NATURALIST [Von L

My examples 92 to 96 mm. Spawning fishes of this spe-

cies were obtained from April until the middle of August.

In the West African killifish (Fundulus nisorius) the

male has the outer portions of the anal rays covered with -

little spinules. It is also quite likely that the anal fin is

furnished with spinules in the spawning Fundulus ber-

mude.

The barred killifish (Fundulus diaphanus) common in

the fresh waters of the east, from Maine to Carolina, is

quite brilliant in the spawning-season. In the male the

spinules are arranged as little points, like those of the

mummichog, though as the fish is smaller they are less

conspicuous. The scales and fins other than the anal are

without spinules. Spawning males 60 to 70 mm. In the

female a well-developed basal anal sheath extends around

the front of the anal fin. Spawners of this species in full

color were obtained from April until the middle of Au-

st.

The zebra-fish (Fundulus zebrinus) of the Mississippi

Valley region has long been noted for its prickly appear-

ance. Jordan and Evermann state, presumably with

reference to spawning fish, ‘‘ in males the margins of both

dorsal and anal fins are evenly rounded, the anal the

higher, its rays beset with minute white prickles.’? My

examples show it differs from any of the preceding spe-

cies in the male having the sides with the scales minutely

spinescent along their edges. The area of spinescent

scales extends from the head in some examples, in others

for variable distances, back to caudal base, and always

with its greatest development over the base of the anal

fin. On the back the spinules gradually disappear, and

the same is true on the under surface of the caudal pedun- ~

cle. Further, an additional modification is seen in the

presence of spinules on the inner or hind surfaces of the

ventrals, though these fewer than on the anal rays. On

the front of the anal fin the spinules are best developed,

though irregularly distributed on the segments of the fin-

rays, here and there appearing crowded or sparse.

No. 600] ORNAMENTATION IN THE KILLIFISHES 749

Length 48 to 75mm. The female has a broad basal sheath

around the front of the anal fin. Length 51 to 63 mm.

The little green killifish (Fundulus floripinnis) of the

South Platte River basin has the male with the scales

along the middle of the side, especially above the base of

the anal, with minute prickles along their edges. Similar

prickles also occur on the rays of the anal fin, though

with irregular distribution on the segments. They usu-

ally appear better developed along the front anal edge. I

have also seen a few minute prickles above the eyes. In

length these males were 47 to 57 mm. This species be-

longs to the section Zygonectes Agassiz, so called as the

fishes were said to swim in pairs: Doubtless this would

refer to the spawning-habits or when spawning, for at

other times they do not appear to swim in pairs. As in

the brown killifish (Fundulus lucite), another member of

the Zygonectes group, I have never seen them swimming

in pairs, and Ellis claims the same for the little green

killifish.

In the stud-fish (Fundulus stellifer) thë males have

very minute spinules along the anal rays and along the

edges of their scales above the fin. They are also irregu-

larly placed. Length 82 to 99 mm., and the females 73

mm. long have a well-developed basal anal sheath at the

front of the fin. The related Fundulus catenatus shows

similar ornamentation in the male, though my material is

inadequate for detailed comparison.

In the rainwater-fish (Lucania parva) males in high

color, taken in June, differ from any other killifish I have

examined in the presence of minute spinules on the upper

surface of the snout, in some cases even encroaching on

the interorbital space. No other spinules occur. The

muzzle of the male is also modified or decidedly obtuse,

suggestive of the fat-head minnows (Pimephales).

The pursy-minnows (Cyprinodon variegatus) when in

brilliant spawning-dress, in the case of the males, are ex-

tensively provided with minute spinules. These extend

all along the edges of the scales on the head, front predor-

750 THE AMERICAN NATURALIST [Von. L

sal region, posterior sides of trunk or above anal fin, and

front side of caudal peduncle. All the anal rays are also

minutely and finely spinescent, though I have not found

any spinules on the paired fins. Spawning-males 54 to 57

mm., and the females smaller. The related Cyprinodon

bovinus of the southwest is similar. Jordanella floride

is represented only by one small example with spinules,

these very minute along the edges of the scales above the

anal. No spinules found on any of its fins.

In the four-eyed fish (Anableps anableps) of South

America, the males have an intromittent organ dextral or

sinistral. They also have the scales on the trunk, especi-

ally above the anal and on the predorsal region, with

spinules, though more numerous or dense with spinules

in the former space. Top of head, belly and lower sur-

face smooth. Sides of caudal peduncle with a few scat-

tered spinules. A large female, 244 mm. long, is largely

spinescent on the trunk above, though the spinules not so

dense as on scales above the anal in the male. In females

of smaller size, 124 to 128 mm. long, the spinules are

rather obsolete, sparse and scattered, also only on the

back and sides above. Young 27 mm. long still show the

umbilical sac well developed and are scaleless.

Though I have not examined spawning examples of

small-finned killifish (Fundulus parvipinnis), Jordan and

Gilbert state, ‘‘scales large; in the males in spring rough-

ened or ctenoid by small granulations and prickles, similar

to the nuptial excrescences of some Cyprinide ; fins also

rough,”’

SHORTER ARTICLES AND DISCUSSION

FURTHER REMARKS ON THE INHERITANCE OF

CONGENITAL CATARACT

AN article’ by the writers in a previous number of this publi-

cation dealing with the inheritance of congenital cataract in a

statistical way has been rather severely arraigned by Danforth?

in a more recent issue. In our original paper we presented data

taken from genealogical tables published by Harman in the

‘* Treasury of Human Inheritance ’’ which led us to believe—

1. That congenital cataract could no longer be considered as a

single, dominant, unit character.

2. That Davenport should be criticized for making eugenical

recommendations based on the inheritance of cataract as a domi-

nant character when the method of inheritance is not positively

own.

3. That from the evidence at hand cataract could better be

considered as a single, recessive, unit character, reserving final

decision as to this pom until more soni data should become

available.

Danforth believes with us that sie okt cataract can not be

considered as a simple, dominant character. Nevertheless he

tries to defend Davenport from ‘‘ unjust ’’ criticism while he

does not agree with him. It is upon the assumption of cataract

as a positive or dominant character that Davenport bases his

eugenical recommendation as follows:

The usual method of inheritance is that of a positive character. Af-

fected individuals have either half or all of their offspring affected,

while two unaffected parents will probably not have defective offspring.

However, as cataract usually appears late in life it is not always pos-

sible to predict whether the parent will become affected or not.

The eugenie rule is this: If either parent has cataract at least half of

the offspring will have it also. If a person belongs to a strain that has

cataract but is free from it, advice must depend on the nature of the

cataract. If in the family strain cataract appears early, before the age

1 Jones, D. F., and Mason, 8. L., ‘‘ Inheritance of Son pe Cataract,’’

THE AMERICAN Nirgnitrer, 50: 119-126, February, 1

2 Danforth, C. H., ‘‘Inheritance of Congenital Cutapnet, as THE AMERI-

CAN NATURALIST, 50: 442-448, July, 1916.

751

752 THE AMERICAN NATURALIST [ Vou. L

of the ee who contemplates marriage, then such marriage may be

advised; .. .

As regards congenital cataract, then, Davenport advises that

unaffected persons from affected stock can marry without fear

of producing affected children. Harman’s tables show over

thirty matings of unaffected parents having at least one affected

child

No matter how unsatisfactory is the proof that cataract is a

simple recessive, it should be borne in mind that the data given

in Harman’s tables do not stand the test when cataract is con-

sidered as a simple, dominant character. If the argument that

heterozygous individuals sometimes show the recessive character

is to be used to prove the dominance of cataract, it would be nec-

essary to use the assumption to explain thirty-one exceptional

families which have from one to eleven children of which 40 per

cent. of the total are affected. On the recessive hypothesis there

is only one exceptional family so far known to be explained. As

long as it is not a simple dominant character it makes no differ-

ence whether it is a simple or complex recessive or a dominant

governed by multiple factors, the eugenical recommendation

quoted above should not be made, and we still believe that Daven-

port can be justly. criticized.

Danforth objects to the disagreement between the observed

and the expected. results in our table I, giving the progenies of

matings of normal by normal, and compares the goodness of fit

unfavorably with data given by Usher on retinitis pigmentosa.

In our results the disagreement lies in an excess of the actual

' number of affected children over the expected number. If, as

Danforth says, ‘‘ a certain number of congenital cataracts are

produced by intrauterine poisoning without necessarily any ref-

erence to heredity ’’ the tendency would be to raise the actual

number of affected children above the expected. Also any cases

of origin de novo, to which he believes we did not give enough

consideration, would tend to have the same effect. Moreover, it

should be noticed that Usher has over twice as many individuals

to base his ratio upon, 320 as compared to 153 in our case.

Danforth states two main conditions which he says our as-

sumption of cataract as a recessive character does not meet. The

first is the low probability of an individual carrying the abnor-

3 Davenport, C. B., ‘Heredity in Relation to Eugenics.’’ Henry Holt

and Co., New York, 191, pp. 111-112.

No.600] SHORTER ARTICLES AND DISCUSSION 753

mality in a haploid or a diploid state meeting with a heterozy-

gous normal, in random mating, which would be necessary to

produce affected children. According to Danforth’s calcula-

tions the probability of heterozygotes in the general population

is one in thirty. He shows that Harman’s tables give in some

eases as high as eight out of nine individuals mating with nor-

mals and producing affected children, thereby showing that the

normals are heterozygous on the recessive hypothesis. In the

previous publication we did not give consideration to this point

which is of noteworthy significance and we are indebted to the

writer for calling our attention to it.

This apparently high proportion of heterozygotes in the gen-

eral population would be a serious objection to our simple re-

cessive hypothesis if it were not for the fact that there is a con-

siderable amount of consanguinity recorded in the pedigrees

given by Harman. With each of the pedigrees including from

one to many families there is a definite statement as to whether

a record was made and, if so, whether or not consanguinity was

present. Tabulating these statements shows that in sixty of the

pedigrees no record was made. In twenty-four no consanguinity

and in eleyen consanguinity was definitely recorded. Then in

those cases in which a record was made nearly 50 per cent. of

the pedigrees show more or less intermarrýing. Altogether

there are seventeen cousin marriages.

‘With this amount of intermarrying among affected stocks the

proportion of heterozygous individuals carrying the abnormality

in a simplex condition would be greatly increased over the pro-

portion in the general population, and Danforth’s most serious —

objection to our hypothesis loses its force. Evidently no con-

sideration was given to this point when he says ‘“‘ a more strik-

ing refutation of the assumption could hardly be found ”’ (p.

447)

With regard to the second main condition which is raised

against the assumption of cataract as a single, recessive, unit

character Danforth seems to be partly in error, if we understan

his statement correctly. He states: “‘ If congenital cataract

were recessive the normal children of a cataractous parent should

themselves produce affected children in half as many cases as do

their cataractous sibs, and the total number of affected children

produced should be one half as great in the first case as in the

second ’’ (p. 446). Since on the recessive hypothesis only het-

erozygous normal and homozygous abnormal sibs are produced

754 THE AMERICAN NATURALIST [Vou. L

in equal numbers from matings of Nn X nn included in category

B, and neither can produce affected children in turn, unless

mated to a heterozygous normal or a recessive, we do not see

why the normals should produce affected children in half as

many cases as their cataractous sibs. If the chances for obtain-

ing such mates were the same the number of matings which pro-

duce affected children should be approximately the same.*

The second part of the quotation is correct only when both

normal and abnormal F, individuals have an equal chance to

mate with individuals who are either affected or carry the ab-

normality in a recessive condition. The chances of the two —

classes mating with such individuals are probably not equal be-

cause individuals affected with cataract would have a harder

time to find a mate than their normal brothers or sisters and

there would be a greater tendency towards consanguineous mar-

riages and consequently a greater chance of mating with cata-

ractous individuals. The frequent intermarrying among affected

stocks is well known with other abnormalities. Hence normal per-

sons carrying the affectation in a heterozygous condition would

be more likely to marry into unrelated stocks with a far less pro-

portion of heterozygous individuals than would their affected

sibs. If this is true then the expectation of the number of mat-

ings of affected F, individuals giving affected children in turn

4 In answer to a letter sent to Dr. Danforth asking about the above point

the following was received which shows that we did not understand his

meaning correctly:

Por reply to your letter of September 29, I do not say in the paragraph

ich you refer that of the bares of cataractous parents half as

many normal as cataractous individuals should produce affected offspring —

but, on the contrary, that the normals, HE as a group, should produce

affected children ‘‘in half as many cases’’ (i. e. at half as many births)

the cataractous. The families of numerous individuals in both

normal parents as in families with one cataractous parent. The remainder

of the paragraph to which you refer and the statements in your letter

show, I think, that we are in complete agreement as to the theoretical ex-

pectations; it was my unfortunate use of the word ‘‘cases’’ which no

doubt caused you to raise the question. I meant it to refer to F, you

doubtless suspected it might refer to F.. This suspicion was undoubtedly

strengthened by my ‘‘relation of one to two’’ which is one of those slips

of the pen for which I have no means of accounting. Of course it should

have been ‘‘one to one’’ and that fact was uppermost in my mind at the

time of writing the passage!

No. 600] SHORTER ARTICLES AND DISCUSSION 755

would be greater than that of the unaffected and the total num-

ber of children would therefore be more than twice as great.

Danforth, however, after raising this condition does not deter-

mine the number of affected children from the affected and un-

affected F, individuals, but calculates the percentages of these

two classes of parents which produce at least one affected child.

He finds that eighty-six per cent. of the cataractous children of a

cataractous parent themselves produce some affected children and

thus presumably have mated with heterozygous normals. Of

the normal children from the same F, generation only ten per

cent. produce affected children. If the chances for securing

similar mates were the same these percentages should be approxi-

mately equal. The relation of ten to eighty-six which does not

conform to a one to two ratio as Danforth states that it should

necessarily deviates still more widely from a one to one ratio.

There are two reasons why this deviation from a one to one

ratio can be expected in favor of a larger number of affected

matings giving affected children than of unaffected matings.

The first lies in the fact that matings of affected by heterozygous

normals should give a one to one ratio of affected and normal

children, whereas the matings of heterozygote by heterozygote

should give a ratio of one to three. As was emphasized in our

previous publication the only criterion by which it can be deter-

mined whether the mates to the two kinds of F, individuals are

heterozygotes or homozygotes is the production of at least one

affected child. In families with a small number of children the

matings which promise a one to one ratio would have a greater

chance of producing at least one affected child than matings

which promise a ratio of one affected to three unaffected chil-

dren. Hence more of the families of the latter than of the

former class would be omitted from the data.

The other reason why the deviation that Danforth obtains can

be expected is that which has already been mentioned, namely,

that affected individuals are more likely to marry related indi-

viduals because of the greater difficulty of obtaining a mate than

the unaffected would have. The proportion of heterozygotes in

affected strains would be far higher than in the general popu-

lation, so that the chances of the two kinds of F, individuals mat-

ing with a heterozygous normal would not be equal as irs whee

considers them to be.

It is recognized that these arguments are extremely indefinite

and that it is difficult to determine just how much value to give

756 THE AMERICAN NATURALIST [ Vou. L

them. They are however hardly necessary since the numbers

ninety-six and forty-seven upon which Danforth bases his criti-

cism are too small to make a really critical comparison.

Since the number of affected F, individuals which should give

one half affected children exceeds the number of unaffected F,

individuals which should have only one third affected children,

the actual number of affected children in the two kinds of F,

populations would deviate proportionally farther from a ratio

of one to two. If it is conceded that the chances for the two

kinds of matings are not equal, then this deviation would be -

expected.

` The three cases in category C which we gave as matings of

abnormal by abnormal which theoretically should give only ab-

normal children according to the simple recessive hypothesis can

be found in Harman’s tables in the ‘‘Treasury of Human In-

heritance °% as follows: Table 309, Parents I, 1 and 2—Children

II, 1 to 5; Table 312, Parents II, 3 and 5—Children IHI, 3 to 4;

and Table 342, Parents III, 28 and 37—Children IV, 60 to 66.

Danforth says that he can find only two of these. They are

probably 309 and 342. The one which occurs in Table 312

should not have been used without an explanation. Although

the chart indicates that both parents are affected as well as their

two children, the description of the family shows that the exact

parentage is somewhat in doubt. It was an error on our part-

not to mention this fact.

. With regard to the family in 342 in which part of the children

are normal where only abnormals are expected, Danforth does

not accept our explanation that heterozygotes sometimes have the

recessive character. This is quite frequently shown in other

material. His refusal to accept this explanation to account for

the one exception to the recessive hypothesis is shown in the fol-

lowing quotation: ‘‘a single bona fide case in which two affected

individuals produce normal offspring is sufficient to overthrow

it” (the recessive hypothesis) (p. 447). We can not under-

stand his refusal to accept this explanation to account for one

exception when he is willing to use it to explain thirty-one excep-

tions to the dominant hypothesis! This is evident from the fol-

lowing quotation previously alluded to:

Again, since Jones and Mason elsewhere in the same paper (p. 124)

š Harman, N. B., ‘‘Treasury of Human Inheritance,’’ Eugenie Labora-

tory Memoirs, XI, Part 4, Section XIIIa, pp. 126-169, Pl. XXVII-

XXXITI, Dulau and Co., London, 1910.

No. 600] SHORTER ARTICLES AND DISCUSSION 757

use the same argument that “ heterozygous individuals sometimes show

the recessive character,’ we might, if necessary, use the same argument

cataract is dominant instead of recessive it might be maintained that in

those cases where both parents of affected individuals seem to be normal,

one of them is, after all, heterozygous—and affected children are there-

fore to be expected (p. 444).

Perhaps Danforth would be willing to consider another expla-

nation which he suggests, that somatic cataracts of a congenital

origin are not uncommon. If one of the parents in question had

a somatic cataract the appearance of normal children would be

- expected but not of affected children unless the parent was also

heterozygous for hereditary cataract. A probability which would

be rather remote but not impossible.

From the data as they have been gathered up to this time it

seems impossible to arrive at an explanation of the mode of in-

heritance of cataract which will be entirely satisfactory. While

more proof is awaited, we believe that the assumption of con-

genital cataract as a single, recessive, unit character has the best

support from the facts at hand. The article by Danforth has

brought out several important considerations which we neglected.

It is regretted that in this paper which at first sight makes out a

strong case against our recessive hypothesis there is nothing

offered towards a different solution of thè problem.

D. F. JoNES

S. L. Mason

BUSSEY INSTITUTION,

HARVARD UNIVERSITY

THE STATUS OF FOWLER’S TOAD, BUFO FOWLERI

PUTNAM

S. P. FowLFR, of Danvers, Essex County, Massachusetts, ap-

‘pears to have been the first to recognize the fact that this toad

differed in many respects from the common toad. In a letter’

to Prof. F. W. Putnam, Fowler gave a very accurate and com-

plete account of the song and habits of this toad as he had ob-

served it around Danvers.

Cope (see loc. cit.) discussed in much detail Bufo Gohan

fowleri (Putnam). Little was known of this toad at the time

Cope wrote. In fact, Cope stated that it was confined to a few

1Cope, E. D., ‘‘The Batrachia of North America,’’ Bull. 34, U. 8.

National Museum, 1889, pp. 279-281.

758 THE AMERICAN NATURALIST [Vou. L

ponds in northeastern Massachusetts, near the town of Danvers.

He says:

Such a limited distribution for a land vertebrate is remarkable, as is

also the fact of its having so long remained without introduction to

science.

Cope’s work was published in 1889, in the same year that

Allen’ reports having heard Fowler’s toad in New Hampshire.

Speaking of Bufo americanus Le Conte, Allen said:

After the breeding season, the toad’s song changes from a prolonged

pipe to a shorter, lower-toned note, that, at night, has a peculiar weird-

ness, and almost reaches a wail.

Although Allen thought that the common toad was respon-

sible for the two songs, it is plain that he had heard the unmis-

takable song of Fowler’s toad. Allen’s observation extended the

range of this toad well up into New Hampshire.

Although as late as 1889 Fowler’s toad appeared to have a

very local distribution in New England, more recent work has

shown that this toad has an extended range southward.

In 1907 the writer? published a paper showing that Fowler’s

toad is very common around Oxford and Worcester, in Worcester

County, Massachusetts. In a second paper, published in 1908,*

it was shown that the range of this toad extended through Wash-

ington, D. C., and Chapel Hill, North Carolina, into northern

Georgia, where it appeared to be the only common form in the

vicinity of Hoschton and Thompson’s Mills, near Gainesville.

In 1910 Miller and Chapin® gave an excellent discussion of the

range of Bufo americanus and Bufo fowleri in New Jersey and

adjacent regions of New York.

From the observations of Miller and Chapin it appears that

Fowler’s toad occupies practically the entire state of New J ersey,

except, perhaps, the extreme northwestern part. Throughout

2 Allen, Glover M., ‘‘ Notes on the Reptiles and Amphibians of Intervale,

New Hampshire,’’ Proe. of the Boston Society of Nat. History, Vol. 29,

No. 3, 1889, p. 71.

3 Allard, H. A., ‘‘Fowler’s Toad, Bufo fowleri Putnam,’’ Science, N. 8.,

Vol 26, No. 664, Sept. 20, 1907, pp. 383-384.

4 Allard, H. A., ‘‘ Bufo fowleri in Northern Georgia,’’ Science, N. 8.,

Vol. 28, No. 723, Nav. 6, 1908, pp. 655-656.

5 Miller, W. De W., ana Chapin, James, ‘‘ The Toads of the Northeastern

United States,’’ Bidhaa, N. B. Vol. 32, No. 818, Sept. 2, 1910, pp. 315-317.

No. 600] SHORTER ARTICLES AND DISCUSSION 759

central and southern New Jersey it is the only species, as B.

americanus was not found here. Miller and Chapin also found

that Fowler’s toad was the only form to be found upon Staten

Island, N. Y., as well as upon Long Island. In the mountainous

parts of northern New Jersey both B. americanus and B. fowleri

occur.

In 1914 Overton? published an interesting paper concerning

the frogs and toads of Long Island. Overton found that Bufo

fowleri is the only toad occurring on Long Island, where it ap-

pears to be common, while the common toad of the mainland of

New York State is B. americanus.

Various authors have mentioned the song of Bufo fowleri.

S. P. Fowler in the letter to Professor S. W. Putnam, previously

cited, first described its song. His description is particularly

apt.

To my ears the croak is a sharp, disagreeable, unearthly screech, diffi-

cult to describe, as it is unlike any sound I have ever heard. A chorus

of these has been likened to the whoop of a party of Indians.

As none of us at this late day can recall the whoop of Indians,

this comparison, although historically interesting, does not give

us much aid in appreciating the peculiar nature of the sound.

Dr. Nichols, in the same letter, is cited as considering the song

to be a shrill monotone in a high falsetto voice, longer and more

trilling than the voice of Pickering’s hyla. Fowler, however,

states that there is no trill to the note, an opinion the writer also

shares.

The writer has described the note as follows: ‘‘I have heard

nothing in nature so weird and unearthly as the almost agonized

wail of this toad, repeated at intervals,’ and ‘‘The usual note

of Fowler’s toad is a brief, penetrating, droning scream.’’*

Miller and Chapin, in their article previously cited, say of it:

. . . it certainly has much less musie to it than the trill of the Ameri-

ean toad. The notes are more closely connected, so that a sort of

buzzing is heard.

Miss Dickerson? says of the notes of Bufo fowleri:

6 Overton, Frank, ‘‘The Frogs and Toads,’’ Long Island Fauna and

Flora, III. In the Museum of the Brooklyn Institute of Arts and Sciences,

Science Bulletin; Vol. 2, No. 3, Nov. 3, 191

7 Science, N. 8., Vol. 26, No. 664, Sept. 20, “1907.

8 Science, N. S., Vol. 28, No. 723, Nov. 6, 1908.

9 Dickerson, Seep C., ‘‘The Frog Book,’’ 1906.

760 THE AMERICAN NATURALIST [Vou. L

The call of the Fowler’s toad is a metallic, droning sound, not conspicu-

ously vibrated. The pitch of the call may be as high as that of Bufo

americanus, but descends in doleful fashion through several intervals

before the close. Its carrying power is unusually great. The quality

is indescribable; on the whole, the call is weird and mournful and not

especially agreeable to our ears.

Overton (previously cited) says:

Its song is a combination of a low whistle and a moan, and the two

sounds do not melt into a chord. The combined sound is discordant

and decidedly unpleasant to a musical ear, but at a distance the sound

is more pleasant for the moan is not apparent and only the whistle is

heard. The sound lasts from two to three seconds and may be repeated

at intervals of about ten seconds.

Overton says the song of Bufo americanus is prolonged about

thirty seconds.

‘Dr. Andrew Nichols,*® of Danvers, Massachusetts, is quoted as

saying:

There is no sound in bog, pond, fen, forest, or air at all like it.

Although Nichols referred to the toad as Bufo lentiginosus Shaw,

it is extremely cae that he had in mind Bufo fowleri.

Miss Hinckley™

The bleat of B. fowleri, with its far reaching, metallie ring, is usually

heard after sunset. I have seen the latter give voice on the land, while

the trill of B. americanus, heard at all times of day and night during

the mating season, I have only seen given in the water.

In the field the writer has found little difficulty in recognizing

Fowler’s toad throughout its range. Its note at once distin-

guishes it from B. americanus. Color characters, while fairly

definite, do not, perhaps, always serve to distinguish B. fowleri

from B. americanus. According to Miller and Chapin, the color

of the eye alone will distinguish B. fowleri from B. americanus.

These observers state that in the former the iris is silvery, while

in the latter it is bronze. There is some question in the writer’s

mind as to the value of this character as an identification mark.

The question is now under investigation.

10 Nichols, Andrew, Proc. of the Boston Soc. of Nat. History, Vol. 1,

Aug. 2, 1843, p. 136.

11 Hinckley, Mary C., ‘‘On Some Differences in the Mouth Structure of

Tadpoles of the Anourous Batrachians Found in Milton, Mass.,’’ Proc. of

the Boston Soc. of Nat. Hist., Vol. 21, 1882, pp. 307-314, .

No. 600] + SHORTER ARTICLES AND DISCUSSION 761

Miss Dickerson states that the eggs of Bufo fowleri are often

arranged in double rows, but that, so far as known, the eggs of

B. americanus are always laid in single strings. If these char-

acteristics hold true for the two toads it would appear that the

toad with which Gage’ worked was Bufo fowleri, rather than

Bufo lentiginosus americanus. -Speaking of the toads with

which he worked, Gage states that they lay their eggs from the

middle of April until the middle of June, and that the eggs were

laid in two strings, one from each oviduct. The lateness of the

tags season adds to the probability that Gage worked with

B. fowleri rather than with B. americanus.

From the observations of various observers, it is evident that

Bufo fowleri is a widely distributed toad and is extremely

abundant in many places from New Hampshire, throughout New

Jersey, the District of Columbia, southward at least as far as

Gwinnett, Jackson and Hall Counties in northern Georgia.

Cope (previously cited) records a specimen of this toad from

New Harmony, Posey County, Indiana. He also states that a

specimen of the variety B. lentiginosus var. americanus from

Nebraska approximates so nearly B. fowleri, that the latter can

not be regarded as under all circumstances separate and specific

in its rank.

Miller and Chapin have found that toads taken on the Pali-

sades and on the northern end of Manhattan Island sometimes

show forms intermediate between B. americanus and B. fowleri.

These observers have suggested that such intermediate forms may

represent hybrids, but, as they state, it is a question for experi-

mental study.

For a long time the writer has had in mind the question of

experimental hybridization between typical forms of B. fowleri

and B. americanus. It would be of considerable interest to de-

termine whether or not these two toads can be hybridized.

Although B. fowleri is more sensitive to lower temperatures than

B. americanus, and lays its eggs later in the season, it should

not be especially difficult to provide conditions that would bring

the mating season of the two toads together under temperature

conditions required by B. fowleri. It is very probable that the

hibernation period of B. americanus could be prolonged by arti-

ficial refrigeration until the mating and egg-laying period of B.

12 Gage, S. H., ‘‘ Hibernation, Transformation and Growth of the Com-

mon Toad (Bufo lentiginosus americanus),”? Ithaca, N. Y., Proc. of Amer.

Assoc. for the Advancement of Science, 47: 1898.

762 THE AMERICAN NATURALIST [ Von. L

fowleri had arrived. If experimental hybrids could be obtained,

it would be especially interesting to compare the voices of the

hybrids with the voices of the parents, as well as to determine

the hereditary behavior of various other characters.

n those localities where both toads are found, differences in

behavior peculiar to each species tend to prevent natural cross

mating. Bufo americanus is the first toad to appear and, at

least around Oxford, Massachusetts, has completed egg-laying

and left the water long before B. fowleri has appeared. Fur-

thermore, the preference that B. fowleri shows for certain ponds

from year to year is rather remarkable.

Fowler (letter previously cited) noted that only certain ponds

around Danvers,- Massachusetts, were visited by B. fowleri. In

the region of the writer’s early home, Oxford, Massachusetts,

the same rigid preference was shown for certain bodies of water

during the mating season. Here it was indicated that these

toads traveled very long distances to reach a certain quiet bend

in the Maanixit River. Although other permanent bodies of

water were near, these, for some reason, were never visited by

these toads.

The writer hopes that an interest in our common toads will

finally lead some one to investigate the possibility of experi-

mental hybridization between B. americanus and B. fowleri, and

the question of the relationship of these toads. Batrachian hy-

bridization seems never to have been undertaken. It would ap-

pear that such investigations would throw much light on the

question of geographic variation, intergrading forms, ete. Few

creatures are more companionable and harmless in their behavior

and more useful to the agriculturist as insect destroyers, than

the toads. Knowledge of their habits, relationship, ete., is not

only of scientific, but also of soundly practical interest.

ADDITIONAL REFERENCES IN THE LITERATURE TO

FOWLER’S TOAD

Holbrook, J. E. North American Herpetology, Vol. 5, 1842. Speaking of

Bufo lentiginosus Shaw, he says the males seek the females in the

month of May when hundreds may be seen together in some stagnant

pool depositing their eggs. Of the notes he says: ‘‘The males at this

season are extremely noisy, though at other times they are silent, or

make only a slight chirp when taken ’’ (p. 9).

Gorman, Samuel. The North American Reptiles and re Bull.

Essex Inst., Vol. 16, 1884. On page 42 he says of B. fowleri Putnam:

‘‘ This is an americanus of moderate size and with ribs ridges low,

No. 600] SHORTER ARTICLES AND. DISCUSSION 763

close together, and nearly or Gute parallel. Voice peculiar. Manitoba

to Winnipeg; Massachusetts. ’’

Cope, E. D. Check List of North American Batrachia and Reptilia with a

Systematic List of the Higher Groups and an Essay on Geographical

Eea meg on the Specimens contained in the U. 8. Nat.

Muse 1875. P. 29, B. lentiginosus, subspecies fowleri is r

a inti bavon from Massachusetts to Lake Winnipeg.

fowleri (page 86) as confined mostly to the Canadian District i ‘the

Eastern Region.

Hay, O. P. The Batrachians and Reptiles of Indiana, 17th Ann. Rept. of

Sree of Geol. and Nat. Resources of Indiana, 1891. B. fowleri con-

ered a variety of B. lentiginosus. Range given as Danvers, Mass.,

ii the fact that Cope reported a specimen from New Harmony,

9).

Sherwood, W. L. The Frogs and Toads Found in the Vicinity of New

York City. Abst. No. 10 of the Proc. of the Linn. Soc. of N. Y. for

year ending 1898. Mentions B. fowleri as a subspecies of the common

toad, stating that it was confined to northeastern Massachusetts.

Jordan, David Starr. A Manual of the Vertebrate Animals of the Northern

United States. 1899. On page 182, B. fowleri is mentioned as a

variety of B. lentiginosus.

Ditmars, Raymond L. The Batrachians of N Vicinity of New York City.

The American Museum Journal, Vol. 5, 1905. Speaking of the com-

mon toad, he says there are four ile BA, one of which occurs only in

northeastern Massachusetts.

Fowler, H. W. A Supplementary Account of New Jersey Aeolian and

Reptiles. Rept. of New Jersey State Museum, Part III, 1911. Bufo

fowleri is mentioned.

Hancock, J. L. The Toad’s Social Life in Nature. Sketches in Temperate

America, 1911. Fowler’s toad is briefly mentioned and illustrations

are shown.

Surface, H. A. First Report on the Economic Features of the Amphibians

of Pennsylvania. Zoological Bulletin of the Div. of Zoology, Penn-

sylvania Dept. of Agriculture, Vol. III, Nos. 3 and 4, May-July, 1913.

On page 114, B. fowleri is discussed. Statement made that it has been

recorded from New England and New York.

H. A. ALLARD

WASHINGTON, D. C.,

May, 1916

INDEX |

NAMES OF CONTRIBUTORS ARE PRINTED IN SMALL CAPITALS.

Aleoholized Mammals, Hereditary

Transmission of Degeneracy and

Deformities by the Descendants

LES R. STOCKARD and

GEORGE PAPANICOLAOU, 65, 144

Alge, ed, Li Histories in,

BRADLEY Moore Davis,

LARD, H. A., The Status er Fow-

ler’s Toad, 757

Angiosperms, Hybridism and the

Rate of Evolution in, Epwarp C.

JEFFREY, 12

Baco, Hatsey J., Individual Dif-

agri "i Family Resemblances

in Animal Behav

BO Pe ai virgatum,

attached to a Fish, Diodon hys-

with especial reference to

thera, 51

Beans, Inheritance of Eye Pattern

in and its Relation = Type of

Vin RA OTT

pee The Pla ia Ge pharm ong

fe:

Banai perei et abasic Dif-

neti

> Mice of the Color Varieties

‘ °“ Black and- Tan” ? and ‘‘Red,’’

L o.

Dunn, 664

aye and Mendelian Inheri-

W. E. CASTLE, 321 —

a , Distribution of, with WE

Rôle played by

he aoe? Response to Soil Tem-

perature and Soil Moisture, W. A.

CANNON, 435.

ALKINS, Gary N., General Biology

of the Protozoan Life oes 257

ee hk À., er tribution of the

reference to

tha Role Pa by the Root Re-

sponse to Soil Temperature and

Soil Moisture, 435

CASTLE, W. E., Degree’ under

ia and Crossbreeding,

; Can Selection “sal gn

endelian Inheri- ri-

nce, ree Tables of Linkage In-

chats BT 5

Cataract, Congenital Inheritance of,

D. JO

ES and S D MASON.

119, 751; The onsen of Con-

genital, C. H. DANFORTH,

ae scx fat of the, E. A. MIN-

CHIN, 5, 106, 271

Character, Eis and Phylo-

eer Evolution of tie iig T. WATER-

23T:

P oara some Studies e the’ p oak

Drosophilide, CHARLES W, METZ

587

neat Pink-eyed Bien

Occurrence of

LITTLE, 335; rt

“t Black-and-Tan’? and ‘‘Red’’

Mice, ae Genetic Behavior of,

L. C. DUNN, 664

Correlation, and Contingency, An

tline of Current Progress in the

Theory . ARTHUR HARRIS

Variation koa tahkorntante of Fa

tility in the Mammals, J. ARTHUR

Cows, Ka Registry, Statistical

TENE for Age of, C. W. HOLD-

AWAY,

Crossbrecding a and Inbreeding, Vari

ability u nee E. CASTLE, 178

Raag ore” Mechanism of,

HERMA J. Minin. 193, 284,

350, 42

W. J., On a Barnacle, Con-

choderma virgatum, attached to a

Fish, Diodon hystrix, 636

DANFORTH, O. as The Inheritance of

VENPORT AS. Form of

Evolutionary Theory that Modern

765

766

Genetical Research seems to favor, ,

Davis, BRADLEY Moore, Life His-

tories in the Red Alge, 502; Œno-

thera _neo-Lamare kiana, Hybrid of

L. franciscana Bartlett X O. bi-

ennis Linnæus, 688

DETLEFSON, J. Å., Pink-eyed White

Mice, carrying ‘the Color Factor,

Diptera, Chromosome Studies on the,

III, Drosophilide, CHARLES W.

MET

DUNN, E he Genetic Behavior

CoTr

of Mice of the

Color Varieties

‘* Black-and-Tan’’ t: Red,’

664

and

Egg ae ope and Selection, H. D.

GOODALE, 479

Tiisa R. A. The se ig aren of

Lin nkage Intensities, 4

a rt Hybrids of the Genus, R.

LDEN, 243

foaia of the Cell, E. A. MIN-

Ohin, T. E TERMAN, 237;

Various Plant Types, Comparative

Rapidity of, EDMUND W. SINNOTT

466

ree Theo ory, The Form that

Moder

to favor, AS. RT,

ignificance e Osmotic

ll il ge of the Blood, Guana

Scorr,

EWING, Trifolium ovine

quinquefolium, 370

Family Resemblances and Individ-

al Behav-

ior, HALSEY J. Bage, 222

Fasciation in Maize Kernels, T. K.

Ww 306

LFE,

ET Dispersal, FRANK E. LUTZ,

Fecundity in the Domestic Fowl and

ag Pee age Problem, RAYMOND

Fertility, Inheritance of, Variation

and elat

Correlation, in the Ma mmals,

J . ART Harris, 62

OWLER, HENRY W., S Features

- Ornamentation in the Kili-

es or To

in Plants, L. B.

ALTON, 498

THE AMERICAN NATURALIST

[ Vou. L

Gates, R. RuceLes, Huxley as a

COAN rete

Theory that “it seems to fav

DAVENP apr

havior of Mice of the Color Va-

rieties ‘* Black-a

‘‘ Red, ; N

GOLDSCH Ric Experi-

mental “Intersexuality a and the Be

Problem,

ODALE, Egg Production

and Selection, 479 i

HARRI Si ARTHUR, An Outline of

t Progress in the Theory of

Correlation and scien Ti 53;

tance o rtility the

mals, 626; Btátittical Studies ae

the Number of ipples in the

Mammals, 696

Hibernation, Theories of, ANDREW

ASMUSSEN,

Honpaway ,0:W. e a

ing for "Age of Advan d Reg-

gahi Cows, 6

R., ’ Hybrids of the Genus

Epilobiun, 243

Te ey as a anu R. RuG-

S GATES

Hybridiein an and net Rate of Evolu-

maie in hr pe iar EDWARD C

JEF

keeta of the Genus Epilobium, R.

HOLDEN, 243

| Inbreeding and a payin Vari-

ability under, W. E. ÇASTLE, 178

Individual Differences om: Family

esemblances in Animal Behavior,

ALSE G, 222

Inheritance, of Congenital Cataract,

D. F. JoNEs and 8, MASon,

Grape, W. AL’

554; of Eye Pattern in Beans pe}

"Relation to of Vine,

K M. SURFACE, 577; of Fer-

tility, Variation and Correlation

No. 600]

in the Mammals, J. ARTHUR HAR-

RI |

Intersexuality, Experimental, and

the Sex a a RICHARD GOLD-

SCHMIDT,

JEFFREY, EDWARD C., Hybridism and

the Rate 2 Evolution in Angio-

spren k

JONES, D. P. and S. L. Mason, In-

r of Congenital Cataract,

119, 751

KROEBER, A. L., e Cause of the

Belief” in Use gi rol eet 367

mA Cycle, Protozoan, a! of

, Gary N. CALKINS, 257; His

‘ile in oe Red Alge, BRADLEY

Moore Davis, 502

Linkage Intensities, Caleulation of,

R. E "41 1; Tables of,

W. = Ce 575

LITTL The Occurrence of

Three Rasogniend Color Mutations

in Mice, 335

Lutz, FRANK E., Faunal Dispersal,

374

McDoweELL, E. C., Piebald Rats and

poh oy Factors, 719

Mam Variation, Correlation

REFI AM ni of Fertility in

26; S

Nipples in the, J. ARTHUR HAR-

RIS, 696

Mason, 8. L., and D. F. Jones, In-

heritance of Congenital Cataract,

119, 751

Mendelian oe tep oa Blend-

ee E. Cas :

Metz, CH. Skog e SE osome

Studies on the Diptera, III, Dro-

sophilide, 587

Mice, Pink-eyed White, carrying _

Color Factor, J. A.

46; The Oceurren ce of Three i

‘Red, 17 he Genetic Behavior of,

Ba, NN,

Mimicry in Butterflies, Poulton on,

sai i H. GEROULD,

MINc BoA; The Evolution of

ae “Cell, 5, re 271

ORGAN, T. H., The Bagster Gynan-

dromorph Bees, 3

MULLER, ANN r, The Mecha-

nism of Crossing-over, 193, 284,

350, 421

INDEX

767

be ge a an Piebald Rats,

Mutation Theory with Especial Ref-

erence to Ginothera, HARLEY

RIS BARTLETT, 513

Mutationist, Huxley as a, R. RuG-

Mutations in ao The Occurrence

of Three Sore e Color, C. C.

LITTLE, 335

Newt, the Vermilion- -spotted, :

viridescens, raine ons in, ALBERT

EESE, 316

NICHOLS, JoHN TREADW n

Primarily Voadapiive Variants,

565

Notes and Literature, 53, 184, 374,

502

(Enothera,

Mutation Theory iad

especial ARLE

reference

Har BARTLETT me

marckiana Hybrid o

= Bart

us, BRAD Lis Mo OORE Davis, 68

Ornameniation in the Killifishes or

Too bei innows, HENRY

Fow

Platha ae of the Blood, The

Pe anpas 4 Significance of the,

GEORGE G. ScoTT, 641

PAPANICOLAOU, GEORGE and CHARLES

. STOCKARD, urther Analysis

of the Hereditary Transmission of

Degeneracy and Deformities by

the Descendants of Alcoholized

Mammals tt

Parthenogenesis and Sexual Repro-

duction in Rotifers, D. D. WHIT-

NEY,

PEARL, RAYMOND, Fecundity in

Donestia: Fowl and the Sel ion

vt Namen 89

EVITCH, LEXANDER, The

Blips of the Sternum in Scorpi-

ons as a Systematic and a Phylo-

genetic Character, 600

Petunia, Selective Partial Sterility

throwing Sto

SAUNDERS, 486; Breeding Experi-

ments with, EDITH R. SAUNDERS,

548

Piebald Bi and Multiple Factors,

E. C. McDowELL, 719

ns, Sex Control and Known

RIDDLE, 385

ite Mice, carrying the

Color Factor, J. A. DETLEFSON, 46

768

Pina Inheritance AAS in, OR-

WHITE

Plant. p ari sine Rapidity

Evolution in Various, EDMUN

W. SINNOTT 6

Poh. in Butterflies, The

COMS of Seasonal, JOHN

, 310

Poulton on Mimicry in Butterflies,

v H. GEROULD, 18

Protozoan Life Cycle, nae had of

the, Gary N. CALKIN 257

is ee of Bvolution in Various

Plant Types, EpMuND W. SIN-

* 468

RASMUSSEN, ANDREW T., Theories

of Hibernation, 609

Rats, Piebald, and Multiple Fac-

tors, res bed MoDo » 719

, Variations i in the

Vermilion spotted Newt, D. viri-

desce 6

Reproduction, Sexual, and “ig peat

gen sis in Rotifers, D. D. WH

Gees. scar, Sex Control =

Known Correlations in Pigeon

385

Rotifers, Parthenogenesis and ss

ual eproduction in, D. D. WH

SAUNDERS, > li , On eri

Partial Sterility as an E

° Do uble- throwing Stock

and the Pelok e Results

of “sid Breeding Tarinat

with Petunia, 548

rare cla "The Shape of the Ster-

num in, asa Deere and Phy-

logenetic Charac EXANDER

facto gt 600

SCOTT, GEORGE G., The Evohitiovary

Si anes of the Osmotic Pres

sure of the Blood, 641

Selection, Problem and Fecundity it in

e Domestic Fowl,

Change ef, W. E. ;

Egg Production, H D. Goon.

479

"o Brag sa and Known Correla-

RIDDLE,

rape,

ALLEAU, 554; Problem and

Experimental Intersexuali ]

ARD GOLDSCHMIDT, 7

Shorter Articles and Disdjasias, 46,

THE AMERICAN NATURALIST

[ Vou. L

119, 178, ey eg 367, 435, 486,

575, 626, 688,

SINNOTT, EDM Co ig sia AD

Rapidity of ‘Evolution in Variou

Plant Types,

Soil “Temperatare ac Soil Moisture,

e to and the Di st

bation of ‘ne Cacti, W. A. Can

Statitial, Weighting fi Ape of

Adv Registry Cows, C.

76; Stu dies of the

in Scorpi-

s a Systematic and a Phylo-

genetic Character, LEXANDER

PETRUNKEVITCH, 600

STOCKARD, CHARLES R., and GEORGE

PAPANICOLAOU, ry Further Analysis

the ig rA of Alcoholized

econ S, o

CE, FRANK Ha Note on the

geeis of va Pattern in

Beans <r “ani Relation to Type

of Vine

Trifolium piaga quinquefolium,

H. E. EwING, 370

Use ery ar The Cause of the

Belief in, A. L. KROEBER, 367

VALLEAU, W. D., Inheritance of Sex

in the Grape, 554

Variability under Inbreeding and

Crossbreeding, E. CASTLE, 178

WALTON, E: B; Gametogenesis in

Pla

WATERMAN, T. T., Evolution of the

Chin, 237

WHITE, ORLAND a Inheritance

‘Studies in Pisum

Wuirney, D. » Parthenogenesis

and Sexual Reproduction in Roti-

fers, 50

T. K., Fasciation in Maize

Kernels, etek