VOL. XLV, NO. 541 : JANUARY, 1912

THE
AMERICAN
NATURALIST

A MONTHLY JOURNAL
Devoted to the Advancement of the Biological Sciences with
Special Reference to the Factors of Evolution :
CONTENTS
Page
I, The Inheritance ef Color in Short-horn Cattle, H.H.Lavemzi~n - - - b

IL. Supplementary Observations on the Development of the Canadian Oyster.
De. J.Starrorp 29

I. The Effects of Alcohol not Inherited in Hydatina senta. Dr. D. D, WHITNEY 41

yr
THE SCIENCE PRESS
LANCASTER, PA. GARRISON, N. Y.

The American

Naturalist

MSS. intended for publication and books, etc., intended for review should be
sent to the Editer of THE AMERICAN NATURALIST, Garrison-on-Hudson, New York.

Articles containing
tien are especially we

One hundrea reprints of contri
Further miee will be supplied at

Subscriptions and advertisements should be sent to the publishers.
dollars a year.

subscription price is four
Canadian g -five
thirty-five cents. The advertisi

cents additional.
ng rates are Four Dollars fo

on the problems ee organic evolu-

research work beari
oman and will be given Fakri in publicat
butions are supplied to authors free ef charge.
cost.

The
een oS postage is fifty cents and
The “erat for single copies is

page.

THE SCIENCE PRESS

_ NEW YORK: Sub-Station 84

Lancaster, Pa.

Entered tter, April 2, 1908,

tèh

Garrison, N. Y.

, Pa., under the Aet of

Con ongress of March 3, 1879.

THE BULLETIN—Por bargains in Ethnolograph-
ical and Pre-historic Specimens. Books on Natural
History, Science, Travel, Voyages, etc. See THE
BULLETIN post free for 3 cent stamp.
4 Duke St., Adelphi—London—Engiand

TENTH EDITION.
THE MICROSC COPE,

by Siow HENRY G om, of Cornell Univ enter Over
A ni

ai eao e pe aa T a o emi
COMSTOCK ALNA K A N. Y.

BIRDS’ EGGS

W. F. H. ROSENBERG,
57 Haverstock Hill, London, N. W.,

ka announce the publication of a new |
3 Price List (No. 15) of Birds’ Eggs, con- |.

Back or Current Numbers
of any American or Foreign technical
or trade journal and magazine fur-
nished on short notice at moderate
rates; all kinds of Government and
State Reports in stock. Clippings on
special subjects furnished promptly.
Large stock of American Naturalist,
Science and Popular Science on hand.

Magazines, Books and Papers _
| awe

THE TE

AMERICAN NATURALIST

A MONTHLY JOURNAL
DEVOTED TO THE ADVANCEMENT OF THE BIOLOGICAL SCIENCES

WITH SPECIAL REFERENCE TO THE FACTORS OF EVOLUTION

VOLUME XLVI

wo Lot. Garden
1913
NEW YORK
THE SCIENCE PRESS
1912

THE
AMERICAN NATURALIST

Vou RIN January, 1912 No. 541

THE INHERITANCE OF COLOR IN SHORT
HORN CATTLE

H.

H. H. LAUGHLIN

CARNEGIE STATION FOR EXPERIMENTAL EVOLUTION,
‘OLD SPRING HARBOR, N. J.

In order to adduce further evidence bearing upon the
problem of the inheritance of color in cattle the follow-
ing observations on the occurrence of dominant and reces-
sive whites in other animals are reported. Many species
of animals have some strains with solid white coats and
others with coats made up of both white and pigmented
areas. The white in such latter coats is always possessed
in a somewhat definitely arranged and progressive system
of areas characteristic of each species, spreading from
the first of these areas until the entire body is covered.
Thus in the guinea pig the whitening process begins with
the underline and large ‘‘centripetal’’ body blotches and
spreads until the hair, skin and body pigments are
entirely removed, the eye and the ‘‘centrifugal’’ coat pig-
ments persisting the longest. In the domestice cat the
process begins with the anterior underline and collar,
from which areas it spreads in large blotches. The
rabbit’s pigment areas behave similarly. With the parti-
colored dog of any or no breed, the whitening process
begins with a white line down the middle of the face, a

5

6 THE AMERICAN NATURALIST [ Vor. XLVI

white chest, a white collar and generally a white tail tip.
With cattle of all breeds and crosses possessing broken
color patterns the process begins with a white belt at the
rear flank, continues with another at the fore flank, a
white underline and a white forehead. With horses of the
English breeds it begins with a blaze face and white feet,
continuing with large body blotches. With the French
and the desert breeds it seems to begin witha ‘‘dappling’’
of the body hair—dark pigment persisting longest on the
legs—and continues through a lighter ‘‘dappling’’ to
white, the skin remaining black, while the sharper whiten-
ing process seems to follow the sequence observed in Eng-
lish breeds. In the case of the piebald negro the median
face line is quite noticeable. Thus, while there is for each
species a characteristic pattern, there is in reality a some-
what common pattern in all species of mammals possessing
particolored individuals. This common pattern is de-
seribed as follows: White line down the face, white under-
line, white anterior belt or collar, white rear flank belt
and white feet and switch. These white areas may, in
animals possessing but little white, be represented by
several smaller areas, but always near the median line
of the areas white in the larger pattern, which smaller
areas may fuse as the pattern becomes coarser. Thus
the white nose and forehead in some cattle may in
others make a continuous white line down the face.
The white spreads from the areas just defined with much
the same sequence as a fire would spread over a ‘‘hide-
shaped’’ meadow, starting at the centers homologous
to the first white areas of the coat. In all cases the
pigments seem to persist the longest in and about the
eyes and ears and at the buttock—the ‘‘centrifugal
areas’’ mentioned by Castle.

The coat of a white Shorthorn may consist of: (1)
Solid dominant white covering red (quite rare); (2)
some definite coat areas dominant white covering red,
others albinic white (very common); (3) some areas
albinic white, others dominant white not covering red

No. 541] INHERITANCE OF COLOR IN CATTLE 7

(quite common). In eye color the white Shorthorn is
either blue or brown; the roan and red Shorthorns are
always brown eyed. The characteristic pigment areas
may, in domestic animals, become subjected to rigid selec-
tion, resulting in modified forms—as the white belt of the
Dutch belted cattle and the white head, neck and under-
line of the Hereford. Approximations to the former and
to the reciprocal of the latter modifications (not of the
species pattern) are commonly observed among Short-
horns. In general, however, color patterns are quite
characteristic of the species and are quite persistent.
Greatly modified patterns are seldom seen and, moreover,
the reciprocal coloration is never seen.

Not only does the whitening process begin at definite
centers quite specific for each species, but it also seems
to be definitely progressive in tissues carrying pigment.
Thus, the whitening process begins in man with the skin,
extending to the hair, the iris, and finally the choroid.
Partial albinos often have blue eyes—the absence of the
iris pigment but the presence of the choroid. In the
guinea pig the hair and skin pigments generally seem to
disappear before those of the eye. Castle’® reports a
guinea pig with an area of red hair underlaid by a patch of
black skin, and observes that a white dog may have a
patch of pigmented skin somewhere under the hair coat.
Similar phenomena are found in all species having parti-
colored strains. It is a matter of common observation
that a white patch of skin or hair unsymmetrically cover-
ing but one eye of a dog or horse sometimes gives this
animal a ‘‘glass eye,” i. e., unsymmetrical eye color, the
blue eye being surrounded by white hair and skin while
the dark eye is surrounded by dark tissues; this, however,
is not always the case, for many times both eyes are dark.

In horses and cattle the hair is first whitened, then the
skin, the iris and the choroid follow in the order named.
A black horse or cow may have a spot of white hair on
some portion of the body; it may be entirely underlaid by

# í Heredity of Coat Characters in Guinea-pigs and Rabbits,’’ p. 46.

8 THE AMERICAN NATURALIST [ Von. XLVI

black skin—this is especially apt to be the case with small
spots, or under the center of such an area there may be a
pigmentless skin—generally characteristic of the larger
areas, while its margin is underlaid by black skin, but
never the reverse. In some instances, however, the pig-
mentless skin and hair areas exactly coincide. The hoof
of the white foot of a horse or cow is generally white,
that of the dark foot always black; however, quite often
a white patch or streak will extend to the hoof and then
come to an abrupt end, the hoof continuing in a vertical
line the same pigment possessed by the skin immediately
underneath the lighter hair patch giving rise to the hoof ;
thus the hoof is as dark or darker, but never lighter, than
the hair patch immediately above. In spotted horses it
is observed that a white coat spot crossing the mane will
sometimes whiten it, while in other instances it will not.
Mr. Chas. E. Burns, the pony breeder of Peoria, Ill.,
writes:

We naturally expect spotted Shetlands from spotted ancestors but
ean say that very frequently I have bred a spot to a spot and the off-
spring has been a solid color. On the other hand, I have very fre-
quently had spotted colts from solid-colored parents. The fact that
there is spotted blood in the ancestors of the solid colored ponies ac-
counts perhaps for the spots, and vice versa. There seems to be no
sure rule in governing the color of a Shetland. The mane and tail are
not always the same color as the adjacent color patches of the coat.
Very frequently I have seen a white mane come right out of a black
patch, although as a general rule the color of a mane is the same color
as the adjacent coat of the pony. This is general also as regards the
tail, but very frequently, as I say. a black tail comes out of a white
spot, or a white tail out of a black spot, and often the tail is both
blaek and white.

Mr. C. R. Clemmons, of Coffeyville, Kans., writes:

I have been breeding spotted Shetlands for twenty-five years. I find
that many mares of a solid color will bring spotted colts quite regu-
jarly when bred to a spotted stallion having considerable blood from
these colors but every now and then there will be a foal of perfectly
plain color as the result of this same mating. I am of the opinion,
however, that a spotted color breed could be obtained by breeding in
these colors and perhaps inbreeding.

No. 541] INHERITANCE OF COLOR IN CATTLE 9

The mane and tail of a spotted Shetland are not always of the same
color as the adjoining patches of the coat, there is sometimes a dis-
tinct color line between the mane and tail and the coat.”

Mr. W. A. Long, of Greeley, Ia., offers the following
evidence :

We have no recollection of seeing any roan Belgian stallions with
white manes and tails. Our experience with the roan stallions has
been that the colts are principally roan Je have in mind one roan
stallion that we imported that sired 68 elke one year and they were all
roans out of all colors of mares. We handle many chestnut horses
with white manes and tails, and they sire principally all chestnut
colts, but there will be some bays and other colors, as all colors are
represented in the Belgian breed. They do not all sire the white mane
and tail but many of the colts are so marked.

Mr. A. W. Hawley, of Pioneer, Ia., says:

I had a beautiful light mane and tail chestnut from a black Belgian
mare and a black Percheron stallion.

The roan stallion above referred to doubtless corre-
sponds exactly to either type No. 7 or No. 8 of Shorthorn
cattle, the dominant duplex white always persisting and
the hairs of the second network remain either red or
black, making the familiar ‘‘red-roan’’ or ‘‘blue-roan,’’
depending upon the gametiec composition of the dam,
epistacy and the laws of chance. Belgian horses resemble
Shorthorn cattle in that they are a breed of many colors,
including the interesting roan. Indeed, the roan, silvered,
barred, ‘‘agouti,’’ mottled, piebald, flea-bitten and other
variegated types of animals of all species so charac-
terized seem to behave in inheritance in a manner typified
by the roan Shorthorn.

While the mane and tail are generally of the same color
as the adjacent body coat, there is often. a pigment
differentiation—the coarser hair whitening first. This
phenomenon is also exemplified in the case of black or —
: spotted eats, which often have white ‘‘mustaches’’ grow-
ing from black or dark skins and coats, but never the —
reverse. Among wild animals the silver fox and the

10 THE AMERICAN NATURALIST [Von. XLVI

silver-tipped bear present instances of the whitening
process beginning with the hair tips. Recall, in this con-
nection, the lighter colors sometimes present on the hair
tips of the cattle crosses reported by Professor Went-
worth, and the ‘‘albinic’’ superficial tissues of the Silkie
fowl with its pigmented deeper tissues. Thus it seems
that with mammals and with some birds the whitening
process begins with the more superficial tissues and con-
tinues to the deeper ones; with mammals, coarse hair,
fine hair, skin, nail, sclerotic, iris, choroid, being the order
followed.

Permit a short digression into the plant kingdom. Of
the two or three hundred varieties of the dent corn, all
of the yellow and red varieties have red cobs and all of
the white varieties have white cobs, with the exception of
St. Charles County white, which has a red cob.

Jack-stock breeders in America are making an effort to
establish a ‘race of black animals with white points, and
the following evidence, while primarily bearing on this
problem, is typical of the behavior when not involving
the whitening process of pigments of domestic animals.
In the Breeders’ Gazette, May 10, 1911, a breeder states
this problem:

A jack is of good size, well made in every way but he is of maltese
color. He is exactly the color of his sire and his sire was a popular
jack in his locality and a first-class mule-getter. Is this color a real
objection? What is the prevailing color of mules sired by the mal-
tese-colored jacks?

To which L. M. Monsees, of Sedalia, Mo., the jack-stock
breeder and authority, responds:

I have seen some extra good mule jacks of the maltese color. A
maltese or blue jack, if from a good, large family of good blood, and
himself a good individual, will no doubt prove a good breeder. He
should be expected to get good solid colors—bays, blacks, browns,
blues and chestnuts.

Thus it seems probable that, when different parental
pigments, but not the whitening process, are involved, the

No.541] INHERITANCE OF COLOR IN CATTLE 11

pigments of the offspring are due to either a simple Men-
delian mixture of various dilutions of parental pigments
with their resultant hypostatic effect, or to minor reaction
between the determiners presented by the two parents
resulting in modified pigment bodies.

Barrington, Lee and Pearson’s study of color in the
eray-hound—Biometrika, 1904—presents evidence that
might well be given such an interpretation. Their elabo-
rate tables measure accurately the correlation of the
color of ancestry and offspring in this animal, but they
do not explain what takes place in the zygote upon its
creation by the union of two somewhat differently
organized and differently descended gametes ; nor do they
clarify the conception of gametic organization. It is,
however, primarily a study in the mathematics, not the
chemistry, mechanism nor biology of inheritance.

The black mane, tail and feet of the bay and of some
blue roan horses, and the white mane, tail and feet of the
chestnut Belgian seem to indicate that in horses the
whitening process may proceed somewhat out of synchro-
nism in its tissue and area sequences. The white mane
and tail seem to be causatively correlated with the chest-
nut coat of the Belgian, which white seems to be recessive
to the heavier pigments. Moreover, when the destroying
process attacks highly organized pigment bodies, is the
destruction always complete? May there not be resting
stages in this destruction and may not the series—blacks,
browns, bays, chestnuts, sorrels, duns and creams—be-
sides being different dilutions and hypostatic effects of
different pigments, represent these stages? Further-
more, may there not be a pigment sequence as well as
an area, tissue and ontogenetic sequence involved in the
whitening process? And are the chestnut, sorrel, dun
and cream pigments the ones most readily destroyed by
the antibody?

Note in this connection that in dogs and other mammals
having some individuals with black, tan and white areas, _
the black areas are quite often bordered by a zone of tan,

12 THE AMERICAN NATURALIST (Von. XLVI

and often small tan but rarely small white spots are
found within the larger black areas. These are the con-
ditions expected if the tan were an intermediate product
resulting from the attack of the destroying antibody
upon the determiner for the heavier pigmentation. The
sequence of color bands along the hairs of the wild and
agouti cavies, viz., heavily pigmented brown tip, yellow
band and leaden base, is also suggestive of the same
derivation of the yellow.

Besides an area progression and a tissue progression
involved in the whitening process in animals, there is
also an ontogenetic progression of the same process.
In man and in many pigmented animals a progressive
grayness, called ‘‘senile white,” comes with old age, in
some strains earlier than in others. White horses—
dominant white—are always born pigmented, but soon
change to white—juvenile white, it might well be called.
White Leghorn fowls are hatched white and, save for a
senile deposit of pigment, remain so.

The observed facts seem to demand intra-zygotie in-
hibition and reaction quite closely approximating the
following hypothetical processes: In a germ cell of
some heavily pigmented animal, say, of a black Angus
bull, let there be a specific chemical determiner (N) for
black pigment in the entire skin and hair coat and in the
sclerotic, choroid and iris. This determiner reacts like
and indeed may be a body closely related to the enzymes,
in that both may be weakened, exhausted, or totally in-
hibited without being impaired or destroyed by the pres-
ence of varying amounts of an antibody of some sort,
still greater amounts of which set up chemical reaction
resulting in partial or total destruction, depending upon
the relative quantity and intimacy of the two bodies in
much the same manner as trypsin is totally inhibited
but not destroyed by .05 per cent. of lactic acid, but is
totally destroyed by .1 per cent. of hydrochloric acid."
_ In the germ cell of a white mate of the aforementioned

1 Green, ‘The Soluble Ferments and Fermentation,’’ p. 198.

a
No. 541] INHERITANCE OF COLOR IN CATTLE 13

animal let there be an antibody (W) (analogous to the
acids in the above illustration) substituted for and
placed homologously to the determiner (N) for black
pigment, which antibody is capable of weakening, in-
hibiting and finally of totally destroying the deter-
miner according to the relative quantity and intimacy
of the two bodies. Let this antibody (W) exist in a
quantity large enough to totally inhibit the ontogenesis
of N, but not to effect its destruction. Now let fertiliza-
tion take place; the F, generation is white. A white so
behaving is said to be dominant. Because there was
only inhibition of N with no chemical reaction between
N and W, and segregation may take place in later gen-
erations according to the familiar formula, F, is said to
be simplex in reference to this unit character. If, how-
ever, the antibody in quantity sufficient for inhibition
makes its intrusion de novo into a gamete possessing a
determiner for N and this mutant germ cell meets
another of similar origin or descent, and the total
amount of the antibody is still sufficient to cause reac-
tion, a duplex dominant white offspring results, which,
mated with one of its own kind, will establish a race of
white animals, inhibiting somatically in heredity until
further disturbance by extraneous intrusion or by hy-
bridizing the determiner N. Animals of this sort upon
hybridizing—as Davenport has shown in his white Leg- |
horn crosses—may be made to yield the ancestral colora-
tion. If the antibody exists in quantity sufficiently great
to inhibit absolutely all of the determiner, with an excess
sufficient to cause chemical reaction destroying a por-
tion of N, then partial albinism results and the off-
spring, although entirely white, will possess some definite
areas of dominant white covering pigment, and others of
albinie white, breeding exactly like the white Short-
horns designated in this study as type No. 6. If, how-
ever, in the germ-cell of the white mate a still larger
quantity of W be present (exactly large enough to ef-
fect the total destruction of N) upon fertilization N and

a
14 THE AMERICAN NATURALIST [ Vou. XLVI

W react and are destroyed, the F, generation is white—
this time albinic white, mutants. W and N both being
destroyed, these animals are nulliplex and breed accord-
ing to the familiar formula.

As still another alternative, let W exist in still larger
quantities and the mating take place; not only is all of
N destroyed but there is an excess of W which gives
some areas of duplex dominant white not holding the
pigmented color as a recessive trait—in quite the same
manner as the Shorthorns designated in this study as
type No. 9 possess a coat solid white, some areas of
which are dominant white not covering the red and the
remainder of the areas are albinic white. A still greater
amount of W will apparently effect the total destruc-
tion of N, making the offspring in the entire coat duplex
dominant white, not holding N latent in the gametes and
not capable of ‘‘reversion.’’

Let the antibody exist in very small quantity, insuff-
ciently large to inhibit the ontogenesis of N, and let fer-
tilization take place. It is conceivable that the antibody
in such small quantity might have the same effect upon
N as alcohol has upon an enzyme, in which case N would
play its usual part in ontogenesis, but, being constantly
harassed by W, would finally be inhibited or destroyed.
The F, generation would then show senile grayness, as
in man; here again the most superficial tissues are first
attacked. If the antibody (W) is a trifle more concen-
trated the F generation will be born pigmented, but will
develop juvenile white, as with the white horse, which
as previously described is born with pigmented hair
and skin—the skin remaining black and the hair turning
white. Thus the process seems to be progressive, de-
pending upon different intrusions de novo—‘‘muta-
tions’’—and different inheritance lines for the pre-
sentation of various quantities of the antibody effecting
the destruction of N in a definitely progressive onto-
genetic, area and tissue sequence.

This is the hypothetical picture of intra-zygotic reaction

No.541] INHERITANCE OF COLOR IN CATTLE 15

demanded by the somatic behavior in inheritance of coat
pigments and patterns in Shorthorn cattle and in the
other instances above cited.

Now, let some further observations be reported and
then fitted to this conception for its support or rejection.

In the Breeders’ Gazette (April 12, 1911) in response
to an inquiry concerning the behavior in inheritance,
with special reference to the possibility of spotted off-
spring of a white stallion, described as follows:

He is white with pink skin and would be albino but for a very few
small specks in the skin and his dark eyes.

Dr. W. E. Castle answers:

The dark-eyed white condition is closely related to the piebald con-
dition. It may indeed be regarded as an extreme variation of the
piebald state in which the white spots cover the entire body except the
eye. Most black-eyed white animals produce a certain number of
piebald offspring, even when bred to animals exactly like themselves.

In reply to a request, W. P. Newell, of Washburn, Ill.,
the owner of the white stallion, supplies the following
data:

The albino offspring of my stallion do not have pink eyes, but have
“ glass” or “ watch ” eyes.

Their hoofs are white or flesh color; there are no spots in the skin
and not a colored hair on them. Not all of his white colts are albinos,
some of them have a few colored hairs in mane or ears; these I do
not refer to as albinos.

As a two-year-old this stallion was bred to six mares. Each one of
these six produced a white colt..

As a three-year-old he had thirty-nine mares; got thirty-three in foal.
About half of these were white, the others solid colors. These mares
were very ordinary and of all colors, every size, shape and age. Fol-
lowing are a few of the instances: Bay mare got white colt; bay got
black colt; two blacks got white colts; black got black colt; white and
buckskin spot got pure albino; dapple gray got white colt; flea-bitten
gray got white colt; three or more brown mares got white colts; two or
more brown mares got brown colts; brown mare got pure albino; one
sorrel got pure albino; one sorrel got brown colt. This will give you
an idea of how his colts are colored.

Nothing is given and not much can be deduced con-

16 THE AMERICAN NATURALIST [ Vou. XLVI

cerning the gametic make-up of these brood mares, but
this interesting stallion seems to be barely on the domi-
nant white side of the critical border between dominant
white and albinic white. Had the whitening factor been
a little more concentrated in the zygote giving rise to
him, doubtless the ontogenesis of his choroid would have
been inhibited or destroyed, the determiner for much
of his more superficial pigmentation would have been
destroyed and he would have been a true albino. Some
of his germ cells seem to contain the antibody W in
quantity and distribution adequate to inhibiting the
quantitatively definite determiner for pigmentation
found in some of the gametes of many pigmented mares ;
others of his gametes seem to lack this specific anti-
body, having in its place a determiner for dark pigmen-
tation, hence, he is apparently simplex with reference
to his dominant white determiners. If one of his
gametes possessing W unites with a mare’s gamete pos-
sessing pigmentation determiners greater than the
quantitatively definite determiner above referred to,
the inhibition will either not take place or it will take
place incompletely—in the latter case resulting in some
modification of the solid-color coat and skin condition.
If the mare’s gamete possessing less of the pigmenta-
tion determiners than the optimum quantity above re-
ferred to meets one of the stallion’s gametes possessing
W, the offspring will be white—dominant if the relative
concentration of the determiner and the antibody is such
as to cause only inhibition; recessive, i. e., ‘‘albinie” if
reaction occurs.

Let us consider the criteria of albinism. The general
conception among investigators and writers on the sub-
ject seems to be that all strains of albinos have origi-
nated through dropping from the germ-plasm deter-
miners for pigmentation previously possessed, rather
than to have descended from ancestral types never pos-
sessing such pigmentation. Generally an animal is
designated as ‘‘albino’’? when inhibition and reaction

No.541] INHERITANCE OF COLOR IN CATTLE 17

have covered the entire skin, hair, nail and eye pig-
ments. Castle'® in his ‘‘Heredity of Coat Characters
in Guinea-Pigs and Rabbits,” excepts ‘‘centrifugal’’
areas. There is, moreover, no reason to believe that the
pink eye of an animal may not result from the inhibition
of the pigment determiner as well as from its destruc-
tion. In the progressive development of whiteness from
senile white, juvenile white, dominant white covering
pigment, albinie white, to dominant white not covering
pigment, there seems to be, as we have seen, a species of
tissue resistance as well as of area progression to this
inhibition and reaction; the pigment of the deeper
tissues being more generally resistant, or at least slower
or later in succumbing to the attacks of the antibody.
These deeper tissues, when dark and covered by the
pigmentless tissues, give rise to a condition that is
proved by experimental breeding generally to be domi-
nant white. This is quite consistent with the present
conception, for if the skin below is still pigmented it is
quite probable that the hair pigments are only inhibited
and not destroyed, and by the time the inhibiting proc-
ess reaches the choroid, the destroying process is prob-
ably quite complete in the hair, and the animal is quite
‘properly designated as an ‘‘albino’’—recessive white.
It must be borne in mind, however, that albinism may be
either partial or complete; it may affect, the entire coat
color or it may affect only a limited area or a specific
tissue. In partial albinism the eye is often blue—the ab-
sence of superficial pigments but presence of the deeper.
Thus albinos become of great interest, and the study
of their behavior very complicated, on account of this
nascency of mutation. The intricate organization of the
_ gamete can be determined only by the study of its onto-
genetic sequence and end which, however, strongly sug-
gest that chemical bodies within the germ cell behave
exactly as such bodies within the test tubes of the lab-
oratory. The disturbance of a single determiner may
* << Heredity of Coat Characters in Guinea-pigs and Rabbits,’’ p. 9.

18 THE AMERICAN NATURALIST [Vou. XLVI

cause an accompanying correlation in something like the
following manner: Consider the determiner for some
definite somatic structure in a germ cell of one parent
to be destroyed by an antibody analogously placed in
the germ cell of the other parent; this chemical reaction
must leave a product, which product it is conceivable
may cause considerable havoe in so intricate a mechan-
ism. There is no reason to believe that this product
would of necessity confine itself to reactions with de-
terminers first attacked; it might indeed be conceived to
disturb or to destroy certain determiners for other
tissues and forms.

The Silkie fowl seems to have received a very
severe and peculiar upset in its determiners for
pigmentation—note its black eyes and black deeply
seated body pigments, together with its ‘‘albinic’’
plumage. Neither is there any probability, except
by chance, of parallelism or similarity between the
mechanical or chemical cause of such reaction and
the resulting determiners—a notion savoring some-
what of the earlier conceptions prevalent in some
quarters, of ante-natal influence—for Weismann?’ ex-
perimenting with Vanessa appears to have effected color
changes by means of temperature and Tower” to have
permanently upset that portion of the germ plasm of
Leptinotarsa determining pigmentation by means of
humidity and temperature. Thus, units may be made
and unmade, and thus a foreign body or force entering
a germ cell may conceivably cause a long series of reac-
tions, each product becoming a new reagent affecting
the determiners of many forms and tissues, if by chance
lethal damage is not done before equilibrium is reached.

Moore in his paper ‘‘A Biochemical Conception of
Dominance,” says:

When fertilization occurs, the germ cells bring into contact certain
substanees which are set free to react upon each other. Some of these

1 cí Germ Plasm,’’ p. 379.
%<í An Investigation in Chrysomelid Beetles of the Genus Leptinotarsa.’ ’

No. 541] INHERITANCE OF COLOR IN CATTLE 19

substances may react simply with cther substances and obey the Guld-
berg-Waage law of mass action, while others are of the nature of
enzymes (ferments) and accelerate reactions which are already going
forward at a very slow rate.

It has been many times demonstrated that a positive
determiner in a gamete of a simplex individual is not as
‘‘pure”’ as one from a duplex individual; furthermore,
a soma developed from a zygote made up of a gamete
containing a positive determiner, and another char-
acterized by its absence, is not as strong in the char-
acter in question as one produced by two duplex parents.
Thus, Davenport”? has shown that in mating dominant
white fowls with pigmented fowls there is often an ‘‘im-
perfection of dominance,’’ giving rise to some more or
less scattered pigmentation in F,; this he demonstrated
experimentally and, among other things, finds that

Two white Leghorns crossed by a black Minorea produced only white
hybrids, but the female hybrids at least had some black feathers. .. .
No barring resulted from crossing white Leghorn with . . . black
Minorca. . . . Of 26 hybrids between black Cochin and white P NE
8 were barred black and white.

And he concludes that—

alongside of dominance we must place an important modifying fac-
tor—the factor of the strength or potency of the representative of the
given character in the germ plasm. This is clearly a variable quantity.
If it is very potent we get a typically Mendelian result but if it is
weak, we will have imperfect dominance or failure to develop alto-
gether.

Thus the determiner for pigmentation in the black
Cochin seems to be more concentrated than the same
determiner in the black Minorca. Or is it possible that
the antibody, although present in quantities theoretically
in excess of the amount necessary for complete inhibition,
fails to effect such inhibition completely for the same
reason that the analogous phenomenon, due to some

aA Biochemical Conception of E University of California
Publications in Physiology, Vol. 4, No. 3, p. 11.

=‘<*The Imperfection of be reg American Breeders’ Magazine,
Vol. 1, No. 1, p. 42.

BO THE AMERICAN NATURALIST (Vou. XLVI

mechanical necessity, is commonly observed in chemical
experiments?

Obviously, the mass of the determiner for pigmenta-
tion is as potent a factor in determining the end result as
the mass of the destroying antibody. The kind or quality
of the pigment seems also to be a factor; the yellow or
sorrel pigments seem to be destroyed more readily than
the black or brown. It is also apparent that, due-to a
difference in the relative mass of the determiner and the
antibody in the zygote, one cross may affect total destruc-
tion of the pigment while another parallel or reciprocal
one may not. Thus, as above mentioned, Davenport’s
white Leghorn on black Minorea cross gave only white
or nearly white offspring, while his parallel cross, viz.,
white Leghorn on black Cochin, gave considerable black
pigment in the offspring. It has also been observed that
the barred Plymouth Rock male, which is much less
heavily pigmented than the female, when mated with a
white Leghorn female gives only white offspring, but
- the reciprocal cross, viz., the white Leghorn male on the
barred Plymouth Rock female gives barred, mottled,
gray, creamy and white offspring regardless of sex. In
this latter mating the two gametic elements, viz., the de-
terminer for pigmentation and the destroying antibody,
seem to be present in quite closely chemically balanced
masses and it would be interesting to know whether in
this cross the fluctuations across the color line are due to
accidental variations in the strength of the individual
gametic elements in question or to the Mendelian phe-
nomenon.

There is still another white possessed by birds and
mammals known as ‘‘structural white,’’ characterizing
some arctic animals such as the arctic fox, which is white
the year around, and the arctic hare and the ptarmigan,
which are pigmented at one season and white at the
other. It would be interesting to know whether the fur
and feathers of these animals in their unpigmented
phases possess oxidized pigments. There are, more-

No. 541] INHERITANCE OF COLOR IN CATTLE 21

over white pea fowls. The gorgeous hues of the common
pea fowl are due both to pigments and to defraction and
it would be interesting to know whether the white pea
fowl has lost its pigments or defraction surfaces, or both.

Animals of heavy pigmentation—as the blackbird,
the crow and the negro—are said to be more subject to
pre-senile and albinic white than others less heavily pig-
mented. Enzymes may be inhibited or destroyed by an
excess of their own products. May it be indeed that the
antibody (W) is itself a product of the determiner (N)?

To throw further light upon the whole problem, among
other things, a careful study should be made of the be-
havior in inheritance of the age of graying of the hair
and beard in man. If the conclusion of this paper pre-
sents a true picture, early graying of the hair and beard
will be found to be dominant over the later manifestation
of the same phenomenon. It is further anticipated that
a chemical analysis of senile white and juvenile white
tissues will show the same absence of somatic pigment as
Gortner2* has shown in his study of albinic and dominant
whites.

In this study of Shorthorn cattle nine theoretical game-
tic coat-color types are defined. As previously stated, the
striking fact is this: The roan of type No. 3 (which is
reciprocally colored as compared with the ordinary color
pattern of cattle) is never observed, and quite probably
the red of type 2 is also missing. The reason is appar-
ently as follows: The antibody inhibiting and destroying
the determiner R (for red pigment) first attacks through
mechanical and chemical necessities the determiner for
coat pigment in the somatic areas of Set 1 (roughly—
flank, heart girth, forehead) and progresses systemat-
ically through the areas of Set 2 (roughly—underline,
barrel, legs and quarters, head and neck) according to the
following scheme:

= íí Spiegler’s ‘White Me’anin’ as related to Dominant or Recessive
White,’? THE AMERICAN Naturatist, Vol. XLIV, p. 501. ;

22 THE AMERICAN NATURALIST [ Vou. XLVI
TABLE VIII
| Gameti e |. Ga ametic | Number of, Examples of Some
Class | Formula for| che ses ar ‘Inheritance Breeds of Cattle
ve | Ament of Antibody | Areas of Unitsin | Representative of
tage | resent | First At- | pe At lEntire Coat the Respective Rest-
ieee ack. | ack, | Pigmenta- | ing Stagee of od
| (Set 1) (Set 2 J tion Whiten ng Proces
—| —— ——— | --

1. (None or too little to w,P, w,P, | One. Angus and solid
| start inhibition. black breeds
| | enerally.

2. | Enough to inhibit the W,P, | wP, | |Two or |Holstein and
determiner for pig- | two spotted breeds
mentation of the areas groups. | generally.

of Set 1.

3. | Enough o Wi WLP, | WPi | One. White park
hibit the | rete bod | | cattle of Bri-

| tain.

4. nough more to start| w.p, | W,P, | Two or |Not fe neat
reaction and to de- two b i breed
stroy the determiner | | groups. | nor ever ob-

for pigmentation of
the areas of Set 1.
5. nough more to con-| Wp

served in mon-

One. Casita albi-

WP2

Two or|White Short-
horns of Type 9

|
W2P2 |
| tw
ody in place of deter- | | | groups. | of this paper.
e j
|

6. | Enough more to de- | WP:
it anti- |

7. |Enough more to de- W,p, W.p, | One. Remotely pos-

sible that some

y in place of deter- | | strains of British
miner for pigmenta- | | white park ss
tion of areas of Set 2. | are of this type.

W = presence of antibody. P = presence of determiner for pigmentation.
w = absence of antibody. p = absence of determiner for pigmentation.

In Shorthorn cattle, classes 4, 5 and 7 of this table are
not met with, neither are conditions parallel to class 4
ever observed in any other mammals. The further expla-
nation may be as follows: Reaction between W and R does
not begin until an excess of W is present (a condition not
hard to parallel in the chemical laboratory) but when
reaction does begin it is quite rapid, destroying all of R
and most likely leaving an excess of W at the point of
first attack. This would eliminate Class 4 (type 3 of
the series previously described) and Class 5 (pure
albinos) of this table. There may be ‘‘albino’’ cattle;

No.541] INHERITANCE OF COLOR IN CATTLE 23

Pearson** reported a rumor of a herd of such but he was
unable to locate it. Wilcox and Smith” describe a race of
white cattle—Polled Albino—made by crossing a white
Shorthorn cow with a polled bull of unknown breeding.
The Swedish cattle were thought to possess ‘‘ pink eyes’”’
and if so were probably albinic in their entire coat; the
Polled Albinos are doubtless ‘‘partial albinos.” White
Shorthorn cattle are generally blue-eyed, however, a con-
siderable percentage are brown eyed.

The following chart of the ancestry and offspring of
‘White Rose,” the first cow purchased by Mr. J. F
Hagaman, of Leonard, Mich., is prepared from data sup-
plied by him:

Far ass set
Koan
a Cot toet Banes re Spring wood
Rr,
=
Ue
Whitt
Roars.
Cuart No. 3 ý

He aiso writes:

I purchased another cow, Daisy Dean, red and white. All her an-
cestors were red, red and white, or roans. She was bred to Park Farm
Prince (roan) and produced twin bull calves both white. They were
exactly alike and were made into steers. A drover took them to Bos-
ton where they sold for $500. ... All the white calves had blue eyes,
flesh-colored noses and light skins.

Dr. D. M. Kipps, of Fort Royal, Va., writes:

I feel sure I never had a white Shorthorn with a black nose; I had
one or two that had slightly cloudy noses. I think every one had pink

“(On the Inheritance of Coat-Colour in Cattle,’’ Biometrika, 1905-06,

p. 436. É
21‘ Farmers’ Cyclopedia of Live Stock,’’ p. 369.

24 THE AMERICAN NATURALIST [ Vor, XLVI

skin underlaying the white coat and nearly every one had slightly
reddish hair on the inside and around the outer rim or auricle of the
ear.

Mr. J. H. Hawkins, of Xenia, O., writes:

Will say I have never seen a white Shorthorn with pink eyes. My
white Shorthorns have pure white coats, pink skins and brown eyes.
As to black noses, they are not a rare thing to see . . . now and then.

Shorthorn cattle were made from the Anglo-Saxon reds
—Class 1 of the above table No. VIII; the Flecking—
Class 6; the Romano-British—Class 3, and probably some
other primitive types. Evidently none of the breeds of
domestic cattle has yet reached stage 7, i. e., solid domi-
nant white not capable of reversion. The Shorthorns of
to-day present all the possible combinations of Classes 1,
2, 3 and 6.

In reference to the fact that the race of duplex yellow
mice has never been produced and in view of what Castle®
says,—viz., that the union between germ cells carrying
only yellow pigment is doubtless affected, still all such
germ cells from some cause are doomed to destruction,
may it not be that in so delicately adjusted a mechanism
two of these specific determiners present a lethal dose?
May this not be one of the causes of the limits of hybridi-
zation and of the sterility of hybrids? The germ cells are
doubtless distinguished by both a specifie architectural
and a specific chemical organization of the greatest nicety
of adjustment and balance. The closest approach in the
chemical world to their behavior is that of the enzymes,
which, though not entering into reactions, may bring them
about; while in the course of its own continuity the germ
plasm gives rise to cells of its own kind, supplying them
with bodies behaving in an enzyme-like manner sufficient
for their own continuity and for a long series of onto-
zenetic processes.

It is obvious that a disturbance of some consequence
would follow the advent of a foreign body or of unusual

æ ‘í Modified Mendelian Ratio amorg Yellow Mice,’’ Science, December
16, 1910, Vol. XXXIT, p. 868.

No. 541] INHERITANCE OF COLOR IN CATTLE 25

quantities of a normal body presented either by hybrid-
izing or by osmotic intrusion; perhaps it may clarify the
conception to make analogy to the degree and sequence
of reactions in test-tubes or other containers of more
complicated design holding the same chemical in varying
quantities, places and degrees of nascency, wrought by
the addition of varying quantities of the same reagent.
The inhibitions and reactions expected from such condi-
tions would begin at definite places, would continue in a
more or less definite suecession characteristic of each set
of conditions, would complete a reaction first in definite
parts and would proceed with varying degrees of speed,
might effect a reaction and deposit an excess of reagent in
some parts before even reaching other parts. Let there
be an equilibrium following reaction; then add more of
the reagent or of the chemical acted upon and it is easy
to picture subsequent reactions all of which are closely
analogous to the processes which the study of Shorthorn
cattle leads us to believe have taken place within their
gametes and zygotes. The behavior of their coat color
and that of many other animals demand such behavior
within the zygote. Thus such processes seem to account
for the coarse mosaic or the spotted, and the fine mosaic
or the roan color coat, the imperfection of dominance,
reversion, the origin of the mottling and barring of fowls,
the progressive dappling of horses, the peculiar behavior
of ‘‘albino’’ guinea pigs, the characteristic behavior of
coat pigment and patterns in Shorthorn cattle, and other
similar phenomena. The stag, but not the doe, caribou
possesses a beautiful white collar, and it may be that sex-
limited characters are wrought by a sort of ‘‘havoec’’ or
series of progressive reactions, preceding chemical equi-
librium caused by the introduction of the essential sex-
determiners. :

A human family is recorded? in which a pre-senile
gray spot oceurs in the beard of the left cheek of many
of its male members. In possible explanation, it is sug-

= Files Eugenies Record Office, Cold Spring Harbor, L. I.

26 THE AMERICAN NATURALIST [Vou. XLVI

gested that a small quantity of some antibody somehow
inhibited or destroyed a portion of the determiner for
pigmentation in the germ cell from which this family
sprung. This indeed points toward a possibility that unit
characters may arise from a partial destruction of larger
units; that a determiner for a unit character behaving
precisely in unit fashion may be a complex capable of
being shattered into a large number of independently be-
having ch ters. Small as the germ cell is and quanti-
tatively insignificant as the determiner for the skin and
hair pigment must be, the facts demand that this body
consist of many molecules arranged in definite structure,
each one destined for a somewhat definite ontogenetic
process leading to a definite somatic end. Thus the often
inherited specific color mark seems to indicate that a
color pattern once produced—no matter how intricate or
complex—will reproduce itself exactly until its deter-
miners are disturbed by unbalanced bodies or forces pre-
sented by fertilization or otherwise.

The Shorthorns are a race of white cattle caught in the
making and preserved in the nascent state by a rigid selec-
tion. It is thus conceivable that mutations may arise
constantly, and that they may be progressive in char-
acter. Complications resulting in somatic effect are
legion, but nothing occurs in the germ cell giving rise to
new ch ters, splitting up and combining others and
dropping out still others, that can not be analogously
pictured with the simple operations of the chemical
_ laboratory, and as Shull ’s?s illuminating ‘‘Simple Chem-
ical Device to Illustrate Mendelian Inheritance’’ seems
to indicate, the analogy is too constant and too far-reach-
ing to be cast aside as a mere pedagogical device. It may
indeed be a simple statement of facts of intra-gametic
and zygotic behavior and the analogy may no longer be
needed to picture the actual conditions.

*The Plant World, Vol. 12, pp. 145-153, July, 1909, and companion
paper, ‘‘The ‘Presence and Absence’ Hypothesis,’’ THE AMERICAN NAT-
URALIST, Vol. XLIII, No. 511, pp. 410-419, July, 1909.

No. 541] INHERITANCE OF COLOR IN CATTLE 2

The evidence of this study of Shorthorn cattle is to
support that theory of unit segregation incompatible with
a somatic blend in the ultimate unit, and that theory of
heredity permitting intra-zygotic inhibition and reaction
in response to specific set conditions.

The mutually corroborative evidence of the authentic
history of this breed of cattle, the behavior of their coat
pigments and patterns as recorded in the most extended
authentic records of pedigree breeding of domestic ani-
mals, analogy to the occurrence and behavior of pigments
in other animals, and the close fitting of the final work-
ing hypothesis, amply justify the following conclusions:

1. Shorthorn cattle as a race possess two kinds of white
hair. (A) White, dominant to all pigments (analogous
to the white of the Leghorn fowl) in a series of areas
varying somewhat: in size and shape but in a given indi-
vidual always definite and genetically independent—a
few at the front flank belt, a larger number or larger
areas about the rear flank belt, a few along the underline
and a fine network covering the remainder of the body.
A few animals from their Romano-British ancestry have
the entire coat of dominant white. An area of dominant
white may be duplex or it may be simplex. In the former
case its possessor will throw only gametes with deter-
miners for dominant white; in the latter alternately ga-
metes with determiners for dominant white and for red.
(B) White, recessive to all pigments (analogous to the
white of the Silkie fowl) in a series of definite areas gen-
erally smaller than those of the dominant white, forming
a fine network about the neck and head, the sides and
back, and the hind quarters and legs—quite precisely
excluding the areas of the dominant white network.
From their Dutch ancestry, this mosaic may in some
strains be quite coarse. It is doubtful if a strain albinic
white in its entire coat exists within the Shorthorn breed.

2. The color effect of an indiviđual Shorthorn is deter-
mined by the registering of fortuitously one of the alter-
nate color phases of each of the genetically independent

28 THE AMERICAN NATURALIST | Vou. XLVI

color areas gametically possessed by each of the two par-
ents, together with such intra-zygotic inhibitions and re-
actions between the determiner for pigmentation (R) and
the antibody (W) as may result from definite concentra-
tions and intimacy of these two bodies presented by the
two parents upon the formation of the zygote.

SUPPLEMENTARY OBSERVATIONS ON THE
DEVELOPMENT OF THE CANADIAN
OYSTER

J. STAFFORD, M.A., Pu.D.
BIOLOGICAL STATION, DEPARTURE Bay, B. C.

In the Amertcan Narurauist of January, 1905, Janu-
ary, 1909, June, 1910, I have given some account of obser-
vations (in 1904) on the development of the oyster at
Malpeque, Richmond Bay, Prince Edward Island, Canada.

Opportunity to verify, continue, and extend these ob-
servations was again afforded in 1909, when I studied
the oyster in the most important centers along the east
coast of New Brunswick.

In the present summer, 1911, being occupied at the
Pacific Biological Station of Canada, in Departure Bay,
near Nanaimo, Vancouver Island, I have the privilege of
observing some of the Prince Edward Island oysters
transplanted to this vicinity in 1905, as well as adding to
my acquaintance the little British Columbia oyster, so
different in size, appearance, habits and reproduction.

In the intermediate years, not being located in oyster
regions, I devoted a good deal of time to other bivalve-
larve, largely with a view to making my studies of the
oyster more secure, the main results of which have been
given in a paper ‘‘On the Recognition of Bivalve Larve
in Plankton Collections,’’? unreasonably delayed in publi-
cation at Ottawa.

Tn all this work I have kept sample preservations with
dates and ‘localities, which have often proved of great
service in judging of questions that subsequently arose.

My first work began where that of Brooks left off, and
showed for the first time that later stages of the oyster-
larva undoubtedly exist, and when, where and how they

30 THE AMERICAN NATURALIST [ Vou. XLVI

may be procured, as well as the length of the period of
their free-swimming life. The larve obtained by Brooks,
Rice, Ryder, Winslow, and others were obtained by cul-
ture from fertilized eggs, and were at most six days old,
and in the young straight-hinge stage. In Europe larve
of a similar age, size and structure had been taken from
the infra-branchial cavity of the parent oyster by Da-
vaine, Lacaze-Duthiers, Costé, De la Blanchére, Gwyn
Jeffries, Saunders, Salensky, Mobius, Horst and Huxley,
but the older, later or larger stages were quite unknown.
This left room for some speculation as to the exact time,
place and manner in which the succeeding stages should
be found, as well as occasioned the prevalent mistake that
the free larva settles down at this period to become a
fixed spat. Brooks wrote. ‘‘All my attempts to get later
stages than these failed . . . and I am therefore unable
to describe the manner in which the swimming embryo
becomes converted into the adult, but I hope that this
gap will be filled, either by future observations of my own
or by those of some other embryologist.’’ In a similar
way Jackson, at a later period, speaks of ‘‘a blank in the
knowledge of the development of the oyster.” This
‘‘gap” or ‘“‘blank’’ is now completely filled. My studies
prove that the larva continues to live as a larva in the
sea-water about oyster-beds for two or three weeks
longer, where it swims about, feeds, grows and changes
in structure, and that it first settles down to become a
sedentary spat, fixed to shells or other objects, at an
age of three to four weeks from fertilization—the length
of time depending to some extent on temperature, food,
individuality or such causes. This information has been
gained through the method of procuring oyster-larve
from the waters of oyster-areas by means of a plankton-
net, and connecting them in series with younger stages
obtained by fertilization and culture and with older
stages obtained by catching spat on glass, shells, ete., so
as to make out the complete life-history.
The discovery that the hitherto unknown stages of the

No. 541] THE CANADIAN OYSTER $1

oyster-larva can be conveniently obtained by a plankton-
net carries with it the possibility of a practical applica-
tion of inestimable value in the culture of oysters. From
the time of the early Roman Empire it has been known
that oyster-spat can sometimes be obtained on ropes,
anchors, piles of wharves, stones, shells or other natural
or artificial objects in the sea, and some sort of method
of culture has long been in use in many countries. At
times men have risen to exalted conceptions of the possi-
bility of finding a practicable, safe and sure method of
catching, retaining, and rearing the young spat. I quote
Winslow to the effect that ‘‘Thousands of dollars would
be annually saved by the Connecticut oystermen if they
could determine, with even approximate accuracy, the
date when the attachment of the young oyster would
occur. Hundreds of thousands would be saved if they
had any reliable method of determining the probabilities
of the season.” This is now possible.

It is well known that oyster or other shells dried and
whitened in the sun form the very best oyster-collectors
or cultch. To put these back into the water haphazard
has often resulted solely in the loss of all the labor of
preparation. In even a few days they may become cov-
ered with a slimy coating which reduces or largely
destroys their efficiency. The point is to be able to deter-
mine with accuracy, for each season and for every local-
ity, when oyster-larve are present in the water full-
grown and ready to settle as spat, so as not to run the
risk of losing adequate value for the laboriously pre-
pared cultch.. A man instructed and qualified in the
method of taking plankton and in identifying oyster-
larve can tell almost to a day when is the proper time to
put out cultch so as to obtain an abundant and copious
set of spat. It is not enough to know about the time, or
to know the time for certain previous years, or to know
the average time.

Three methods are open to the expert: (1) Examina-
tion of the genital organs of adult oysters to determine

32 THE AMERICAN NATURALIST [Vou. XLVI

when the eggs are ripe, (2) examination of the sea-water
to learn if oyster-larve are present and in what stage,
(3) examination of natural or improvised objects in the
water to discover if young spat are already formed. The
first is not immediately determinative because of the long
period of development separating spawning and spat-
ting. The last is not very practicable because of the diffi-
culty of finding and recognizing the youngest spat before
the period is gone by for putting out cultch. The second
is the only practicable and conclusive method and its
efficiency is proportionate to the number, care and ac-
curacy of the observations. Its success will increase with
experience.

This method makes use of the colossal number of larvæ
lavishly provided by nature to offset the exigencies and
accidents of life and insure a reasonable chance of keep-
ing up the stock. I believe that all the larve an army of
men could raise up and turn into the sea would not ma-
terially alter the number of successful individuals in the
set of spat. But on the other hand a few culturists could
enormously increase the chances for a successful catch
by spreading an abundance of suitably prepared cultch
at the proper time and place.

In the paper of 1909 I have described the method of
obtaining plankton, the appearances and measurements
of the oyster-larve to be recognized, the time of the year
to begin making observations. In the paper on ‘‘Bivalve
Larve’’ I distinguish in sizes, shapes, colors, the com-
monly occurring associates of the oyster-larve which
might be taken for the latter. In the present paper, after
long reflection, I suggest a practical application of the
knowledge acquired.

I should not omit to mention that the paper of 1910
connects the larva, through the youngest microscopic
spat, with the macroscopic spat of fishermen and finally `
with the adult. Similarly in 1909 I performed extensive
artificial-fertilization experiments, while at Shediac,
Caraquette and Malpeque, in order to connect the small-

No. 541] THE CANADIAN OYSTER i

est plankton stages of oyster-larve with culture-stages
and through these back to the egg. Larve by the million
were reared in beakers of sea-water at a temperature
little above 20° C. and with a specific gravity (salinity)
varying somewhat under 1020. I also carried Caraquette
oysters to Malpeque and raised up larve from eggs cross-
fertilized between two such obviously different varieties
as the small, narrow, curved, thick, hard and heavy Cara-
quette oyster and the fine, large, broad, straight, clean,
smooth specimens from the Curtain Island beds.

In 1896 and again in 1905 the Canadian Government
had Atlantic oysters transshiped to the Pacific and put
out at selected places. In the latter year some of the
places were chosen by Captain Kemp, expert in oyster
culture.

Being occupied this summer at our Pacific Biological
Station, I have taken advantage (although not requested
to do so) of my proximity to three of these places to
search for the transplanted Prince Edward Island
oysters, and to examine plankton taken in the vicinity.
At the first place, Hammond Bay, being a small bay and
close to hand, I could easily over-run all the beach at low
water, and soon discovered the dead shells that had been
deposited too far above low-water mark. At Nanoose
Bay, some twelve miles away, perhaps five miles long and
a mile and a half wide, with extensive flats at low tides,
this was not so easily done. Having spent three summers
with Captain Kemp, I thought now to test my judgment
of where he would select to deposit the oysters. As the
tide was unfavorable at my first visit I used the dredge,
and was afterwards surprised to learn that I had actually
calculated to within a few rods of the place. At the
second visit I went to look at other parts of the bay, but
on the third returned and, with a favorable tide, could
wade and pick up some of the oysters. This was at 3
P. M., July 17, and I took 16 fine living specimens of the
Malpeque oyster for examination—two or three of them
with pieces of Prince Edward Island red-sandstone still

34 THE AMERICAN NATURALIST (Vou. XLVI

attached to them. They varied from two and three
fourths to five inches in length, some of them showing
considerable growth. This proves that Atlantic oysters
can be transplanted to the Pacific and remain healthy
and grow. Upon reaching home I proceeded to examine
some of the oysters and it turned out that only one had
already spawned while the other fifteen were ripe and
generally somewhat distended with eggs or sperm.
This proves that the transplanted oysters can come to
maturity and ripen the reproductive elements.

At 7.10 P. M. of the same day I put together eggs and
“sperm in a tumbler of sea-water and at 7 A. M. next
morning there was an abundance of segmentation stages
and free-swimming larve. This proves that the oysters
can spawn and that the eggs can develop into young. I
make these statements because of a prevailing opinion
that the transplanted oysters have all died, and the few
people who think there are still some living are dogmatic
in their assertion that they do not breed.

Plankton taken at intervals at Hammond and Nanoose
Bays had not yielded any oyster larve, which became ex-
plainable upon finding the condition of the reproductive
organs. A further observation on this was afforded on
the 26th of July, when I examined a second lot (obtained
at a very low tide the day before) from Nanoose Bay.
The forty-seventh oyster examined was the first to yield
good ripe eggs—all previous ones were spawned with the
exception of four or five which were ripe males. The
interval between these two visits had been the hottest of
the summer and the oysters had nearly all spawned in
this period—slightly later than is usual on the Atlantic.
On the 27th I made a trip to Oyster Harbor (Ladysmith),
about fifteen miles from here, where I had better luck in
getting track of the few transplanted oysters. In a-
similar way I examined several individuals and took
plankton which for the first time contained larve of the
Atlantic oyster—recognizable by their shape and meas-
urements but not presenting such a deep pink or brown

No. 541] THE CANADIAN OYSTER 35

coloration as in their native home. For comparison
with my former papers I will give the measurements of
a single specimen with the characteristic postero-dorsal
high umbos, the large convex left valve, and the smaller
and flatter right valve, velum, foot, pigment spot and the
rest. Ocular V, objective 4, 42 long by 37 high
(—.289 x .255 mm.). This proves that larve grow up.
There is only one other bit of evidence possible and that
is to find spat. This I have not done as yet. It is too
early for this year’s spat and I have not seen any un-
doubted specimens of a former year’s spat. One can
judge that the comparatively few descendants of two
and a-half barrels deposited at Hammond Bay, five
barrels at Nanoose Bay, and one barrel at Oyster
Harbor, when dispersed over the broad areas at their
command, would not prove very conspicuous objects,
which is again complicated by the presence of millions of
British Columbian oysters of varying sizes, shapes, and
complexions.

I regard my findings as conclusive and would urge the
transplanting of Atlantic oysters (Ostrea virginica
Gmel.) to the Pacific in greater quantities. The At-
lantic clam (Mya arenaria L.) has propagated enor-
mously here notwithstanding the fact that it has more
competitors in its particular habit than in its original
home.

Ostrea lurida Carp.—Even before making any head-
way in the foregoing researches, I had begun to gather
information on the occurrence, size, shape, color, struc-
ture, breeding, etc., of the British Columbia oyster.

This species is not common in Departure Bay, or in
Hammond Bay, but a few specimens may be found under
stones exposed at about one hour from low water in
front of the C. P. R. cable house in the former, and just
inside the far point of the latter, and are usually so
broadly and solidly attached (with the left valve against
the under side of the stone and hence uppermost) that it
is scarcely possible to separate them without destroying

36 THE AMERICAN NATURALIST (Vou. XLVI

the attached surface. But on the extensive flats at the
upper ends of Nanoose Bay and of Oyster Harbor they
occur free on the surface by thousands and more or less
covered with barnacles.

Good specimens reach two inches in length by an inch
and a half in breadth, with a straight dorsal margin and
a semicircular ventral curvature. The right, upper or
smaller valve is nearly flat or but little convex and fits
into the margins of the larger, convex, lower or left valve,
the greater part of the lower and posterior margin being
scalloped, while the left valve has corresponding ridges
and points. The color is usually dark (those under
stones lighter) with the older parts weathered grayish
and the umbonal region of the left valve is often attached
to a small stone or another oyster or bears a scar. Imn-
ternally the shell is extensively pigmented, dark, with
smaller bands or blotches of lighter pearl, while the
muscle sear is rather lighter and banded. The mantle is
broadly margined with dark, which may also creep up
on to the abdomen.

The most interesting feature in connection with the
Pacific oyster of Canada is its divergence in some re-
spects from the mode of breeding of our Atlantic species.
In the British Columbia form there is no primary sepa-
ration of individuals into males and females—the sexes
are united in each individual. In other words each in-
dividual is bisexual, monecious or hermaphrodite. In
this respect it is identical with the English or common
European species (Ostrea edulis L.).

My first observations were made on July 12, on
specimens procured under stones near the Biological
Station. Nearly all appeared to be males, and, as they
were of small size, I took it that, as commonly occurs,
the males had ripened earliest. But one was of medium
size and contained eggs that at once attracted my atten-
tion on account of their large size, opacity and rare ex-
hibition of nucleus. Measured exactly as all my former
measurements, these gave: Oc. V, obj. 2—6.5; Oc. V,

No. 541] THE CANADIAN OYSTER 37

obj. 415; Oc. V, obj. 772. Another individual, ob-
tained since, with an abundance of eggs oozing from the
oviduct, pure and ripe, gave the almost unvarying meas-
urement of the egg as: Oc. V, obj. 775. This when
calculated is 75 X 1.45»—108.75» = slightly over .1
mm. = slightly over 450 inch — fully twice the diameter
of the egg of the Atlantic oyster, and perhaps identical in
size with the egg of the English oyster.

In making measurements it is important to use only
ripe eggs, as in this case, and to select those that are
spherical or nearly so and not flattened by the weigth of
the coverslip, as well as to extend the measurements to
many individuals in order to exclude all possibility of a
slip. The nucleus is between one half and two thirds .
the diameter of the egg.

Upon turning particularly to spermatozoa I found
them in every individual—even between the eggs of those
containing eggs in the gonad. The younger individuals
had no ova, but all sperms. Some of the older ones had
a few big, soft, opaque, irregular, elliptical, oval or
nearly spherical eggs, scattered among irregular masses
of less than half their size, which are balls of spermatids
on the way to development into spermetozoa. One of
these measured 46» 40», and each one is kept in a
dancing or rolling movement, somewhat like that of many
infusoria, by the flapping of the tails of the ripening
sperms on the surface. Between these masses are mil-
lions of mature, free, dancing spermatozoa, of which the
tails are rarely visible until one searches for them with
a high power. I have not yet made extensive measure-
ments of the sperm on account of the difficulty of measur-
ing such exceedingly small objects with certainty, but I
believe the sperm of the British Columbia oyster is
smaller than that of the Prince Edward Island oyster,
which may have some relation to the particular mode of
fertilization, such as being introduced by the respiratory
current. In some parts of the gonad ova may be plenti-
ful, while at other parts there are only sperm-balls.

38 THE AMERICAN NATURALIST Vou. XLVI

Later, in the warmer weather, the sperm may be pretty
well run off and the reproductive organ contain mostly
eggs. In this way the younger oysters, and the older
oysters at the beginning of the season, may be physio-
logically males, while older oysters at the height of the
breeding season may be physiologically females.

Oysters from Hammond Bay showed the same phe-
nomena.

Upon finding an abundance of larger oysters on the
surface at Nanoose Bay, I brought home a pail-full of
picked specimens to serve as a convenient stock for ob-
servation and experiment. On July 16 I found a speci-
men with perhaps half a teaspoonful of eggs in various
stages of segmentation, lying free in the lower valve—a
mass of white granules. The ripe eggs ooze into the
infra-branchial cavity and lie on and between the gills,
i. e., between the two folds of the mantle, where they are
retained apparently without any retaining, sticky matrix.
I suppose that it is here they first meet with ripe sperms
from other individuals, for I do not believe that at this
time the sperms of the same individual are physio-
logically capable. The whole oyster appears exhausted,
the gills rent, the flesh collapsed, soft and parts of it
almost rotten. On July 24 I opened one hundred of
the stock supply and found six with eggs, embryos or
conchiferous young, in the infra-branchial cavity. All
the others were in process of spermogenesis and
oogenesis.

An experiment that has often seemed possible to me is |
to do the same with the European oyster, by way of
artificial fertilization, as Brooks did with the American
oyster. Now that I had an oyster essentially the same as
the European I tried it, and with seeming success, but
of course it is difficult to be sure that sperm from another
had not already had access to the eggs. Unripe eggs
are no good; eggs already freed from the gonad may
have come in contact with sperm. This restricts one to
finding a specimen just before but just on the point of

No. 541] THE CANADIAN OYSTER 39

extruding its eggs. I also tried Atlantic oyster eggs
with Pacific oyster sperms, as well as Atlantic oyster
sperms with Pacific oyster eggs, but without success, as
one might suppose. I put eggs, embryos and larve of
both species together under the same coverslip for com-
parison—those of the small British Columbia oyster
looking like giants beside those of the large Prince Ed-
ward Island oyster. This is a curious phenomenon
which I have several times observed on other species,
e. g., the very large eggs of Astarte compared with the
small eggs of large species like Mactra.

For the study of segmentation, ete., the Atlantic species
is of advantage on account of smaller size and greater
transparency. The order of segmentation appears to be
the same in both—both subject to variations such that it
would require a great number of painstaking observa-
tions to decide exactly what is the normal mode in good
healthy eggs. I have, on both sides of this continent,
spent considerable time in trying to determine the order
of segmentation, the cell-lineage, the planes of cleavage,
the succession of nuclei, the effect of gravitation, the
constant and continuous orientation of successive stages,
the origin of the shell-gland and the mode of formation
of the shell, etc., but can not discuss such subjects here.
I may briefly state, however, that I believe Brooks failed
to observe the shell-gland, in his original work, and at
one particular stage mistook the relation of the shell-
valves to the blastopore which made it necessary to re-
verse his orientation of the embryo—hence his use of the
terms dorsal and ventral are misleading. The polar
bodies are dorsal at first—later, if they persist, they may
become displaced anteriorly. The blastopore is ventral,
the velum anterior, the shell-gland dorsal, the mouth
ventral. There is no foot, nor rudiment of it, in pre-
conchiferous stages.

I have found conchiferous young of the British Colum-
bia oyster retained within the parent’s shell until their
own minute shells were .138 mm. in length. I believe

40 THE AMERICAN NATURALIST [ Vor. XLVI

they remain longer, for, according to Möbius, the young
of the European oyster leaves the parent at a size of .15
to .18 mm. (Horst gives .16 mm.; Huxley 459 inch). I
have taken larve of O. lurida in plankton (identified by
comparison with those from a parent, and also by the
structure, shape and size) of a length of .165 mm. as well
as different larger sizes. They still had a straight-hinge
line of half the length of the shell—unlike the O. virginica
which at this size is already passing into the umbo-stage
and with a much shorter hinge-line. The larve of O.
lurida are not pink or brown but have five or six dark
blotches in the region of the liver and in the velum, in
contrast to the general light shade.of the rest of the
animal.

THE EFFECTS OF ALCOHOL NOT INHERITED
IN HYDATINA SENTA

DR. D. D. WHITNEY

WESLEYAN UNIVERSITY

Many experiments have been performed and much
published concerning the effects of alcohol upon living
organisms. Hodge, Calkins, Lieb, Woodruff, Estabrook,
Matheny, and others, have observed its influence on the
rate of growth and reproduction in certain unicelluar
organisms. Abbott, Hodge and others have carried on
some experiments with mammals by which they have
demonstrated that the resistance to certain bacterial in-
fections is lowered by the influence of alcohol. Hunt and
Woodruff found an increase of susceptibility to certain
poisons in the animals subjected to alcohol. Abel and
Welch have summarized in general the pharmacological
action and the pathological effects of aleohol upon man
and some of the other mammals.’

Stockard has produced abnormal fish embryos and
Féré has produced abnormal chick embryos by the use of
aleohol, while Hodge, Newman, Sullivan and others have
demonstrated the harmful influence of aleohol upon the
embryos of mammals and man during pregnancy.

The evidence taken altogether with a few exceptions
shows that when living organisms in any stage of their
life are subjected to alcohol in appreciable quantities eaer
are as a whole or in part unfavorably affected by it.

In nearly all of the previous work observations have
been made especially upon the organisms themselves
which have been directly subjected to the influence of
alcohol at some stage of their life. As the harmful effects

1I am greatly indebted to Dr. F. E. Chidester for placing at my dis-
posal his bibliography and notes of his forthcoming paper, ‘‘Cyelopia in
Mammals,

41

42 THE AMERICAN NATURALIST (Vou. XLVI

of alcohol on the organisms subjected to its influence have
been so conclusively demonstrated, it seems desirable to
determine whether the offspring of alcoholic individuals
in the subsequent generations are normal or show any
of the weaknessess of their alcoholic ancestors. In other
words the problem is to find out whether the descendants
of alcoholic parents are in any way inferior to the normal
individuals of the species and, if so, for how many genera-
tions the weakness continues.

That the parental use of aleohol in human beings affects
some of the offspring in the first filial generation is un-
doubted by many observers, yet Pearson and Elderton
have recently shown that the school children of alcoholic
parents are as normal as the children of sober parents
in physique and intelligence. However, the results set
forth in this paper do not purport to have any relation-
ship with the effects of alcohol upon man and his descend-
ants.

While working with the rotifer, Hydatina senta, obser-
vations have been made which show that while alcohol
decreases the rate of reproduction and increases the sus-
ceptibility to copper sulphate, still these harmful effects
of alcohol disappear in the second generation after the
alcohol has been removed from the culture water. The
grandchildren show none of the alcoholic weaknessess of
the grandparent, but are as normal as the individuals
whose grandparents never were subjected to alcohol.

Hydatina senta can be readily reared and controlled in
the laboratory in the manner described in a former paper.
Alcohol can be added directly to the liquid medium in
which the animals live. A large amount of the liquid is
drawn through the mouth, indirectly by means of the pul-
sating bladder, into the alimentary canal, and the dialyz-
able parts pass through its walls into the body cavity and
then finally out through the excretory ducts to the exterior
of the body. In this way the animal is bathed both on the
outside and on the inside of the body by the solution in
which it is living. Consequently all internal parts and all

No. 541] EFFECTS OF ALCOHOL IN HYDATINA 43

organs of the animal are subjected to whatever dialyzable
chemical substance there may be in the solution.

The young females grow to maturity very rapidly and
lay eggs which develop and hatch within a few hours.
This extremely short life-cycle, from egg to egg in forty-
eight hours, more or less, makes this animal a very favor-
able form with which to work. Many generations can be
reared in a short time and as much information gained in
a few weeks as it would require years to obtain from
some of the other forms.

Experiments were first carried out to determine what
influence a } per cent., 4 per cent. and 1 per cent. alcoholic
solution had upon the race when it was subjected to it
continuously for many successive generations. Precau-
tions were taken to have all conditions, excluding the alco-
holic conditions, in each generation exactly identical.
The experiments were conducted in the same room so
that the temperature was always uniform for each genera-
tion. The same amount of food culture from the same
jar was always mixed with the same amount of water or
with the same amounts of the various alcoholic solutions
thus making the proportion of food culture to the mixture
always alike. This mixture was then poured out into
watch glasses and one young female rotifer put into each
glass. At the end of forty-eight hours the young female
had matured, layed eggs some of which had hatched, and
young daughter-females would be found swimming in the
dish. One of these daughter-females was isolated to start
the next generation in the same manner as the mother was
originally isolated. This was continued for twenty-eight
consecutive generations. The twenty young females
which were isolated to form the first generation were the
grandchildren of the same grandmother, thus making the
control, and the other three strains or groups all start
originally from one female of one race. This was a very
vigorous race hatched from a winter egg which was taken
from a general mixed culture jar in the early spring.

Table I shows the detailed and summarized data of the

(Vou. XLVI

THE AMERICAN NATURALIST

44

uo;yes9Uay

gst erg s ee co ees z9 1 =e ror A pa bi
fee ee Sa cia} re. pala L og el ae Ge
wiria se Jaje ie Pees 8 oF ed fe ee ee ae
ove | st | » ot | op | + get | s | s H | tL e ee
tisi n aes o | | 9 g oz ees re Be:
eg | or | zo «| te | ¢ ot | ols | exer | s 7s
ove | st | » se | or | g sy | wi ¢ sor | e C En a
cs | I ee | 6t | 9 F+ | zz | s ¥6 Le s; : Ri 9 an a
O ta S ls l siai) rz zI e e Ra An a
pea ea | g zz ir] ¢ t ii na zor 1g St oe
0 ;} 8 | 8 vy | ew | ¢ c9 | g | >» 9 r oe ee
rites | ovo | ge | * co. Ali eg zz eee ae a
VF Ee 3 ial? S | o | ¢ 9: | ge Al ne he Bey
ee | SO te} ot dee | eee os Po ee eee A
Bapadeyo von. ato Supsdsyo vey on Suradeyo eeg ptt Satadsyo paonporg page
Jo ON ‘Ay — Pe JOON ‘AV — SR Jo ON “AY Bop na J0 ON 'AV | Bapids90 | unor owr,
uopnjog opoqoarv FT | uopnjogopoyoay go | uopmog onoyoory # ozo | l01}u09 |

dasn 'JTOHOO'IY dO LNOOWY AHL OL NOMMOAOU NI NOLLONGOUdAY
40 ALVY FHL SUVILA SNOMLVYANA) FDAISSHOONG NO 'IOHOOTY dO TONAATANI SQOONILNO) WHL LYHL DNIMOHS

I WTavi

45

EFFECTS OF ALCOHOL IN HYDATINA.

No. 541]

Fi | 281, ii| We | ver) Bet +009 | 99 Sit] eS | ot] Ter | **"**"***Areurmng
IIAXX rel 6.4 62) pote 1 tee £ er oe a
nax | 60 Jos f? se Elsi yor | z e e ae ia Sat
TARY} eee Ee oe ee ee jaja sa | e E cre ae Ay
AXX Ha 4A Es ia FL mie get | 89 a EN a
AIXX reS re als geia 69 69 w pee a ay
IIXX morj? st |e Ja lwr ela o | o ee ee ay
XX ils js | Jela si | aus est | 16 "EE Pp Bs Goa aa
IX ojo Jy ¢ Joris | ee lola | oy in gt
XE Oo ee ee ne: eiela z9 Tf ot eae as
XIX so le l3 I g 1.9 z'I T 9z > e E lane ea
IIAX oge Ile oran t gni | u A BE Beeps, bares Spy
mx | ot |e }a l stale l= | | e e a
IAX oe | et | i | 99'e eee p io ia i o eh et ee ae
AX cet. | 8 | soe i5 3 w ala 9 og eo AE n ae

sniega | Ce RE nino | HE T a, | eae a Sido ee paierai
EEE a. JOON ‘AV ae Pia L jo ON AV a seni J0 ON “AV pg ap 10 "ON “AY | 3updsyo Zunox ti

TOL NOS IOOF % T | aor NOS oTLOyOoTY % e'o ~ dor ntog IFoo y % 80 | 101}u09

(panuyuog) I ATAVAL

46 THE AMERICAN NATURALIST (Vou. XLVI

observations made upon the twenty-eight generations
while they were subjected to the influence of the alcohol.
One can not compare strictly the number of individuals in
the different generations because of the changed condi-
tions, especially of the temperature, and in some instances
the length of time between the generations. However,
the ratios between the strains or groups in all the genera-
tions may be compared and will show general uniformity.
The first few generations of the + per cent., and the 4 per
cent. alcoholic strains show a fluctuation in the rate of
reproduction above and below that of the control; but this
rate of reproduction in the 4 per cent. alcoholic strain
never rises to that of the control after the sixth genera-
tion. In the 4 per cent. alcoholic strain the rate of repro-
duction never rises to that of the control after the third
generation. In the 1 per cent. alcoholic strain the rate
of reproduction even in the first generation does not equal
that of the control. The summary shows in the average
number of offspring for each female that the alcoholic
strains differ in the rate of reproduction according to the
amount of alcohol used. The more alcohol used the lower
the rate of reproduction.

Another test to show the influence of the 1 per cent.
alcohol in this same series was made by removing some of
the individuals from the alcoholic solution and placing
them in a 1/14,000 G. M. copper sulphate solution? and —
comparing the resisting power, or the ability to live, of
this strain with that of the control when both were sub-
jected to the copper sulphate solution. Table II shows
the detailed data and Table III shows the summary. In
the control 96.8 per cent. of the individuals lived forty-
eight hours and produced young, while only 15 per cent.
of the individuals taken from the 1 per cent. aleoholic
strain in the XIII-XV generations lived forty-eight hours
and produced young. This shows that the susceptibility
to copper sulphate is greatly increased by the alcohol.

* Various solutions of copper sulphate were tried and the one employed

was selected because it appeared to be of the maximum strength which the
control could withstand.

No. 541]

TABLE II

EFFECTS OF ALCOHOL IN HYDATINA

47

SHOWING THE LOWER RESISTING POWER TO COPPER SULPHATE OF FEMALES
R

|

EARED THIRTEEN . TO

FIFTEEN GENERATIONS IN A 1 PER CENT

. ALCO-

HOLIC SOLUTION, AND ALSO SHOWING THAT THE RESISTING POWER
HAS BEEN REESTABLISHED IN THE SECOND GENERATION AFTER

REMOVED

THE ALCOHOL HAS BEEN
(See Table III for Summary)

©
g os yahoo G. M. Copper Sulphate Solution
E PE
= = pe
2 Ree 24 Hours 36 Hours 48 Hours
| E
1 | Control 5| Alive Alive + young | Alive + young
Second water generation 5| Alive Alive + young | Alive + young
1 v alcohol 5| Aliv All dead
2 | Control 5| Alive Alive + young | Alive + young
Second Faroes generation 5| Alive Alive + young | Alive + young
Sine 5| 2 dead ll dea
3 10| Alive Alive + young | Alive + young
Second water generation | 10 ive Alive + young Alive + young
4 ontr 5| Alive Alive + young | Alive + young
a water generation 5| Alive e + yo Alive + young
5 | Control 10| Alive Alive + young ve + young
tear Habeo generation | 10) Alive e + young | Alive + young
6 5| Alive ive live + young
Second by gl generation 5| Alive 2 dead 3 alive + young
1%a 5| Alive 4 dead 1 dea
7 | Contr so 5| Aliv Alive Alive + young
Second water generation 5| Alive Alive Ae F zomg
1 % alcohol 5| Alive 4 dead
8 | Control 5| Alive Alive + young | Alive g ges me
ng Araind generation 5| Alive Alive + ss Alive + y
A Jaj | 5| 3 dead
9 5| Alive | Alive ry young | Alive + young
ial water generation 5| Alive | Alive + young | Alive + young
1 % alcohol 5| Alive Alive but in | 1 dead ers
poorer condi- nearly
ion. Fe Fewer young
aÁ
10 | Control 10/ Alive Alive + young | Alive + young
Second water generation | 10| Alive ive + young | Alive + young
1 % alcohol 10| 5dead | 5 dead, fewer 6 dead +fewer
5 young young
11 | Control 10; Alive Alive + young | Alive + young
Second water generation | 10| Alive Alive + young Alive + young
12 | Control 10| Alive Alive + young Alive + young —
Second water generation | 10| Aliv Alive + young | Alive + young
1 % alcohol 10) Alive 6 +young
13 | Control 10) Alive | Alive + young | Alive + young ©
Second water generation | 10| Alive Alive + young | Alive + young
1 % alcohol 10! 3 dead 5 dead Ty
14 | Control 20. 4 dead 4 dead 4 dead + young
Second water generation | 20) 3 dead 3 dead - 3 dead + young
1 % alcohol 20 19 dead All dead
15 ntro! 10| Alive Alive Alive + young
Second water generation | 10| 6 dead 6 dead 6 dead, young
1 % aleohol 10) All dead

48 THE AMERICAN NATURALIST (Vou. XLVI

On July 1, these four strains of rotifers were carried to
Woods Hole, Mass. Owing to the high temperature diffi-
culty was experienced in growing proper food cultures
and consequently by July 4 many of the animals had died
and those that had survived were in a very bad condition
and had very few offspring. It is interesting to note that
more of the animals in the alcoholic strains died at this
time in the twenty-eighth generation than in the control
strain. The experiments were discontinued on account of
these unfavorable conditions.

TABLE III

SHOWING SUMMARY OF TABLE IT

Copper Sulphate Solution

|
No. of | r Cen
Young No, p Disa in No. Died in| No. Died in St End of Living ar
Females | Hours | 36 Hours | 48 hours Dr pwa
Isolated | | : ts Toate
Control ....... H |. 4 4 96.8
Second water |
generation ... 125 | 9 11 He 91.2
1% aleohol.... 1900 -1 52 78

The data in these three tables seem to show that alcohol
from 4 per cent. to 1 per cent. has a decided influence in
lowering the rate of reproduction ‘and also in lowering
the power of resisting copper sulphate in the individuals
of the 1 per cent. alcoholic strains. Presumably the
resisting power to copper sulphate of the 4 per cent. and
the 4 per cent. aleoholic strains was similarly lowered,
but this was not determined.

This decrease of the reproduction rate and the in-
creased susceptibility to copper sulphate can be consid-
ered as an indication that the ‘‘general vitality’’ of the
race had been lowered in that the individuals were much
inferior to the control individuals in their ability to cope
with adverse conditions and to leave offspring with which
to continue the race.

Since it is shown that alcohol decreases the ‘‘general
vitality’’ of these animals the condition of their offspring
now remains to be considered. Table IV gives the de-

No. 541] EFFECTS OF ALCOHOL IN HYDATINA 49

tailed and summarized data of experiments showing the
comparative reproduction rates of the offspring from the
three alcoholic strains between generations XI and gen-
eration XXIII, special emphasis being laid upon the off-
spring from the parents in the 1 per cent. alcoholic strain,
and the reproduction rate of the control or normal strain.
Young females were isolated from the three alcoholic
strains placed in media containing no aleohol and reared
two generations parallel to the alcoholic strains. In this
way they were under exactly the same conditions as the
control strain. The isolations of young females from the
ł per cent. and the $ per cent. aleoholic strains were dis-
continued after a few experiments and the time devoted
to experiments with the 1 per cent. strain. As this strain
in Table I showed the lowest reproduction rate and was
decidedly susceptible to the influence of copper sulphate,
it was assumed to have suffered the most of any of the
three strains subjected to aleohol and therefore was con-
sidered to be the most favorable to show the effects of
alcohol upon the offspring. For the sake of clearness all
the data are so arranged in Table IV as to show the three
rates of reproduction in the same generation, of the con-
trol, alcoholic strains, and the aleoholie strain with the
alcohol removed. In the first water generation the young
females were isolated from the preceding alcoholic gen-
eration soon after hatching and reared in media contain-
ing no alcohol. Thus the formation of the egg from which
each young female hatched and all the embryonic develop-
ment occurred in the alcoholic solution. After being
transferred to culture water lacking alcohol they grew to
maturity and reproduced. This generation is called
Water Generation I. Some of the young daughter-fe-
males from Water Generation I were isolated to form
Water Generation II.

In comparing the rates of reproduction in the Water
Generation I with the rate of reproduction in the same
generation of the alcoholic strains it is seen that in all
eases the rate of reproduction is higher in the Water

THE AMERICAN NATURALIST [ Von. XLVI

50

II | XIX

1] IIAX

II IIAX

I IAX

I AX

E] AIX

II | ne

I | IIX
|

Li IX
|

= |

5

5

2 suor

5 | -gaouan opoqoory

z

F

A

|
z9 (S3 AEE e 9 i L4 I e Gi == amre 2 |
Pg IZT ¢|r01'2|9¢9)| Fe i A er e Ee B iE |
| '
|
8'8 [PF 919118) 9488 « g rA Be hast E e TR
|
9 2 | | 9S | SI ge 9 Br | E [99'S IT | $ | GE PIF 0 0
9 sz F Sot |} ¢ | P979 os P| FIZ |S | 9 GE] S| 9S {EI
GLP let Pi SS sea Se GiS ee x G 9L'g SI | FP | Se 196
ee |S Lee et os Pr i gS 6 OLS LiSé | S| HS et
9'9 ££ | bay es 6e ie ey SS et G |$3°9 SZ | ¥ Sah LI
GL PS | G GZ’ | SI] ¥ | PZO | g 8 o g nie ¥ | 9E1/89 |
|
et Be ee dee Beem > z =
ca | Biel = S “los | Bel 2/§ | of aE A S
se etga] 3 |e |_s/ Ba esgar] 8| | S| Ss ef Se) § |e
TR SS Bt] ata S jas Ye |9 5]® r af = aE gs SD Bx) 2 ® >
By ARETES R ER] Bz BEST gS E BS] oe REET ES |a
BS AE] Bz Brae | eb Sz En Es 5| 2z |%
: 3| Ela? gigs 6 | 3] SEST Bar: “EFE =
A 3g TE a 5 E LFS or’ A AE A | g 2. z
TOA ` ‘paaou) o | & $row 'poaom) oO 6 F| iam pasow) o §
suono | A | & yoqoy | R |Ê oy oqoy | RJA

104091 “3499 dod T

10409] V *}090 19d 9'0

[OOTY “4000 19d €7"0

: | TIGI ‘gT oun ‘aA
2 | 99 (82 17 { TT6T ‘TT oung “oa
TIGI ‘IT oung ‘aa
. : TIGI ‘6 əunpf ‘aA
| 6d 8/3 { II6T ‘4 oung ‘oa
II6T ‘4 unf ‘aa
1/98 &F ¢ { ars Seat ara &
ITIGI ‘p oung ‘aA
o} 9 (08/9 {| irere ome “oa
; TIGI ‘g oun ‘oA
s| z9 he |s | 116 ‘Te AUN ‘oat €
L161 ‘Te AB ‘aa
Se Trac ee ae oa
| TIGI ‘6g ABI “0A
IP Sere { TOT ‘26 AUI OAH Z
TT6T ‘23 AEN “IW”
g | vaiet |a {] Ter se ow vag 1
> wd
ʻA ©
a SiE
AERAR:
5| F 85) Ba lg]
Tn = Am ge »
eel a ia] E : 3
Srno E e 3
SE al? g g E
g g
101}u09

CTAONGY NATA
SVH TOHOOTYV AHL AV NOILVYANTO ANODU AHL NI AAAOLSAY SI NOILONAOYAJAY JO ALVY IVNWYON AHL LVHL PNIMOHY
AI ATAV LE

EFFECTS OF ALCOHOL IN HYDATINA 51

No. 541]

lam]
Laa)

SUOPJBIIUJE) 194V AA

S XXN |
ITXX IXX XIX 0L RUL- 26; L
‘ITAX ‘AX OIX |
XXS UX |
‘XX ‘IIIAX ‘TAX čte 981-90 29°T
“AIX TIX TX he
oe
AIXX __|68°8 [892 6z I
mxx |i9z e] er FT
IIXX sijoi L| #1
|
MIXE 0106 |6| —
| |
IIxx L 7 k} I
Ixx met pr e o
|
Xx let K a g'o
EEE wl
of | BE 4
28 RHES 28
Tz fog ae pole
pS -a5 24
suore REN Ae R
-19095 O1[OYooTy AT x Tg A
19438 M PIou ©
-9Y [OOO V R

104094V ‘4099 Jad T

| |
T PPL PEI ST
mas 6¢°9 ie
g P = j= =
7 ° m A-
L S s Pe Ge
|i f
a Gru tas
| |
ie. Ce Geet aoc
|
0 ? ae a
fe | — ae
T ae pl ol <
© = aS
(æ) Z% =
MEFR:
A ESR |Z | =
Bp |2@5| 40 3 ope ro
a (Sel oz ARISE
eena | mae
slas a| 3| F
& fd
= $|101BA\ “peaowu
£ -ƏH 10Q0]y

96'I

N

<+

peonporg Burdsyo | = Eh S

Sutids
-HO JO ‘ON o3¥s0ay

68T

{Oqoo[V “3Ua) 13d £0

g8

po BOs]
soyeme.y Sunox

4 1901| ST

Z icīTT 91I
TEE:

< ©
28 | Fine
Pe (Faj sn
fa RR
a4 |e
R’ F àE
CAL U g
aj M 'pəsom
-34 [04001 V

|

Suads
-PO J0 ‘ON o3t190ay

£01

66

peenpoig Zuladsyo

66/062 (She | Areuruing
8z 88°01 SOF sf S1euumg
we TT6T ‘93 ounf “Wy
s | 6° leo |ot{ | ther ‘ve omn "IEY
T16I ‘FZ OUNL WY
a aa N
TIGI ‘Pz OUNE “Wy
yo s C a nr AY
Lly TIGI ‘gg əunf ‘9AN
gorte |2 {| Tret ee ounp "w'y
a TIGI ‘Ig ounf “oa
¢ | Z'STi16 |S { a a,
. TIGI ‘ST oung “oA
¢ | oF ite |¢ { ETGT “OT sung on
. TIGI ‘OT ƏUNf ‘9A
e| eo je js { TI6T ‘FI oune aag) 9
id Be ee
<4 c
© R =]
S aS EE
=z TR ES) py
an| Z. a
e 52 în es
Bo) * 2 s aS 3
zz] e| i F ou d
8 5
101}u00

104091 Y "309 19q £70

(panwiquog) AT T'IAYL

52 THE AMERICAN NATURALIST (Vou. XLVI

Generation I than it is in any of the corresponding
aleoholic strains. However, it never reaches that of the
control. In the Water Generation II the rate of repro-
duction is much higher than it is in Water Generation I
and equals the reproductive rate of the control. This
demonstrates that one of the ill effects of alcohol is par-
tially eliminated in the first generation and is entirely
eliminated in the second generation after the alcohol has
been removed.

Very probably the rate of reproduction of Water Gen-
eration I, which is lower than that of the control, is due
to the fact that the females in their embryonic develop-
ment were subjected to the influence of alcohol from which
they never fully recovered after they were transferred
to water solutions containing no alcohol. They were thus
perhaps influenced while in the growth and maturation
stages of the egg inside the body of the mother or, after
the egg was laid in the alcoholic solution, in the embry-
onic stages which occurred inside the egg membrane be-
fore hatching. Soon after hatching the females were put
into the solutions free from alcohol.

This transmission of a low reproduction rate to Water
Generation I is in reality not a hereditary transmission
of a characteristic, but is probably the result of the direct
influence of the aleohol upon the mother during her em-
`- bryonic development. —

Tables II and III show the effect of copper sulphate
upon Water Generation II. Of the individuals tested
91.2 per cent. lived forty-eight hours and produced young
in the copper solution. This is only about 5 per cent. less
than the number of individuals of the control living and
producing young for the same length of time in the cop-
per solution. It can be concluded from these observa-
tions that the resisting power to copper sulphate has
been restored practically to normal and that these indi-
viduals are no more susceptible to its influence than the
individuals of the control.

Billings states to the Committee of Fifty in his report,

No.541] EFFECTS OF ALCOHOL IN HYDATINA 53

which is based on over five thousand reports of cases of
insanity, that : ‘‘ Inherited tendency to insanity, due to the
use of liquor by parents, is reported in one hundred and
twenty-two cases ... while six cases were ascribed to
the intemperance of the grandparents. These statistics
must be received with caution as showing possibilities
rather than as definite evidence. To prove that the
insanity of one generation is due to alcoholic excess of a
previous generation, and is not merely a coincidence,
requires that other causes of degeneration shall be care-
fully studied, and duly allowed for.’’

It is, however, evident from the six cases reported that
some, at least, of the medical examiners believe in the
transmission of alcoholic weaknessess from grandparents
to grandchildren.

Bunge, from an investigation extending over two thou-
sand families, found that chronic alcoholic poisoning in
the father was the chief cause of the daughter’s inability
to suckle and that this inability was not usually recovered
from in subsequent generations. These results have been
severely criticized by Bluhm and their validity ques-
tioned.

Mariet and Cambemale gave considerable quantities of
alcohol to a female dog during the last week of her preg-
nancy. She gave birth to a litter of seven puppies, of
which four were dead, two apparently healthy but men-
tally backward, and one, No. 7, both physically and men-
tally backward. No.7 was a female and grew to maturity
free from the influence of alcohol and mated with an
apparently healthy dog. All of the puppies of her first
litter were abnormal to such a degree that they were con-
sidered worthless. One had club feet and a clefted
palate, another had a eonspicuous ductus Botalli, and
another developed muscular atrophy in its hind legs.

If these observations and interpretations are correct
they may demonstrate either the same fact that is shown
in Water Generation I of Table IV, namely, that when a
mother is subjected to the influence of alcohol during her

54 THE AMERICAN NATURALIST (Vou. XLVI

own embryonic development she shows some sign of
weakness at her period of reproduction, or, that the
grandchildren are affected by the influence of alcohol
upon their grandparents. However, only one experi-
ment alone like the above is not sufficient to prove any-
thing, and furthermore, Hodge in speaking of dogs says:
‘We do not attach much importance to the greater per-
centage of deformity, since this is of somewhat common
occurrence in kennels.”’

If the transmission of an alcoholic weakness to subse-
quent generations is possible in any living organism, it
ought to be actually demonstrated in some manner, but
if it is a delusion, the sooner it is dispelled the better.
These experiments with Hydatina senta are an attempt
to determine, in one race of animals only, whether cer-
tain aleoholic weaknesses are truly hereditary and the
evidence found is negative.

It by no means follows that these results would be
found to be true in man. Alcohol primarily affects the
nervous system and may have a very different action on
the highly organized nervous system of man than it does
on the lowly organized Hydatina, whose nervous system
is extremely simple. Furthermore, the germ substance in
man is probably very different from the germ substance
in the rotifer and alcohol might have a very different
effect upon it.

SuMMARY

1. Four strains of parthenogenetic rotifers originally
descended from the same female were observed throughout
twenty-eight successive generations. One strain was kept
as a control and the other three strains were kept in a
+ per cent., a 4 per cent. and a 1 per cent. solution of
alcohol. The rate of reproduction was lower in the alco-
holic strains than in the control and it was proportion-
ally lowered according to the amount of aleohol used.

2. The individuals of the 1 per cent. alcoholic strain in
the XI-XV generations showed a decidedly increased
susceptibility to copper sulphate.

No. 541] EFFECTS OF ALCOHOL IN HYDATINA 55

3. When the alcohol was removed in generations XI-
XXII, the rate of reproduction increased noticeably in
the first generation and in the second generation the
reproduction rate equaled that of the control.

4. Individuals of the second generation after the alco-
hol had been removed were no more susceptible to copper
sulphate than individuals which had never been sub-
hres to alcohol.

. The general conclusion is that alcohol in 4 per cent.,

4 per cent., and 1 per cent. solutions is danaa to this
race of rotifers when it is subjected to it continuously for
many generations. The weaknesses developed by the
parental use of alcohol are partially eliminated in the
first generation after the alcohol has been removed, and
practically completely eliminated at the end of the second
generation after the alcohol has been removed. In other
words, the grandchildren possess none of the defects
caused by alcohol in the grandparents.

6. These results in general show that alcohol in the
percentages used affects only the somatic tissues of the
animal, and if they are subjected to its influences indefi-
nitely, generation after generation, the race would prob-
ably become extinct because of its ‘‘lowered resistance
power’’ to unfavorable conditions. However, if the alco-
hol is removed it is possible for the race to recover and to
regain its normal condition in two generations, thus
showing that the germ substance is not permanently
affected by the aleohol.

BIBLIOGRAPHY

Abbott, A. C. 1896. The Influence of Acute Alcoholism upon the Vital
EED of Rabbits to Infection. Jeur. Exp. d., Vol.

Abel, J. J. 1903. A Critical Review of the Pharmacon Aetion of
Ethyl Alcohol, with a Statement of the Relative Toxicity of the Con-
stituents of Alcoholic arenge Physiological Aspects of the Liquor
Problem, Vol. 2. Houghton, Mifflin and Co., Boston and New York.

Ballantyne, J. W. 1902. pen e Pathology and Hygiene. William
Green and Sons, Edinburgh.

Billings, J. S. 1903. Relations of Drink Habits to Insanity. Physiological
Aspects of the Liquor Problem, Vol. 1. Houghton, Mifflin & Co., Boston
and New York.

56 THE AMERICAN NATURALIST (Vou. XLVI

Bluhm, Agnes. 1908. Die Stillungsnott. Zeitschrift fiir Sojiale Medizin.
1909. Sexual Probleme.
Bunge. 1907. Die Zunehmende Unfähigkeit der Frauen ihre Kinder zu

Calkins, G. N., sa Lieb, C. C. 1902. Studies on the Life-history of Pro-
tozoa—II. e Effects of Stimuli on the Life-cycle of Paramecium
caudatum. pos fiir Protistenkunde, Vol. I.

Estabrook, A. H. 1910. Effects of Chemicals on Growth in Paramecium.
Jour. Exp. Zool., Vol. 8.

Féré, Ch. 1899. Cinquantenaire de la Société de Biologie. Vol. Jubilaire.

Hodge, C. F. 1903. The Influence of Alcohol on Growth and Development.
Physiological Aspects of the Liquor Problem, Vol. 1. Houghton,
Mifflin & Co., Boston and New York.

Hunt, R. 1907. Studies on Experimental Alcoholism. Bull. No. 33, Hyg.
lá b., U. S. Pub. Health and Mar.-Hosp. Serv., Washington

Mairet, Ac et Combemale. 1888. Influence PRLR de Vileohol « sur la
descen iihi, C. R. de Se. de L’ Acad. des Se. Paris. Vol. 106, pp. 667-

Matheny, W. A. 1910. Effects of Alcohol on the Life-cycle of Paramecium.
Jour. Exp. Zool., Vol. 8.

Newman, G. 1906. Infant Mortality, pp. 72-77.

Piati. K., and Elderton, E. . A First Study of the Influence of
Parental Alcoholism on the Saiga and Ability of the Offspring.
London, Dulau and Co., 37 Soho Square, W.

1910. A Second Study of the tease of Parental Alcoholism on the
Physique and Ability of the Offspring. London, Dulau and Co., 37
Soho Square, W.

Pearson, K. 1911. Questions on the Sy and of the Fray. No. III. Lon-
don, Dulau ay Co., 37 Soho Squar

Stockard, C. R. 1910. The Influence a Aleohol and other Anesthetics on
Embryonie Development. Am. Jour. of Anat., Vol. 10.

Sullivan, W. C. . A Note on the Influence of Maternal Inebriety on

the spr " Jour. cf Mental 8c., Vol. 45, pp. 489-503.

1906. Aleoholis

Welch, W. H. n06e: The Pathological Effects of Alcohol. pego
Aspects of the Liquor Problem, Vol. 2. Boston and New
Houghton, Mifflin & Co.

Whitney, D. D. 1910.

bd
to
(a

e Indice = oa Conditions upon the Life

ol.
1911. The Poisonous Effee ects,of acini Beverages not Proportional to
their Puree Contents. Science, Vol. 3
Woodruff, L. 1908. Effects of Alcohol on ie Life-eyele of Infusoria.
Biol. Bull., ‘Vol. 15.

ARY, 1912

FEBRU.

of
y

XLVI, NO. 542

VoL.

THE
AMERICAN

The American

MSS

Naturalist

intended for publication and books, etc., intended for ctr should be

sent to a Editor of THE AMERICAN NATURALIST, Garrison-on-Huds n, New York.

icles containing research work _bearing on the problems = ‘organi evolu-
tion are er wa welcome, and will be given preference in publica

rea reprints of contributions are supplied to authors hes of charge.

One hund
Further reprints will be supplied at c

ion

Subscription:
ao price h four

an postage
thirty-five cents,

dollars a

s and advertisements should be sent - the yepe
year. Foreign pos

twenty-five cents additional.

The advertising rates are Four Dollars for a page.

The
age is fifty cents and
The ort for single prer is

THE SCIENCE PRESS

Lancaster, Pa.

Garrison, N. Y.

NEW YORK: Sub-Station 84

Entered as second-class matter, April 2, 1908, at

t Lancaster, Pa., under the Act ot

Congress of March 8, 1879.

THE BULLETIN —For bargains in Ethnolograph-
ical and Pre-historic Specimens. Books on Natural
= History, Science, Travel, Voyages, etc. See THE
_ BULLETIN post free for 3 cent stamp.
4 Duke St., Adeiphi—London—England

TENTH EDITION.
THE MI CROSC COPE;

an introduction to

by OTOR TARE GAGE, OF Cornell U apts Over ;
dys miea pa e and above 250 a T ape

pared edition. Price $2.00, postpaid. ;
COMSTOCK PUBLISHING CO., Ithaca, N. Y.

= ergs EGGS

w if: H ROSENBERG,

} G : uates. ‘Daily Field excursions.

| Botany,

A U n E E, De ae 3

University of Colorado

Mountain Laboratory
Fourth summer session June 24 to Au-
1912, Courses in Field Biology
(animal a
Biology of Pind and Streams. i |

r d građ-
o.

ee OER UE

The laboratory is at Tolland, Colo- ii t

| | the Rocky Mountains at an altitude of |

| 8889 feet. Climate cool and healthful. pee |

and plant), Ecology, Systematic l

THE
AMERICAN NATURALIST

Vou. XLVI February, 1912 No. 542

SOME ASPECTS OF CYTOLOGY IN RELATION
TO THE STUDY OF GENETICS!

PROFESSOR EDMUND B. WILSON

CoLUMBIA UNIVERSITY

THe consideration of genetic problems from the stand-
point of cytological research sometimes encounters a
certain opposition or prejudice which seems to me to be
due to a misunderstanding of the position that is actu-
ally held by many cytologists. It probably grows pri-
marily out of a conviction that the heredity of particular
traits is not to be explained by referring them to the
operation of particular cell-elements or ‘‘determiners,”’
but results from an activity of the whole cell-system, or
of the whole organism, regarded as a unit. With this
view, as will appear, I am essentially in agreement. In
the second place, the opposition is a kind of reaction or
protest against the theory of pangens or biophores and
the too elaborate logical constructions that have been
built upon it, especially by Weismann. I also consider
this theory untenable, or at least unnecessary. I will
therefore attempt to outline a point of view from which
I think genetic problems may reasonably be regarded
from the standpoint of the cytologist.

The most essential result of modern genetic inquiry I
take to be the proof of the independence of the so-called

1A paper read before the American Society of Naturalists at the
Princeton meeting, December 28, 1911.

57

58 THE AMERICAN NATURALIST [ Vou. XLVI

‘<unit-characters’’—that is to say, that they may be inde-
pendently combined, disassociated and re-combined in
many different ways. The independence of these char-
acters often seems to be complete; more rarely it is
limited by definite phenomena of ‘‘coupling’’ or ‘‘repul-
sion.’’ The interesting facts recently brought to light by
Bateson and Punnett in case of certain unit-characters
in plants, and by Morgan in case of sex-limited char-
acters in flies, demonstrate that coupling or repulsion,
as exhibited in the F, generation, are a consequence of
an original association or separation in the grandparental
gametes. In the cases referred to, characters that enter
the F, zygote in the same gamete tend to ‘‘couple’’ (re-
main in association) in the gametes produced by this gen-
eration; while if these same characters enter the F,
zygote in different gametes they tend to ‘‘repel’’ each
other (remain separate) in the ensuing gamete-forma-
tion. This is almost a proof that the factors for coupled
characters are borne by a common vehicle or substratum
in the germ-cell, while in repulsion they are borne by
separate ones. Not alone such facts, but the whole
history of unit-characters points unmistakably to the
conclusion that they are in some way connected with
material substances or bodies; and that it is the com-
binations, disassociations and recombinations of the latter
that explain the corresponding behavior of the former.
For example, in sex-limited heredity the peculiar linkage
of certain unit-characters with sex becomes readily intel-
ligible, as several writers have recently pointed out, if
factors necessary for the production of these characters
are associated in the same material body with a factor
that plays a certain necessary rôle in the production of
sex. In this particular case, as it happens, we are actu-
ally able to see a material body (the ‘‘X-chromosome’’)
which undergoes precisely such a mode of distribution
with respect to sex and sex-limited characters as is de-
manded by the hypothesis. The question must here be
squarely faced, in a very real and concrete form, whether

No. 542] SOME ASPECTS OF CYTOLOGY 59

unit-characters are in fact dependent upon separate ma-
terial bodies or substances, and whether the chromo-
somes can be regarded as such bodies or the carriers of
such substances.

Without entering upon the evidence in detail I shall
take it for granted that both these questions may be an-
swered in the affirmative. Accepting this (if only for
the sake of argument) how can such a conclusion be
reconciled with the ‘‘action of the whole,’’ and how form-
ulated so as to escape the pangen-theory? The latter
theory has fallen into discredit for two reasons. One is
because of the quasi-metaphysical assumption that the
‘‘physical bases” or ‘‘determiners’’ of unit-characters
are organized, self-propagating germs. Let us lay this
assumption altogether aside as incapable of verification,
and think of the ‘‘determiners’’ in a more vague way
only as specific chemical entities of some kind. The
second and more serious objection lies against the notion
that the ‘‘determiners”’ are to be regarded as ‘‘bearers’”’
of the corresponding characters. This is a fundamental
error, as may be made clear, I think, by a specific illus-
tration. It has been proved that many ‘‘unit-characters”’
‘are not units, but require for their production the co-
operation of several factors, as is shown with especial
clearness in the heredity of color. In such cases, as is
now generally recognized, we should not speak of ‘‘unit-
characters” but of unit-factors. Different ‘‘unit-char-
acters’? come into view as particular unit-factors are
added to or subtracted from a given combination in the
zygote. The factor for gray color in mice, to take a
familiar example, operates by inducing a reaction of the
germ that can only take place in the presence of several
antecedent color-factors lower in the scale ; and the latter,
in turn, are only operative in the presence of still another
factor that is necessary to the production of any color.
The first and most obvious suggestion given by such
facts is that what is added to or subtracted from a given
combination or state of equilibrium in the zygote is some

60 THE AMERICAN NATURALIST (Vou. XLVI

kind of chemical entity which induces a specific reaction
of the germ sooner or later in its development. But
beyond this it is perfectly evident that however far back-
wards we may* follow such a series of unit-factors, at
every stage they play their specific rôle only in so far as
they form part of a still more general apparatus of
ontogenetic reaction that is constituted by the organism
as a whole. The whole of this apparatus, the entire
germinal complex, is directly or indirectly involved in
the production of every character. We find it convenient,
indeed necessary, to treat particular factors of reaction
(i. e., the ‘‘determiners’’) as if they were concrete and
separate things. Such, in fact, they may be, as already
indicated; but when we speak of them as ‘‘bearers’’ of
the corresponding characters, we are using a figure of
speech that may be highly misleading. The reactions
(characters) which they call forth are not ‘‘borne by”’
them. They appear as responses of the germinal organi-
zation operating as a unit-system; and it is to this sys-
tem as a whole that every character belongs, or by which
it is ‘‘borne’’—if indeed we may permit ourselves to
employ the latter expression at all.

The point of view thus indicated may, I think, be made
entirely clear by a chemical illustration. A number of
writers, among them Adami, Guyer and Kossel, have
of late called attention to the parallel that may be drawn
between the physical basis of heredity and the complex
molecular groups of the proteins and other organic com-
pounds. It is a most suggestive one, though it is not to
be taken too literally—indeed I shall employ it only as a
kind of allegory or illustrative fiction. No one can fail
to be struck with the really remarkable analogy, in
method and in results, between the precedure of modern
genetic experiment and that of modern organic chem-
istry. Just as the qualities of a particular protein may
be definitely altered by the addition, subtraction or the
substitution one for another of particular side-chains or
molecular ‘‘Bausteine,’’ so the addition, subtraction or

No. 542] SOME ASPECTS OF CYTOLOGY 61

substitution of particular ‘‘determiners’’ or ‘‘factors’’
in the zygote calls forth specific responses that lead to
the production of corresponding characters. The reason-
ing that applies to the first of these cases seems equally
applicable to the second. No one, I suppose, would hold
in the first case that the particular molecular groups or
‘*Bausteine’’ concerned in the change are ‘‘bearers of’’
(i. e., are alone responsible for) the resulting new quali-
ties. The qualities of any protein, as Kossel has recently
urged, belong to the molecule as a whole, and are not to
be regarded as the sum of the qualities of its constituent
‘*Bausteine.’? Why should we regard in a different light
the ‘‘determiners’’ (chemical substances?) concerned in
the second case? They are, clearly, not to be regarded
as ‘‘bearers’’ or ‘‘physical bases’’ of the characters
which depend upon their presence or absence. They are,
I repeat, only differential factors of ontogenetic reac-
tions that belong to the germ considered as a whole or
unit-system.

In all this I am but expressing what I believe to be the
point of view of many recent writers on genetic prob-
lems; but what I desire to emphasize is that the prob-
lems of cytology should be regarded from the same point
of view. It is our task to see whether an apparatus
of ontogenetic response can be discovered in the cell that
fits with such a conception of the general process of de-
termination. Is there cytological evidence of the exist-
ence in the germ-cell of such specific factors of reaction
as I have referred to—in the nucleus, in the protoplasm,
or in both? I think that observation and experiment
alike have produced such evidence. Such experiments as
those of Boveri on multipolar mitosis, and of Herbst and
of Baltzer on the relations of the chromosomes in recip-
rocal crosses in sea-urchins have almost conclusively
shown that the chromatin does in fact play a causative
role in determination. Observation has gradually estab-
lished the existence of a complex process of segregation
and distribution of the nuclear materials in karyokinesis.

62 THE AMERICAN NATURALIST [ Vou. XLVI

maturation and fertilization, that shows a most striking
parallel to that of the factors of determination. That a
somewhat similar apparatus of distribution may exist
also in the protoplasm of the cell is indicated both by
recent observations on the chondriosomes, or plasto-
somes, and by earlier results on experimental embry-
ology. If in the brief discussion that follows I confine
myself to certain phenomena of the nucleus it is because
the history of the chromatin is more fully and accurately
known.

The progress of cytological inquiry tends steadily, I
think, to sustain the view first clearly formulated by Wil-
helm Roux that the nucleus contains many different sub-
stances which undergo orderly groupings and distribu-
tions in the karyokinetic phenomena. These processes
are in some measure made visible to us in the formation
of spireme-threads, in their history in cell-division, and
in the still imperfectly understood but perfectly definite
events of synapsis and reduction. In his well-known
paper on the significance of the karyokinetic figure, pub-
lished in 1883, Roux maintained that the nucleus is the
seat of many different ‘‘qualities.’? He committed him-
self to no definite view as to what these ‘‘qualities”’
really are; but the implication is not far to seek that they ~
have a chemico-physical basis and may be different chem-
ical substances. On this general assumption he based his
well-known interpretation of karyokinesis, of which the
essential postulate was that the ‘‘qualities’’ (substances?)
become arranged in linear series in the spireme-thread,
and by longitudinal splitting of the thread may thus be
equally divided (or otherwise definitely distributed) to
the daughter-nuclei. I believe that many of the later
advances of cytology lend additional support to this con-
ception.

1. In the first place, the evidence gives strong ground
for the conclusion that the chromosomes, to which the
spireme-thread gives rise, are not homogeneous, but com-
pound bodies. I do not here refer to the well-known fact

No. 542] SOME ASPECTS OF CYTOLOGY 63

that the spireme-threads often consist of linear series of
granules. I have in mind the fact that the number and
size-relations of the chromosomes often differ materially
as between different species, even nearly related ones, and
that in at least one case (that of the X-chromosome) it is
an established fact that a particular chromosome which
in some species is a single body may be represented in
other species by two or more components that sometimes
show constant and characteristic differences of size. The
natural interpretation of this fact is that the chromo-
somes are compound bodies, consisting of different con-
stituents which undergo different modes of segregation
in different species. We may here find a rational expla-
nation, both of sex-limited heredity, as I have elsewhere
indicated, and of other kinds of coupling.

2. Recent studies on karyokinesis and maturation em-
phasize anew the importance of the mitotic transforma-
tion of the chromatin-substance, and add weight to
Roux’s original interpretation of this phenomenon.
Nothing in recent cytological research is more interest-
ing than the discovery by Bonnevie, Pinney, Davis, and
others that new chromosomes may arise within the old
ones in the form of tightly coiled or convoluted threads,
which uncoil or unravel to form separate spireme-threads.
In karyokinetic division these threads may be formed
already inside the telophase-chromosomes of the preced-
ing division, as was discovered by Bonnevie. In other
cases, an example of which is given by certain Orthoptera
first reported upon by Miss Pinney, they are first visible
in the early prophases, when they are seen uncoiling
from massive bodies formed from the old chromosomes
and equal in number to them. The same remarkable
process occurs in the early auxocytes, as the chromosomes
are preparing for conjugation in synapsis, as has been
shown particularly by Davis, whose observations, like
those of Pinney, I have recently been able to confirm
and extend. In many insects the presynaptic spireme-
threads do not arise, as has often been described in other

64 THE AMERICAN NATURALIST [Vou. XLVI

forms, directly from a chromatin-network. They arise
from massive bodies, each of which resolves itself into a
closely convoluted thread which then uncoils before con-
jugation takes place. Why should chromosomes that are
already formed as massive bodies delay their division
or conjugation until so remarkable a redistribution of
their substance has taken place? It is not a necessary
condition of conjugation, as is proved by the case of both
the sex-chromosomes and the m-chromosomes, which do
in fact conjugate in the massive condition. All the facts
become intelligible in the light of Roux’s hypothesis that
the formation of the spireme-threads effects a linear
alignment of different constituents in preparation either
for division or for a definite type of association in pairs
in synapsis.

One of the most interesting applications of this view to
genetic phenomena is that suggested by Janssens in his
theory of the ‘‘chiasmatype,’’ which has recently been
applied by Morgan to the explanation of coupling and
repulsion. In the twisting together of the spireme
threads, either in synapsis or at a succeeding stage, and
the subsequent secondary splitting of the thread in one
plane is provided a very simple mechanical basis, both
for the free interchange of factors between the homol-
ogous chromosomes and for the phenomena of coupling
or repulsion, which are otherwise so difficult to compre-
hend. I do not maintain that this particular interpreta-
tion, or the more general one of Roux, is demonstrably
true, or that no other explanation can be found. I only
hold that they are legitimate conceptions which may be
tested by observation and experiment, and which must be
fully reckoned with as intelligible interpretations of the
facts before we can set these facts aside as utterly mys-
terious or as a meaningless by-play.

3. I would lastly recall the experimental proof by
Boveri that the chromosomes differ among themselves in
their physiological relation to development, and the cor-
responding cytological fact that they differ among them-

No. 542] SOME ASPECTS OF CYTOLOGY 65

selves also in respect to size, behavior, or both. In one
case only has it thus far been possible to demonstrate a
constant relation between particular chromosomes and
particular characters, namely in the case of sex and sex-
limited characters. It is true that in this case we are not
able to assert that the sex-chromosomes are the primary
determining cause of sex—indeed, there seems to be good
evidence to the contrary. Unless, however, we are pre-
pared to defend the proposition that the sex-chromosomes
are absolutely functionless we shall not, I think, escape
the conclusion that they form one of the factors in sex-
heredity.

I will not enter upon the analytical subtleties of the
problem whether the chromosomes, or the substances
that they contain, are permanent and self-propagating
elements or are merely temporary products of an unseen
underlying activity—whether they are causes, effects or
mere accompaniments of the specific reactions with which
they are somehow connected. These are fundamental
questions; and some of them can not yet be answered.
But we should not hesitate to adopt what seems likely to
be for the time the most reasonable and fruitful working
view. ‘‘Hypotheses,’’ said Pasteur, ‘‘come into our labo-
ratories by armfuls; they fill our registers with projected
experiments, they stimulate us to research—and that is
all.” In my view studies in this field are at the present
time most likely to be advanced by adopting the compara-
tively simple hypothesis that the nuclear substances are
actual factors of reaction by virtue of their specific chem-
ical properties; and I think that it has already helped us
to gain a clearer view of some of the most puzzling prob-
lems of genetics. But even if we adopt the opposite view
that the formation, segregation and distribution of these
substances are only signs or indices of what lies behind
them, we still have in this direction one of the most prom-
ising paths of approach to a study of the activities of the
germ-cells in heredity. .

It will perhaps be said that such conceptions of the |

66 THE AMERICAN NATURALIST [ Von. XLVI

nuclear organization as have here been indicated are both
vague and artificial. Vague and crude they undoubtedly
are; and so they will remain until we have far more
thoroughly explored a field of inquiry in which we must
for the present make a shift with crude ideas unless we
are content to work with no ideas at all. They may be
artificial too; but it appears to me that in this respect
they differ only in degree from the graphic formulas of
structural chemistry. The chemist does not hesitate to
picture definite topographical or spatial relations in the
complex organic molecule—symmetrical and asymmetrical
forms, cyclic or ring-formations, linear series, side-
chains and other such graphic constructions. It is by
their use that the whole science of organic chemistry has
been built up, and that such men as Emil Fischer and
Kossel have made nearly all of their advances in our
knowledge of those most complex of known organic com-
pounds the proteins. And these constructions are re-
garded by eniinent investigators as something more than
mere figurative expressions or symbols. They are taken
more literally as representations or models—rude, no
doubt, but as far as they go real—of the actual arrange-
ments in space of the various molecular groups or pro-
tein ‘‘Bausteine.’’ If therefore observation and experi-
ment lead the eytologist to postulate definite topograph-
ical relations among the nuclear substances, and if such
a conception help him to explain the results of genetic
studies, he finds himself in good company, even though
his present clumsy notions regarding the nuclear organi-
zation can as yet make no approach to the exact and ele-
gant constructions of the chemist.

The essential conclusion that is indicated by cytolog-
ical study of the nuclear substance is that it is an aggre-
gate of many different chemical components, which do
not constitute a mere mechanical mixture, but a complex
organic system, and which undergo perfectly ordered.
processes of segregation and distribution in the cycle of
cell-life. That these substances play some definite rôle

No. 542] SOME ASPECTS OF CYTOLOGY 67

in determination is not a mere assumption, but a conclu-
sion based upon direct cytological experiment, and one
that finds support in the results of modern chemical
research. Professor Kossel has recently said that every
peculiarity of the species and every occurrence affecting
the individual may be indicated by special combinations
of protein ‘‘Bausteine.’’ The point of view that has here
been indicated is entirely in accordance with such a con-
ception. The results of cytological inquiry fit with the
view that there are many such combinations in both
nucleus and protoplasm; and the interest of cytological
study lies in the fact that we can in some degree follow
out their modes of segregation and distribution with the
microscope. We are still utterly ignorant as to how these
processes are determined; and the more one studies them
the more one’s wonder grows. I would certainly be one
of the last to disparage the brilliant results that have
been attained through the prolonged and patient labors
of cytological observation and experiment. They stand,
I believe, among the most interesting and valuable
achievements of modern biology. But these studies have
as yet made no approach to their limit, and a vast unex-
plored field still lies before us. We may as well recognize
the fact that our present rude notions of cell-organiza-
tion have not yet progressed very far beyond the paleo-
lithic stage of culture; but they are of use in so far as
they help to open new points of view or to discover new
facts, whether in cytologic or in genetic inquiry. It seems
to me that in both regards they have already proved
worth while.

THE CORRELATION BETWEEN CHROMOSOMES
AND PARTICULAR CHARACTERS IN HY-
BRID ECHINOID LARV Æ’

PROFESSOR DAVID H. TENNENT

Bryn Mawr COLLEGE

Tue student of genetics who bases his researches on
‘the development of echinoderms has become quite accus-
tomed to the criticisms of his friends to the effect that his
work can have little point until he can show. them indi-
viduals of a second or a third generation, and he receives
these criticisms at their full value.

Our ignorance on this subject has been clearly defined.
First, little is known of the inheritance in the adult
hybrids of the characters of the parent species, and sec-
ond, it is not known that these hybrids become sexually
mature.

The ready acceptance of these ideas has blinded us, in
a measure, to the fact that for the solution of some of the
problems of heredity, the echinoderms offer material of
unusual advantage. The plants, insects and vertebrates
are peculiarly adapted for the elucidation of the later
aspects of heredity; the echinoderms are available for
information concerning the processes which occur during
and immediately after fertilization.

In the course of this paper I shall show that we are
now in a position to predict certain characters in the
adult hybrids from a knowledge of the germ cells of the
ancestors and, granted the sexual maturity of the hybrid
adults, to predict the character of their germ cells.

The evidence that I shall bring forward is based on a
study of chromosomes. The advantages afforded by
echinoderms for such a study, lie in the fact that not only
may we study the chromosomes during the divisions of

1 Read before the American Society of Naturalists, Princeton, 1911.
68

No. 542] CHROMOSOMES IN ECHINOID LARV 2 69

the normally fertilized egg, but we may study the chromo-
somes of the egg itself, i. e., the maternal chromosomes,
in artificially parthenogenetic eggs; we may study the
chromosomes of the spermatozoan, i. e., the paternal
chromosomes, in fertilized enucleated egg fragments;
and we may study the chromosomes in cross-fertilized
eggs, identifying here those which have come from one
parent or the other.

The first subject of which I shall speak is the correla-
tion of certain chromosomes with sex.

The conclusions reached by Baltzer, from a study of
Echinus and Strongylocentrotus, are known to most of
those present. They find their readiest expression in the
statement that ‘‘in Echinoids the female is digametic
while the male is homogametic.’’ I have expressed else-
where the idea that this generalization is too broad and
that the statement should be limited to the cases for
which the condition was described; but I must do more
than this. I can admit that apparent condition for but
one of Baltzer’s cases, that of Echinus, for his own illus-
trations for Strongylocentrotus indicate that this form
is in agreement with another interpretation, and not with |
the one given.

In the forms which I have studied, the male is digametie,
the female homogametie, as in the insects.

The first case that I shall submit is that of Tripneustes
esculentus (Hipponoé), in which the evidence is espe-
cially clear, although not submitted to the ultimate anal-
ysis that has been given to the second case.

_ In Hipponoé, I have been able to show, by the study of
cross-fertilized eggs, made in connection with straight
fertilized Toxopneustes and Hipponoé eggs by two of my
students, Miss Heffner and Miss Pinney, that a chromo-
some of peculiar form, namely hook-shaped, is trans-
mitted by half of the Hipponoé spermatozoa, and that the
eggs which receive it must become males.

The evidence here is clear. Such an element never
occurs in straight-fertilized Toxopneustes eggs. It is

70 THE AMERICAN NATURALIST [ Vou. XLVI

found in one-half of the straight fertilized Hipponoé
eggs; it is found in half of the Toxopneustes eggs which
have been fertilized by Hipponoé sperm, and it never
occurs in Hipponoé eggs fertilized by Toxopneustes
sperm,

There is then absolutely no appeal from the fact that
in Hipponoé the spermatozoa are of two classes, one
with, the other without, this idiochromosome.

The second case that I shall submit is that of Lytechi-
nus, better known to us as Toxopneustes. Here I am
able to give evidence gained by a study of straight fertil-
ized eggs, artificially parthenogenetic eggs, fertilized
enucleated egg fragments, and four crosses, the recip-
rocal Hipponoé X Toxopneustes and Arbacia X Toxop-
neustes crosses.

The study of straight fertilized Toxopneustes eggs
showed two classes of zygotes, one with two V-shaped
chromosomes, the other with three. This seemed in
agreement with conditions in Echinus, although obviously
there was no chance of tracing the source of the hetero-
chromosome in such material.

Those who have devoted any attention, to the chromo-
somes in Echinoids know the difficulties involved in such
-~ a study. The chromosomes are all small; most of them
have the form of short, straight or slightly bent rods,
while but three or four, at most, may be distinguished by
peculiarity of form. Difficulties attend the study of even
these chromosomes. The V’s in Toxopneustes, for ex-
ample, best seen in the anaphases of division, may have
both arms brought into contact and so resemble a single,
somewhat thickened rod.

Since this is true, it should be easier to determine size
differences when only the haploid number is present than
when the full number of chromosomes is in the division
figure.

So far as I am aware, none of the cytological work on
artificial parthenogenesis in Echinoids has been done
with the idea of an individuality of form of chromosomes

No. 542] | CHROMOSOMES IN ECHINOID LARV Æ 71

in mind. The chief aim has been the determination of the
presence of the haploid number of chromosomes and of
the fact that auto-regulation does not occur. This is true
for results published as late as 1910.

In my study of this material I have found that all of
the eggs are alike, possessing among the others two V-
shaped chromosomes.

In the fertilized enucleated egg fragments, two classes
are evident. Half of the spindles show two V-shaped
chromosomes in each anaphase plate while half show but
one. I have no conclusive evidence for showing whether
we are dealing with an X chromosome or with X and Y
chromosomes. Counts are decisive for either view, and
therefore valueless. I believe that this is a matter which
ean be decided only by a study of the soe cene, ogeieus
of these forms.

For some reason there is a belief in the tradition that
in Echinoids the somatic number of chromosomes is
thirty-six and in one case eighteen. Counts of perfect
polar views of anaphase plates in Toxopneustes give the
numbers thirty-seven and thirty-eight, but even from
these I do not feel warranted in applying a sex formula.

It will be noted that the number of V’s does not agree
with our expectation from the straight-fertilized eggs.
I can only say that my description is of the facts as I
find them in my preparations. From this it would seem
that the number of V-shaped chromosomes in straight-
fertilized Toxopneustes eggs should be three and four.
These facts are further substantiated by conditions in
the Hipponoé X Toxopneustes crosses. |

The second point of which I wish to speak is that of
correlation of chromosomes and larval characters. The
only larval structure which has a form sufficiently definite
to be used for comparison is the skeleton. The most ad-
vantageous forms for this purpose are those in which
the pluteus of one parent has skeletal arm rods of a
simple straight form and the other rods of the latticed
type. |

13 THE AMERICAN NATURALIST [Vor. XLVI

It is well known that an apparent confusion exists
among observations on hybrid Echinoid larvæ, as to
whether plutei of a maternal type, a paternal type or of
mixed form are derived from certain crosses. Different
results have been obtained by different investigators and
by the same investigators working in different regions,
or in the same region in different seasons.

From my own results I am convinced that there is no
reason for believing that any of these observations are
incorrect and that this dissimilarity of results may be
traced to the influence of the constitution of the sea water
at different times and places, to temperature, ete.

If we look into the matter carefully we find that there
is a correlation between the fate of the maternal and
paternal chromatin and the character of the plutei.

It is a remarkable fact that in the general literature of
genetics, authors who have noted the fact of elimination
of chromosomes, as shown by the researches of Baltzer
and Herbst, have not also noted their retention in some
cases, and the influences of this retained chromatin on the
character of the larve.

It may be seen, by the examination of Baltzer’s tabu-
lation, given in his 1910 paper, that in the cases cited an
elimination of presumably the introduced paternal chro-
mosomes is followed by a pluteus of the maternal type,
and that the retention of all the chromosomes may be
followed in some cases by a pluteus of a mixed or inter-
mediate type, in other cases by a pluteus of the maternal
type. To these cases I must add the case of the retention
of chromosomes followed by a pluteus of the paternal
type, this being afforded by the Hipponoé ¢ X Toxop-
neustes 2 cross. :

In this instance we have a clear example of Hipponoé
dominance with respect to larval skeletal character, and
this dominance is correlated with the retention of Hip-
ponoé chromosomes.

I have not asserted that every Toxopneustes egg which
is fertilized by Hipponoé sperm gives rise to a pluteus

No. 542] CHROMOSOMES IN ECHINOID LARVZ 73

with a Hipponoé skeleton. The largest percentage of
the eggs does respond in this manner; a smaller per-
centage dies in the blastula and gastrula stage, and a
still smaller percentage shows little or no trace of pater-
nal influence.

_If we take the eggs of such a fertilization and examine
them during the segmentation stages, we shall see that in
these eggs the greatest number show normal division
figures, and by normal I mean those of the almost dia-
grammatic sort which may be seen in straight-fertilized
Toxopneustes eggs, figures which give no indication of
any chromosome elimination, while the smaller number
show aberrant figures with varying degrees of elimi-
nation.

The normal division figures correspond in number to
the plutei with the paternal type of skeleton and the
aberrant figures to the plutei of intermediate and mater-
nal character.

Clearly, then, in the hybrids of Echioid crosses we
may have dominance of one species or the other with
respect to the skeleton, and this dominance may be trans-
mitted by the egg or by the spermatozoan. Of the con-
siderable number of crosses that I have made, only one,
the Hipponoé & X Toxopneustes oo eross gives this
evidence.

Some facts of interest are demonstrated by the Arba-
cia X Toxopneustes crosses. These crosses are made
with difficulty and I had never succeeded in getting them
until the past summer. The chromosomes in Arbacia
are much smaller than those in Toxopneustes and I had
hoped that material from this cross would be of use when
compared with my experimental Toxopneustes material.
Here, however, I found that there is an elimination of
chromosomes from the first, an elimination which may
involve the rejection of not only the foreign chromo-
somes, but some of those of the egg as well, the result

being that, in some instances, the full haploid number of _

neither is retained, and few — pass meet the |
blastula pa —

74 THE AMERICAN NATURALIST [ Vou. XLVI

A further insight into the effect of the retention of
foreign chromatin is given by partially arrhenokaryotic
and partially thelykaryotic plutei. In such larve, with
their asymmetrical bodies and skeletons, we may dis-
tinguish regional differences by nuclei of different sizes,
and by a study of segmenting eggs we may find how these
nuclei have arisen. In some instances we know that the
fertilization processes have been so modified that subse-
quent divisions of the egg contain only paternal nuclear
material and are paternal in character. In other regions
part or all of the foreign chromatin is present and has
modified development.

The conditions noted in the plutei find correlation in
the following conditions in the egg.

1. Retention of all chromosomes and dominance of one
species over another with respect to skeletal characters.

2. Elimination of part of the chromatin and dominance
of one species over another with respect to skeletal char-
acters.

3. Elimination of part of the chromatin and interme-
diate skeletal cl ters

4. Elimination of part of both paternal and maternal
chromatin and inhibition of development.

The first three, at least, may all occur in a given lot of
eggs, and, since this is true, depend in part on chance,
just as we may have imperfect mitotic figures in straight-
fertilized eges.

The second and third cases may indicate an incom-
patibility between the chromosomes and the cytoplasm
of the egg; the fourth case, not this alone, but an incom-
patibility between the chromosomes themselves.

Finally as to our ability to predict the character of the
adult from the characters of the larve.

It is clear that in Hipponoé and Toxopneustes we may
predict sex, since we have seen that maleness and a
heterochromosome of paternal origin are correlated. It
is thus evident that if chemically fertilized Hipponoé and
Toxopneustes eggs develop to sexual maturity these in-

No. 542] CHROMOSOMES IN ECHINOID LARVÆ 75

dividuals will be females, the reverse of our belief con-
cerning the naturally parthenogenetic egg of the bee. In
cases similar to those described by Baltzer for Echinus,
half the adults will be males and half females.

A prediction as to other characters is difficult to make,
for the fact that one species is dominant over another
with reference to larval skeletal characters give no idea
as to other characters.

Nevertheless, in the cases of partial arrhenokaryosis
which I have mentioned we should expect a region of the .
larval body containing only paternal nuclear material to
place its stamp on that part of the adult body arising
from it. A natural hybrid showing pure regional resem-
blances to one species or another is theoretically possible
and would find its explanation in partial arrhenokaryosis.

For the inheritance of most characters we must fall
back on our knowledge of natural echinoderm hybrids,
which is increasing steadily.

We should expect that where parts of both contribu-
tions of chromatin have been retained and an inter-
mediate pluteus has resulted the adult will be of the
mixed type. One of the supposed Echinoid hybrids de-
scribed by Mortensen fulfils this expectation.

The proof of these predictions lies in the rearing of
hybrid larve to the adult condition. This we know to be
possible. It requires only laboratory advantages of an
unusual type, care, time and an accurate knowledge of
the natural history of echinoids.

DARWIN’S THEORY OF EVOLUTION BY THE
SELECTION OF MINOR SALTATIONS

HENRY FAIRFIELD OSBORN

COLUMBIA UNIVERSITY

THERE is an opinion which is becoming more widely
prevalent daily that Charles Darwin’s theory of selection
rests upon ‘‘fluctuations’’ which may not be heritable.
Nothing could be further from the facts.1

It is true that Darwin finally came to believe in the
inheritance of somatic modifications (in the modern sense
of bodily changes) caused by the direct action of environ-
ment as well as by habit (ontogeny), but in his original
(1859) and final (1872) opinion evolution was chiefly due
to the selection of heritable ‘‘individual differences.”’

Darwin’s true meaning as to ‘‘individual differences”?
is not to be found in his language, but in the cases he
cited; he has been widely misunderstood? as believing in
continuous evolution whereas he chiefly believed in dis-
continuous evolution.

Darwin’s final opinion? may we cited with a transposi-

*De Vries is partly responsible for this general misunderstanding of

Darwin. In ‘‘Die Mutationstheorie,’’ Leipzig, 1901, pp. 21-27, we find a
very full discussion of the opinions of Darwin in which the interpretation
is reached that Darwin was not clear in distinguishing ‘‘variation’’ and
‘mutation. ’’

* Cox, Chas. F., ‘‘ Charles Darwin and the Mutation Theory,’’ Ann. N. Y
Acad. Sci., Vol. XVII, No. II, Pt. III, February 10, 1909, pp. 431-451.
This is a most conscientious and peananive — of Dai rwin’s opinions in
which the following conclusion is reached: ‘ we have seen, he was com-
pelled to concede that what we now call Hest had occasionally taken
place and become the starting point of new races, but he was none the less
unshaken in the conviction that this process was exceptional and extra-
ordinary, and that, as a rule, a new species originated by the gradual build-
ing up of minute and even insignificant deviations from the average char-
acters of an old species, which deviations we now call fluctuations.’’ [P.
ssi = our own. ]

2 ‘‘The Variation of Animals and Plants under Domestication,’’ Vol.
1, p. 397.

76

No. 542] DARWIN’S THEORY OF EVOLUTION 77

tion of sentences and italicizing of important words, as
follows:

-... we have abundant evidence of the constant occurrence under
nature of slight individual differences of the most diversified kinds; and
we are thus led to conclude that species have generally originated by,
the natural selection of extremely slight differences. This process may
be strictly compared with the slow and gradual improvement of the
racehorse, grayhound, and gamecock. As every detail of structure in
each species has to be closely adapted to its habits of life it will rarely
happen that one part alone will be modified; but as was formerly
shown, the co-adapted modifications need not be absolutely simultaneous.
Many variations, however, are from the first connected by the law of
correlation. Hence it follows that even closely allied species rarely or
never differ from one another by one character alone . . . from the
history of the racehorse, grayhound, gamecock, ete., and from their
general appearance we may feel nearly confident that they were formed
by a slow process of improvement; and we know that this has been the
case with the earrier-pigeons as well as with some other pigeons. . . . It
is certain that the ancon and mauchamp breeds of sheep, and almost
certain that the niata cattle, turnspit, and pug dogs, the jumper and
frizzled fowls, short-faced tumbler pigeons, hook-billed ducks, ete., sud-
denly appeared in nearly the same state as we now see them. So it has
been with many cultivated plants. The frequency of these cases is
likely to lead to false belief that natural species have often originated in
the same abrupt manner. But we have no evidence of the appearance,
or at least of the continued procreation, under nature, of abrupt modi-
fication of structure, and various general reasons will be assigned
against such a belief.

Thus Darwin, on the admirable ground that no evidence
had been adduced in nature of such transformation, re-
jected the hypothesis of major saltatory evolution as
causing the natural appearance of entirely new types of
animals or plants or of entirely new and profoundly
modified organs. There was no evidence in 1872 and
there is none to-day of the sudden appearance in nature
of such a breed as the ancon sheep. On the other hand,
Darwin’s meaning as to “‘slight individual differences of
the most diversified kinds,” as clearly conveyed in the
hundreds of observations he cited in ‘‘The Origin of
Species” (edition of 1872) and especially in his ‘‘ Varia-

tion of Animals and Plants Under Domestication,” is : oe

78 = THE AMERICAN NATURALIST [Vou. XLVI

that such individual differences were in the nature of
minor saltations of character, structural or functional,
and always hereditary. Since Darwin observes that
natural selection may be strictly compared with the slow
and gradual improvement of the racehorse, grayhound,
etc., if we collect all the observations which he assembled
as to the genesis of the racehorse and grayhound we may
gain a concrete understanding of his opinions as to the
genesis of species. It will appear that many of the new
characters which Darwin cites, and he used the term
‘‘new characters’’ interchangeably with ‘‘individual dif-
ferences,’’ are clearly identical with minor saltations or
with the ‘‘mutations of De Vries,’’ as generally under-
stood by zoologists to-day.

‘We may therefore conclude,” observes Darwin,*
‘‘that, whether or not the various existing breeds of the
horse have proceeded from one or more aboriginal stocks,
yet that a great amount of change has resulted from the
direct action of the conditions of life, and probably a still
greater amount from the long-continued selection by man
of slight individual differences.’’ Among the examples
he cites in amplification of this conclusion are the fol-
lowing:

(1) Eight incisors or two super-normal; (2) canines in
females; (3) a nineteenth or posterior rib; (4) a supple-
mentary hock-bone; (5) a reversional trapezium and
Mtc.V; (6) horn rudiments in the frontal bones one or
two inches in length; (7) tailless foals which, he observes,
might have produced a new breed; (8) curled hair corre-
lated with short manes and tails and mule-like hoofs.

All the above ‘‘individual differences’’ are cited by
Darwin as hereditary; all are ‘‘discontinuous’’ in Bate-
son’s sense or ‘‘mutations’’ in De Vries’s sense. In other
parts of Darwin’s works most of the numerous references
made to ‘‘individual variations’’ in horses are of the
same character, namely, saltatory and hereditary. Al-
though (5) cited above is a ‘‘reversion,’’ reversions and

*** Animals and Plants under Domestication,’’ Vol. I, p. 55.

No. 542] DARWIN’S THEORY OF EVOLUTION fh: ae

ontogenic modifications are, for the most part, clearly
distinguished as belonging to another category.

Similarly Lutz’ has kindly analyzed for the writer
twenty-one cases cited by Darwin among insects as possi-
ble material for selection and as indicating modifications
by environment. Eight of these cases are doubtful;
seven cases represent ‘‘gradations,’’ or ‘‘slow degrees,’’
probable or possible continuity; six cases represent dis-
continuity. It seems to Dr. Lutz that Darwin made the
most of the cases of discontinuous variation he knew
about, but that for the most part he had continuous
variation in mind.

It is therefore of interest, in view of the neglectful and
almost contemptuous attitude of certain writers toward
Darwin’s observations, to make a fresh summary of the
principles found scattered through the pages of his great
work on ‘‘Variation,’’ reexpressing some of these prin-
ciples in modern terms.

Darwin on ‘‘ New Cuaracters’’ AND ‘‘[NDIVIDUAL DIFFER-
ENCES’’® as EXPRESSED IN MoperX TERMS

(1) Newly appearing characters arise from unknown
causes, either stable in heredity, variable in heredity, or
not hereditary at all.” (2) ‘‘Individual variations’’ are
minor suddenly appearing characters, heritable. (3)
Characters of all kinds, whether new or old, tend to be
‘inherited. Those which have already withstood environ-
ment will, as a rule, continue to withstand it and be truly
transmitted. (4) A change of environment is principally
but not invariably the source of new variations. (5)
There are periods of variability or mutability in which
many new characters appear. (6) New characters are
observed to accumulate in successive generations in the

* Frank E. Lutz, letter to the writer, December 15, 1911, ‘‘ References to
insects in the ‘Origin of Species,’ —_ those bearing on the question
of continuity versus discontinuity.’

*See American edition, 1900, ‘‘Animals and Plants under Domestica-
tion,’’ dated by Darwin January, 1872.

? Vok L 9; 87:

80 THE AMERICAN NATURALIST [ Vou. XLVI

same direction. (7) There is a predisposition to similar
progressive variations in descendants of one stock
(termed ‘‘analogous or parallel variation’’) due to un-
known causes acting on similarly constituted organisms.’
(8) There is a marked prepotency in heredity in saltatory
types, e. g., ancon sheep, turnspit dogs. (9) There is ‘‘in
a certain sense’’ a mosaic inheritance (after Naudin)
observed in hybrids, or character segregation by self at-
traction and self affinity. Similarly there is an exclusive
inheritance, as observed in characters derived exclusively
from the father or mother and lying dormant or latent
for many generations. (10) There is a particulate inher-
itance, for example, of the special quality of vigor and
endurance observed in descendants of the racehorse
‘*Hiclipse’’; i. e., a dominance of single characteristics.
(11) A large number of individual variations cited are
antithetic characters rather than slight or infinitesimal

gradations of similar cl ters, €: g.,
Erect ears Drooping ears
Fertility Sterility
Immunity < Non-immunit :
Resistance to environment Non-resistance to environment

(12) Environment directly affects the antithetic charac-
ters of sterility and fecundity. (13) New characters may
appear and old ones disappear at any stage of develop-
ment.® (14) There is a similar age heredity or inher-
itance at corresponding periods of life. (15) There is a`
sex-limited heredity. (16) There is a correlated varia-
bility or coupling of new characters in heredity, e. g.,
black color of the skin and immunity from disease. (17)
Particulate inheritance as shown in non-blending colors,
e. 9., gray and white mice produce piebald young, or gray
or white young, but never a blend between gray and
white. (18) Old characters lying latent are ready to be
evolved under certain old environmental conditions.
Thus we observe the ontogenic revival of feral characters

*Vol. II, p. 329.
°’ Vol. TI, p. 392.

No. 542] DARWIN’S THEORY OF EVOLUTION Cae

in feral environment, and of domesticated characters in
return to domestication, e. g., barking in dogs. (19)
Ontogenic reactions to environment are observed in all
external and internal characters in animals and plants.
(20) Under selection Darwin anticipated the theory of
organic or coincident selection.!°

SuMMARY

Darwin did not so sharply distinguish in language as
Weismann and Mendel have taught us to between the
different kinds of variation and different factors of evo-
lution, yet in his observations and the cases he cites his
perception is absolutely clear between heredity, ontogeny,
environment and selection as interoperating factors. In
fact, he remarks! on the frequent difficulty of distin-
guishing between the inextricably mingled factors of
ontogeny (e. g., effects of use and disuse), and of corre-
lated or coadapted variability and of spontaneous varia-
tions. A comparison of all the various kinds of varia-
tion cited by Darwin in his two great volumes shows that
they fall into the following four classes:

I. ‘‘Individual variations,’’ ‘spontaneous variations,”’
new suddenly appearing heritable characters, practically
equivalent to the minor mutations of De Vries, believed
by Darwin to be the chief material of natural selection
and evolution.

II. Sports, or major saltations, such as the ‘‘mau-
champ,” ‘‘ancon’’ and ‘‘niata’’ breeds, believed by Dar-
win not to occur in a state of nature.

III. Fluctuations of proportion, congenital and hence
transmissible, equivalent to the quantitative variation of
Bateson, best illustrated by Darwin in his theory of the
evolution of the long neck of the giraffe:

So under nature with the nascent giraffe, the individuals which were
the highest browsers, and were able during dearths to reach even an
inch or two above the others, will often have been EER for they

* Vol. II, pp. 317, 318.
” Vol. II, p. 327.

82 THE AMERICAN NATURALIST (Vou. XLVI

will have roamed over the whole country in search of food. That the
individuals of the same species often differ slightly in the relative
lengths of all their parts may be seen in many works of natural history,
in which careful measurements are given. These slight proportional
differences, due to the laws of growth and variation, are not of the
slightest use or importance to most species. But it will have been other-
wise with the nascent giraffe, considering its probable habits of life;
for those individuals which had some one part or several parts of their
bodies rather more elongated than usual, would generally have survived.
These will have intererossed and left offspring, either inheriting the
same bodily peculiarities, or with a tendency to vary again in the same
manner; whilst the individuals, less favored in the same respects, will
have been the most liable to perish.

IV. Fluctuating variability, clearly distinguished by
Darwin from II and not especially connected by him with
the process of evolution.

December 30, 1911

THE COLOR SENSE OF THE HONEY-BEE: THE
POLLINATION OF GREEN FLOWERS!

JOHN H. LOVELL
WALDOBORO, ME.

CARL NAGELI, in a memoir often cited, after stating the common opin-
ion that insects are especially attracted by the bright coloration of floral
organs, says: “ We understand now why there are no green flowers;
they would be invisible in the midst of foliage,”

We seem to dream when we read these lines written seriously by a
naturalist. As the lists below demonstrate, not only are there a large
number of green or greenish flowers, invisible or scarcely visible in the
midst of foliage, but insects discover them without the least diffculty.

This false notion of the non-existence of green flowers, which can
only be excusable among those unacquainted with botany, has arisen
from many causes; the plants which bear flowers of this nature are only
cultivated when they are of medicinal use or are valuable for food, when
they are frequently regarded as vegetables rather than as flowers. In
foreign countries natural history collectors travelling for horticulturists
disdain everything which has not at least a decorative foliage. Further-
more, the authors who attach great importance to the coloration of the
floral envelopes for the attraction of insects have often neglected to
mention in their works that such and such a species, though fertilized
by winged arthropods, possesses green or greenish flowers. H. Miiller
has committed this fault many times, and Charles Darwin in his cele-
brated work on the fertilization of orchids passes over the color usually

in silence.
` A reaction was necessary; it was important to remind biologists that
there exists a long series of green flowers, greenish, or scareely visible,
which insects find as easily as the others, despite the absence of their
so-called attractive colors.’

Plateau then proceeds to enumerate 91 entomophilous
species, of which 41 had green or largely green flowers,

<The Color Sense of the Honey-bee: Is Conspicuousness an Advantage

to Flowers???’ Amer. NAT., 43, 338-349, June, 1909; ‘‘The Color Sense of
the Honey-bee: Can Bees Distinguish Colors??? Amer. NAT., 44, 673-692,
t

84 THE AMERICAN NATURALIST [ Vou. XLVI

38 greenish and 12 brownish or brown flowers, to the
larger part of which insect visits had been observed by
himself or other floroecologists. Of 72 of these forms he
examined the color himself and observed insect visits to
63 of them, or more than two thirds. Plateau believed
that the visits of insects to green flowers very strongly
supported his views, and in one of his last papers on the
pollination of Listera ovata, an orchid with green flowers,
he again returns to this subject. In a letter to the writer
he states that he had made innumerable observations on
the pollination of green flowers, which had extended over
twelve years. His conclusion as finally expressed is as
follows:

It is not the colors more or less bright of the corollas but other causes.
which guide to flowers their winged pollinators. Green or greenish
flowers notwithstanding their hue similar to that of foliage are as
effectively fertilized by insects as white, blue, red, or yellow flowers,.
consequently, all these floral colorations might disappear from nature
without the pollination and reproduction of plants being diminished.’

It is desirable to examine briefly the pollination of
green flowers and to consider whether this conclusion is
sustained or not.

A familiar example of a yellowish-green flower attract-
ive to insects is offered by the garden asparagus (As-
paragus officinalis). The flowers are mellifluous, pleas-
antly scented and frequented by honey-bees and less
often by smaller species of bees. Müller says that ‘‘in
spite of their inconspicuous color they are easily visible
at a distance”; but Plateau describes them as ‘‘peu
visible,’’ an illustration perhaps that an observer is apt
to be influenced by his point of view as well as by the
color of the flower. As the result of personal trial I find
that they can be distinctly seen at a distance of at least
fifty feet. Another yellowish-green flower is Tilia ulmi-
‘folia, or basswood, which, according to both Müller and
Plateau, is sought by honey-bees in immense numbers.
This is also true of the American basswood (Tilia ameri-

* Plateau, F., ‘‘La pollination d’une orchidée a fleurs vertes ‘ Listera

ovata R. Br.’ par les insectes,’’ Bull. Soc. roy. Belgique, 46, 339,,1909.

-

No. 542] COLOR SENSE OF THE HONEY-BEE 85

cana), which, according to Root, an eminent authority on
American apiculture, furnishes more honey than any
other plant known, with perhaps the single exception of
the white clover. Other greenish, or dull-colored flowers,
which the honey-bee was observed to visit were Ampel-
opsis quinquefolia, Teucrium Scorodonia, Comarum pa-
lustre, Atropa Belladonna, and several species of Scro-
phularia.

But of the 91 greenish or brownish hued species the
honey-bee was seen to visit only 27, that is, there are no
recorded visits of this bee to a little over seventy per cent.
of the listed flowers. To four species there are no rec-
ords of the visits of any insect. The remaining species
were chiefly visited by flies, sometimes minute species,
beetles, and the less specialized Hymenoptera; but He-
dera helix was attractive to wasps and Platanthera bi-
folia to Lepidoptera. How meager the number of visits
was should be noted: Celastrus Orixa was visited
only by the domestic fly and one other species of Diptera;
Alchemilla hybrida only by small Muscide; Alchemilla
alpina in the Alps by one beetle and two Muscide; Alche-
milla fissa in the same mountains by three Muscide; the
sessile green flowers of Herniaria glabra were visited
only by very small insects; the yellowish-green flowers
of Amarantus retroflexus by the domestic fly and one
beetle; Lepidium Smithii by small flies and many beetles
of the genus Altica; Angelica pyrenea by one beetle and
two flies; three species of Euphorbia were visited only
by flies; Salsoda Soda by Syritta pipiens and microscopic
Diptera; among brownish flowers Pelargonium triste was
visited by one fly; Vincetoxicum purpurascens only by
Musca domestica; the brown-violet flowers of Veratrum
nigrum and the reddish flowers of Neottia Nidus-avis
only by flies. According to this list, in a few instances
greenish flowers secreting nectar very freely are visited
by a large number of insects, but the majority of the-
species are evidently not well adapted to entomophily. —

*Root, A. I., ‘The ABC of Bee Culture,’’ p. 38, 1903. |

86 THE AMERICAN NATURALIST [ Vou. XLVI

This conclusion is sustained by an examination of the
green flowers of eastern North America. In the terri-
tory east of the 102d meridian and northward of North
Carolina and Tennessee there are 1,244 green or dull
colored flowers, of which 1,021 are anemophilous or hydro-
philous, while 223 are entomophilous or autogamous.
The wind-pollinated flowers are small and usually grèen-
ish, as in the 705 species of grasses and sedges, from
which the inference is commonly drawn that anemophily
and inconspicuousness are correlated. In the few excep-
tions (about 27 in the flora of the eastern states) where
anemophilous flowers have brighter hues, as in the golden
yellow aments of the yellow birch and the deep red pani-
cles of the field sorrel (Rumex Acetosella), the colors of
the small flowers are usually determined by the produc-
tion of yellow or red pigments in great abundance by the
vegetative organs—the entire plant of the field sorrel
being sometimes red-colored.

Of the 223 green flowers, which are entomophilous or
autogamous, many have no petals, as fifteen species of
the Polygonacee and eight species belonging to the
Caryophyllacee, also in several Rosaceæ, in Acer sac-
charinum and Didiplis diandra. Many are self-fertilized,
as Triglochin and Scheuchzeria, and the orchids Habe-
naria hyperborea and Epipactis viridiflora, and the small
green flowers of Lechea and Penthorum sedoides. Some
have the petals caducous and depend upon their scent to
attract insects, as the Vitaceæ. Many are visited chiefly
by flies and the smaller bees, as various Melanthacex, the
Smilacee, and the green flowers of the Asclepiadacee.
A few species secrete nectar freely and attract numerous.
visitors, as the rock maple, basswood and the dark green
pistillate flowers of Rhus typhina. Large green flowers,
which are fragrant, nocturnal and are pollinated by
moths, are found in exotic Solanaceæ and Orchidaceæ. It
1s obvious that bright coloration is less important to
moth flowers than a strong Scent, since red and blue

No. 542] COLOR SENSE OF THE HONBY-BEE 87

shades are invisible at night. But as a whole green
flowers are small or even minute and attract few insects.’

Since in Europe and America, where the insect fauna
is rich both in species and individuals and the flora dis-
plays a great variety of brilliant hues and delicate shades,
dull-colored flowers are not well adapted to entomophily,
it may be inferred that in a country where the flowers
are largely greenish there would be a scarcity of antho-
philous insects. According to A. R. Wallace and G. M.
Thomson this condition is partially realized in the Islands
of New Zealand. Wallace says:

In New Zealand where insects are strikingly deficient in variety the
flora is almost as strikingly deficient in gayly colored blossoms. Of
course there are some exceptions, but, as a whole, green, inconspicuous
and imperfect flowers prevail to an extent not equalled in any other part
of the globe, and affording a marvellous contrast to the general brillianey
of Australian flowers, combined with the abundance and variety of its
insect life. We must remember, too, that the few gay or conspicuous
flowering-plants possessed by New Zealand are almost all of Australian,
South American, or European genera. .. . After the preceding para-
graphs were written, it occurred to me that, if this reasoning were
correct, New Zealand plants ought to be also deficient in scented flowers,
because it is part of the same theory that the odors of flowers have, like
their colors, been developed to attract the insects required to aid in their
fertilization. I therefore at once applied to my friend Dr. Hooker, as
the highest authority on New Zealand botany; simply asking whether
there was any such observed deficiency. His reply was, New Zealand
plants are remarkably scentless.

After quoting the above passage, G. M. Thomson, who
resided in New Zealand and made a special study of its
fioroecology adds:

It is impossible to differ from this reasoning in toto, because the
statements and facts on which it is founded are to a great extent correct
though in the light of more recent oni they require considerable
modifieation.®

Bees and intone. according to Thomson, are com-
paratively rare; while Diptera, of which it is estimated

* Lovell, John H., ‘‘The Colors of Northern Gamopetalous Flowers,’’
AMER. Nar. i 365-384 and 443-479.

, G. M., ‘‘On the Fertilization of New Zealand Flowering
Plants Si ea Proc. N. Z. Inst., 13, 241-288, 1880; Wallace, A. + Phe
Geopraphinas Distribution of Abia, > 1, 457-464.

88 THE AMERICAN NATURALIST [Vou. XLVI

there may be a thousand species, are here the chief agents
in the pollination of flowers, whereas in America and a
large part of Europe they occupy the second place and in
the Alps the third place. Neither honey-bees nor bumble-
bees were found in New Zealand at the time of its dis-
covery.

The phylogenetic history of green flowers likewise
strongly supports the view that they are not well adapted
to pollination by insects. In the opinion of many emi-
nent botanists all greenish, inconspicuous flowers have
been derived by retrogression from larger entomophilous
ancestors. This theory has been very ably developed by
Professor C. E. Bessey in his taxonomy of the Angio-
sperms, for which he suggests the restoration of the more
appropriate name of Anthophyta. The buttercups
(Ranales), the water plantains (Alismales) and roses
(Rosales) are regarded as primitive and are placed at
the beginning of the Anthophyta. The typical flower was
entomophilous, of large size, and its organs were sepa-
rate and spirally arranged. Engler’s spiral series of
Monocotyledons, which is composed of orders mostly
devoid of a perianth, is derived from a liliaceous type;
while the Apetale are treated as reduced forms and dis-
tributed among the petalous Dicotyledons.”

A similar view is adopted by Arber and Parkin in their
discussion of the origin of Angiosperms. They reach the
conclusion ‘‘that the Apetalous orders without perianth,
such as the Piperales, Amentiferous families and Pan-
danales, can not be regarded as primitive Angiosperms,
_ but have been derived from ancestors with a well-de-
veloped perianth. Entomophily . . . has supplied the
‘motive force,’ which not only called the Angiosperms
into existence, but laid the foundation of their future
prosperity.’’> Even if Engler’s system of classification

"Bessey, Charles E., ‘‘The Phylogeny and Taxonomy of Angiosperms,’’
Bot. Gaz., 24, 145-178, 1897; “A Synopsis of Plant Phyla,’’ University
Studies, 7, 275-373, 1907; ‘‘The Phy
29, 91-100, 1909, ete.

-Arber B. A: Newell, and Parkin, John, ‘‘On the Origin of Angio-
sperms,’’ Journal of Linnean Society (Botany), 38, 29-80, 1907.

letie Idea in Taxonomy,’’ Science,

No. 542] COLOR SENSE OF THE HONEY-BEE 89

is accepted and the apetalous orders be regarded as primi-
tive it does not support the thesis that small, green
flowers are at no disadvantage in attracting insects be-
cause of their inconspicuousness. That reduction and
change from entomophily and conspicuousness to ane-
mophily and inconspicuousness has occurred repeatedly
in widely separated families is not questioned by any eco-
logist or taxonomist. This is illustrated by the genera
Artemisia, Ambrosia, etc., among the Composite; in
Ricinus of the Euphorbiacex ; probably also in the family
Juncacex, and in species of Thalictrum, Fraxinus, San-
guisorba and Poterium. The evidence supplied by the
phylogeny of green flowers is wholly in favor of the value
of color contrast for gaining the attention of insects.

Approaching the problem from another direction, the
Rev. George Henslow in his work on the self-fertiliza-
tion of plants finds that ‘‘inconspicuous flowers sees gi
most invariably self-fertilizing, or else inconspicuous.’

There are several reasons why inconspicuous flowers are not likely to
be intererossed by insects: (1) Their unattractiveness; (2) the absence
of hohey-secreting organs; (3) the want of scent; (4) they frequently
do not expand, or at most remain half open, especially in cold or inclem-
ent weather, while perfectly cleistogamic flowers are, of course, never
open; (5) their structure sometimes would seem absolutely to prevent
the ingress of insects (such appears to be the case with Polygonum
Convolvulus and-P. hydropiper, the flowers of which seem to be always
closed, and with many others.

He regards existing inconspicuous forms not as primi-
tive, but as derived from conspicuous progenitors, which
in turn owed their origin to the selective influence of
insects.° ;

It has been shown that Plateau’s conclusion is not sus-
tained either by the phylogeny and distribution or by the
ecology and manner of fertilization of inconspicuous
flowers, which have almost universally been compelled to
adopt anemophily or autogamy. In the few exceptional
cases there are present other allurements, as odor and
nosar, which sooner or later attract insects; but this

Henslow, George, ‘‘On the Self-fertilization of Plants,’’ Trans. -Pe
Soc. (Botany), Ser. 21, 317-398, 1880.

90 THE AMERICAN NATURALIST [ Von. XLVI

does not prove that ceteris paribus color contrast is not
an advantage. No assertion is made that bees have an
antipathy to green, only that flowers of this hue are not
as readily seen amidst the foliage. Very likely a green
flower opposed to red or yellow leaves would attract the
attention of insects as readily as the reverse contrast.
Following the example of Plateau, I have included the
species of Tilia among greenish flowers, but it is doubtful
whether the inflorescence of the American basswood
should be considered as inconspicuous. The flowers are
of medium size, sweet scented, produced in vast quanti-
ties, and are described by a disinterested observer as
‘‘vellow and rather pretty.’’ The nectar is so copious
that a single hive of bees has obtained 66 pounds in three
days, and its odor is so strongly aromatic that it can be
perceived throughout an entire apiary.° It may, how-
ever, serve as an example of an exceptional species. The
importance of scent as an attractive factor was, of course,
recognized by Miiller, but green flowers are usually odor-
less, as pointed out by both Hooker and Henslow."!

As additional evidence that insects will visit green.
flowers Plateau describes how he placed honey on seven-
teen anemophilous flowers, as grasses, sedges, rushes and
on species of Rumex and Chenopodium, and observed
visits of honey-bees, flies and a few other insects.12 He
also fashioned crude imitations of flowers from the liv-
ing leaves of the red currant (Ribes rubrum) and the
sycamore (Acer Pseudo-Plantanus), in which he put

* Root, A. I., ‘‘The A B C of Bee Culture,’’ p. 397, 1903. A single
colony of bees belonging to the late Dr. Gallup, of Orleans, Iowa, once
gathered 600 pounds of basswood honey in thirty days. Doolittle, G. M.,
‘‘ Honey from Basswood,’’ Gleanings in Bee Culture, 36, 23, 1908.

“A nectariferous flower may be both green and scentless and yet be
found by bees. According to Fritz Miiller the flowers of a species of
Trianosperma in South Brazil are visited very abundantly all day long by
Apis mellifica and species of Melipona, although they are scentless, greenish,

he

quite inconspicuous, and to a great extent hidden by the leaves. ‘‘
Fertilization of Flowers,’’ p. 270.

“Plateau, F., ‘‘Comment les fleurs attirent les insectes,’’? Bull. Acad.
roy. Belgique, ins partie, 3me série, 34, 602-612, 1897.

No. 542] COLOR SENSE OF THE HONEY-BEE 91

honey, and they attracted bees and flies.1* Attention is
also called to the secretion of nectar by extra-floral nec-
taries upon the petioles of Prunus, the stipules of Vicia
and the leaves of various trees. These sources of sweet
secretions are frequently visited by Hymenoptera and
other insects, as well as over-ripe or partially decayed
frut.

A more interesting example than any given by Plateau
of the secretion of nectar by extra-floral nectaries is
furnished by the American cotton plant. Besides the
nectar glands within the flowers there is a small gland on
the center rib on the under side of each leaf, which at —
times secretes nectar very freely. When the atmospheric
conditions are right, says Mr. J. D. Yancey. in a recent
number of Gleanings in Bee Culture, drops of nectar will
collect on these leaf glands so large that they may be
readily tasted; and a bee has to visit only a very few to
obtain a load.

At such times they neglect the blossoms entirely, and the honey
comes in with a considerable rush. I could not tell that this honey was
any different in either color or flavor from that gathered from the
blossoms.

No other plant in this country besides cotton is known
to me which has extra-floral nectaries, which are of value
as a source of honey; but in favorable years there occurs
on a scale of enormous magnitude an illustration that
honey-bees will readily learn to gather sweet liquids from
green leaves. In many parts of Europe and Ameriċa
Aphididæ, or plant-lice, and scale insects (Lecanium)'*
excrete a sweet substance called honey-dew in such large
quantities that not only the leaves of the trees, but even
the grass and the sidewalks, are coated with it as with a
varnish. Honey-dew is attractive to many insects, as
bees, ants, wasps and flies. In California it is sometimes

In Hawaii enormous quantities of honey-dew are produced by a leaf-
hopper. Phillips, E. F., ‘‘The Souree of Honey-dew,’’ Gleanings in Bee
Culture, 38, 177, 1910.

* Loe. cit., 5me partie, p. 868.

* Loc. cit., 4me partie, p. 604.

92 THE AMERICAN NATURALIST [ Vou. XLVI

so abundant that bees gather large stores of it, and Pro-
fessor Cook says that he has sold it by the barrel.'®
From Sevensville, Montana, a correspondent of Glean-
ings in Bee Culture wrote a few years ago that the honey-
dew had been in a continuous flow throughout the whole
season, and dripped on the sidewalk every night in large
quantities. Another bee-keeper at Dupont, Indiana,
states that, in 1884, his bees gathered about two tons of
honey-dew from the leaves of the oak, hickory, beech and
wild grape. But the year, 1909, was in the opinion of
well-informed apiarists the greatest year for honey-dew
ever known in America. Bee-keepers everywhere re-
ported a scarcity of white clover and basswood honey and
that the bees were storing honey-dew. Professor Surface
says:

I have never known a year in all my studies of entomology, and in a
correspondence of thousands of persons each month, during which
plant-lice, or aphids, have been so abundant as they have this year
(1909), and consequently the honey-dew was likewise unusually abun-
dant.”

Many tons of this sweet excretion were consumed by
the bees during the following winter.

I have dwelt at some length on the collection of honey-
dew in order to establish beyond any question, not only —
that domestic bees would, but that they do gather large
supplies of sweet substances from green leaves. If addi-
tional evidence could strengthen this statement it is fur-
nished by every apiary, where bees frequently may be
seen feeding on materials of every hue, or entering dark
supers, hive-bodies, or boxes, through narrow crevices or
small apertures no larger than a bee’s body in search of
honey. Honey-bees require a great amount of stores and
it would be greatly to their disadvantage, if their actions
were dominated by bright coloration to such an extent
that they were prevented from obtaining food supplies

1 Root, A. I., and Root, E. R., ‘‘The A BC and X Y Z of Bee Culture,’ ’?
p. 273, 1910.

* Surface, H. A., ‘‘Sources of Honey-dew,’’ Gleanings in Bee Culture,
37, 623, 1909. l

No. 542] COLOR SENSE OF THE HONEY-BEE 93

from every available source. It is, then, freely admitted
that bees will collect sweet liquids, after they have once
been found, from green or dull-colored surfaces; but
this is very far from proving that bright coloring is not
an advantage to flowers, and it is astonishing that such a
claim based on the above facts should ever have been
made. .

Knuth in reviewing the observations of Plateau on
greenish and brownish flowers very properly raises the
objection that ‘‘Plateau has not compared the frequency
of insect visits to inconspicuous and conspicuous flowers
of the same size, and it is only experiments of this kind
which can help to settle the point at issue.’”!* This omis-
sion is fatal to Plateau’s argument, and it is difficult to
understand why control experiments were not employed.
It is the object of the present paper to present the results
of a long series of experiments, in which honey-bees
under similar conditions were given the choice between a
conspicuous and an inconspicuous object.

s a preliminary inquiry it is of interest to determine
whether plants with diccious inflorescence afford any
assistance in deciding this question. As is generally
known, the staminate flowers of entomophilous and some-
times of anemophilous diclinic species are more con-
spicuous than the pistillate. This is well shown by the
genus Salix. Willow branches bearing staminate aments
are offered for sale in New England cities in early spring,
and are used for decorative purposes in the churches of
England on Palm Sunday. Careful observation and col-
lection of the visitors of Salix discolor (the glaucous or
pussy willow), the earliest species of Salix to bloom in
this locality, shows that the number of insects attracted
by the staminate aments is much greater than by the
pistillate. The difference is, indeed, so marked as to be
readily apparent to any one who will keep an individual
shrub of each form under observation for a few hours.

13 Knuth, Paul, ‘‘Handbuch der Bliitenbiologie,’’ 1, 394, 1898; ‘‘ Hand-
book of Flower Pollination,’’ translated by J. R. Ainsworth Davis, 1, 207, —
1906. a as

94 THE AMERICAN NATURALIST [Vor. XLVI

Another common dicecious plant is Rhus typhina L.
(Rhus hirta (L.) Sudw.).1° The staminate flowers are in
large white panicles. The thyrsoid, pistillate flower-
clusters are dark green; but as they are terminal and
borne well above the foliage they are visible at a long
distance, i. e., they have conspicuousness of position. Two
large groups of this shrub, or small tree, one of which
was pistillate and the other staminate, growing in an
open woodland only a short distance apart, were selected
for observation. During two collecting trips in July,
1909, I secured on the staminate blossoms 77 visitors, but
only 6 on the pistillate. Of the visitors to the staminate
flowers 63 were bees, 2 wasps, 2 flies, 8 beetles and 2
Hemiptera. Of the visitors to the pistillate flowers 4
were bees and 2 wasps. Even a brief inspection is suff-
cient to show that the staminate flowers are more attract-
ive to insects than the pistillate.

The staminate inflorescence of Salix discolor and Rhus
typhina is, then, undoubtedly more conspicuous and
attractive to insects than the pistillate; but is this larger
company of visitors due wholly to its brighter coloration?
Evidently not. Great numbers of honey-bees and many
species of Andrena frequent the staminate aments of
Salix to procure pollen for brood-rearing. At least five
species of Andrena are oligotropic visitors of this genus.
Examination of the polleniferous scopa of the bees taken
on the staminate flowers of Rhus typhina showed that
they all contained pollen except nine specimens of Proso-
pis (P. modesta Say and P. zizie Rob.), a primitive
genus, the species of which possess only feebly developed
brushes on the posterior legs, which are not used for
carrying pollen. Of the eight species of beetles taken on
the staminate flowers microscopic examination showed
an abundance of pollen on the mouth-parts of four. The
other four beetles were of small size, and it was not

definitely determined whether they or the other insects —

mentioned above were feeding on pollen or not. But this

2 The flowers of Rhus typhina are given as polygamous in most plant
manuals, but they are certainly diccious, as is also stated by Müller.

No. 542] COLOR SENSE OF THE HONEY-BEE 95

was of little consequence, since the proof is ample that in
the case of the staminate plants of Rhus typhina, as in
Salix, the pollen is an important factor in attracting
visitors.

It is commonly believed that insects are attracted first
to the staminate flowers of entomophilous diccious plants
by their greater conspicuousness, from which subse-
quently they carry pollen to the pistillate flowers. For in-
stance, Müller says of the dicecious flowers of Asparagus
officinalis: ‘‘This instance confirms Sprengel’s oft-re-
peated rule that the male flowers of diclinic plants are
more conspicuous than the female, whence insects are
likely to visit the two kinds of flowers in their proper
sequence.’’2? But it is clear that many insects never fly
to the pistillate flowers, since if they did the number of
visitors to the two kinds of flowers would be equal. This
is especially true of female bees, which, having obtained
their load of pollen, often return directly to the hive or
nest. Of the four bees taken on the pistillate flowers of
Rhus typhina only two had pollen on the scopa of the
hind legs. After collecting the pollinators of this species
for several seasons I think it probable that some of the
species of bees taken on the staminate flowers are never,
or very rarely, found on the pistillate blossoms.” Spren-
gel’s rule must, therefore, be accepted with considerable
reservation.

The observations on diccious flowers not proving well
adapted for the purpose intended, owing to the presence
of pollen as an attractive factor, the following experi-
ment was tried. The flowers of Gerardia purpurea have
a rose-colored, campanulate corolla and a short bell-
shaped calyx. The species is common in this locality and
is sparingly visited by bumblebees. When a large bou- —
quet of the flowers was placed in front of a hive of black
bees, it received very little attention. Apparently they
contained no nectar. I now placed in the throat of a

” Müller, H., ‘‘The Fertilization of Flowers,’’ p. 549. |

2 This is also true of dichogamous flowers. Robertson, C., ‘‘ Flowers
and Insects,’’ Bot. Gaz., 27, 41, 1899.

96 THE AMERICAN NATURALIST [Vou. XLVI

large number of flowers a small drop of honey. From a
number of other stems I removed all the corollas and all
conspicuous buds, and between the green calyx teeth I
also put a drop of honey. So abundant was the honey on
the green calyces that it could be seen at a distance of
four feet. I could detect no scent in the complete flowers;
certainly they seem to possess none comparable with that
of honey. The two clusters of plants, the one decorallated
and the other with its flowers complete, were placed on
opposite sides of a glass of water, which was set near the
entrance of a hive of black bees. The bees immediately
showed a decided preference for the flowers retaining their
corollas, as many as five visiting them at one time; while
there were no bees on the denuded flowers though they
were on the side of the glass nearer the hive. Later the
bees discovered, as was to be expected, the honey on the
green calyces and removed it. It is evident that to place
honey on small green flowers, as in the experiments of
Plateau with grasses and sedges, and when it is finally
found by insects to conclude that conspicuousness is not
an advantage is unjustifiable. The bees gave a decided
preference to the brighter-colored flowers, and the fact
that they subsequently discovered and removed the honey
from the green calyces furnishes no evidence whatever
against the benefit of color contrast.

But a method of experimenting was wanted, which
would permit of varying the conditions under which the
conspicuous and inconspicuous objects were exposed, and
of counting the number of visits to each. This was ob-
tained in the following manner: A small number of
honey-bees were trained to visit for honey an unpainted,
dull-gray board raised upon a support two feet high. A
short time before the honey, which was placed directly on
the board, was wholly removed, a conspicuous and an
inconspicuous object were placed at equal distances from
the board and at a known distance from each other. As
soon as the bees had consumed the honey they began
describing a series of ever widening circles in search of
a new supply until one or both of the above-mentioned

No. 542] COLOR SENSE OF THE HONEY-BEE 97

objects were discovered. The number of visits made to
each in a given time was then counted, and served as a
basis for estimating numerically the value of conspicu-
ousness.

On October 1, 1909, a small number of bees were accus-
tomed to visit the dull-gray board, on which there was a
small quantity of honey. For convenience this board will
be called the feeder. While the bees were busily at work,
I put a blue slide (prepared by placing the floral leaves
of the bee larkspur (Delphinium elatum) between two
glass object slides, 3 X 1 in.), on the center of which there
were a few drops of honey, on the grass of the lawn about
three feet from the base of the feeder; and on a dandelion
leaf three feet from the base of the feeder and five feet
from the blue slide honey was also placed. As soon as
the supply of honey on the feeder was exhausted the bees
began circling in the air. In a few minutes one bee had
found the blue slide, in ten minutes two bees, and in
twenty-five minutes five bees; but none had found the
honey on the dandelion leaf. I now placed beside the
dandelion leaf an apple leaf with a comparatively large
quantity of honey on it, and at the end of forty minutes
one bee found it and a little later a second bee. I doubt
if they would have found it then had they not for some
time previously been flying low searching for honey in
the grass, having from their previous experience with
the blue slide learned to look for it there. In this experi-
ment the advantage was clearly on the side of the con-
spicuous object. It would appear that if two flowers
were blooming at some distance apart, the one bright
colored and the other green, the former would be the more
likely to be pollinated.

On October 3, at 12:33 r.m., I repeated this experiment.
The blue slide, a dandelion leaf, both on the grass, and
the base of the feeder formed the angles of an equilateral
triangle, each side of which was three feet. Honey was
placed on all three as before. Two minutes after the last
drop of honey on the feeder had disappeared three Italian

98 THE AMERICAN NATURALIST (Vou. XLVI

bees found the blue slide. At 12:40 there were eight bees
on the blue slide, but not one had found the honey on the
dandelion leaf. Five minutes later there was one bee on
the leaf.

If bees are guided by odor exclusively in their search
for nectar, and contrast in color with green foliage is no
advantage to flowers, then it would seem as though they
should find a quantity of free honey as readily as when it
is associated with bright coloration. About thirty Italian
bees were accustomed to visit the gray-colored board, or
feeder, which, as previously stated, rested upon a support
two feet high. Six feet from this support and six feet
apart, the three forming an equilateral triangle, were
placed two poles each 43 feet high. On top of one of the
poles was placed a quantity of honey so large that it ran
down on the side, and was visible at a distance of twenty
feet. To the top of the other pole was attached a cluster
of yellow ‘‘immortelles’’ (Helichrysum bracteatum)
gathered many years ago, and which appeared to be
absolutely devoid of scent. Each of the flowers was about
14 inches in diameter and the cluster was 3 inches long
by three inches wide. At 11:10 a.m., the bees were per-
mitted to consume all the honey on the feeding board. In
three minutes there were three bees and one fly on the
flowers, but no insects had found the free honey. In five
minutes there were four bees and one fly on the flowers,
and one bee on the free honey. At 11:20 the latter bee
left for the hive and five minutes later returned; a second
bee also alighted on the side of the pole and began suck-
ing the honey which had run down from above; two flies,
apparently house flies, also came. At the same time there
were six Italian bees on the flowers. At 11:30 a. m., there
were six Italian bees and one fly on the flowers, but only
one bee on the free honey. The flowers not only attracted
the bees earlier than the free honey, but three times as
many of them.

I now transposed the poles. But to the top of the pole
on which there had previously been the supply of free

pee AE

No. 542] COLOR SENSE OF THE HONEY-BEE 99

honey I fastened a single yellow ‘‘immortelle’’ one inch
across. The individual flower enjoyed the advantage of
position since it stood where the cluster had been before.
Honey was placed on all the flowers. At 11:50 a.m.,
there were nine Italian bees and a Syrphid fly on the
cluster of flowers and three Italian bees and one fly on the
single flower. The larger and more conspicuous object
notwithstanding its changed position received the greater
number of visitors.??

The following experiments were made in 1910, and only
black bees were employed. As in the experiments of the
preceding year, the bees were trained to visit the same
dull gray board placed upon a support two feet high. On
September 14, 1910, at 12:40 r.m., the bees were carrying
away syrup of sugar from the feeder. Nine feet from its
base I put out on the grass of the lawn a dried yellow
flower of Helichrysum bracteatum 14 inches in diameter,
containing a small quantity of honey. On the opposite
side of the feeder at a distance of nine feet from its base
I laid a Red Astrachan apple leaf, 2 inches long by 14
wide, on the center of which there was an ample supply
of honey. There were at least twenty-five bees on the
board and later the number increased. At 12:55 they had
wholly consumed the sugar syrup. At 1:07 a bee came to
the flower, but left almost immediately. At 1:10 a second
bee came to the flower, but soon left, and a few moments
later a third visit was made in the same way. No bees
had found the leaf. As the honey was excellent I could
account for the brief stop made by the bees only on the
ground that they were looking for sugar syrup. In the
next experiment this was offered to them.

At 1:20 r.m., I again put sugar syrup on the feeder,
and removed the flower and leaf from the grass. Another
‘‘immortelle’’ 14 ins. in diameter and another Red As-
trachan apple leaf, 2 inches long by 14 inches wide, were
laid on the grass on directly opposite sides of the feeder,

Cf. Miiller’s remarks on Geranium, Epilobium, Polygonum and the
Alsiner.

100 THE AMERICAN NATURALIST [ Vou. XLVI

but six feet instead of nine feet away from its base. The
leaf was on the same side as before, as was also the
flower. Sugar syrup, which is odorless, was placed on
each. At 1:30 the bees finished the syrup on the feeder.

One bee flew almost immediately to the flower, but made
a brief stay. At 1:34 a second bee came and sucked, and
three minutes later a third bee came. No bees had found
the leaf.

Sugar syrup was again put on the feeder, and the
flower and leaf were moved three feet nearer its base,
each now being distant three feet. At 1:47 the syrup on
the feeder was all consumed, but even previously one bee
had come to the flower. At 1:47 a bee flew over the leaf,
but did not alight. At 1:50 three bees came to the flower,
and a moment later a fourth, and afterwards two more.
At 2 r.m., there were three bees on the flower, a fourth
came a little later and then a fifth. No bees had visited
the leaf.

Syrup of sugar was again placed on the feeder. At 2:5
P.M., I put out the yellow flower and apple leaf used in
the irst experiment. On these, it will be remembered,
honey had been placed. They were laid on the grass on
opposite sides of the feeder, each three feet distant from
its base. At 2:10 the sugar syrup on the feeder was all
removed. A bee soon came to the flower, but did not
stop, a second bee came and sucked, a third bee came, but
did not stop, several bees came but did not stop; but at
2:13 there were three bees sucking honey on the flower.
A bee flew slowly over the leaf I thought it would cer-
tainly be attracted by the scent of the honey, but this was
not the case. The experiment was continued a little
longer and one or two more visits were made to the
flower, but none to the leaf.

The results obtained in the four preceding experiments
are deserving of careful attention. While the yellow
flower containing honey and the one containing scentless
sugar syrup were visited many times by bees, the leaves
remained wholly unvisited, though the supply of syrup

No. 542] COLOR SENSE OF THE HONEY-BEE 101

or honey on them was plainly visible at a considerable
distance. According to the reiterated statement of
Plateau all flowers might be as green as their leaves
without their pollination being compromised, and color
and form are of little consequence in comparison with
odor. But the experiments showed that color contrast is
of great value, and in these particular experiments indis-
pensable. If the leaves provided with an ample supply
of honey or syrup could not obtain a single visit under
the conditions described, where a large number of bees
were brought into their immediate vicinity, how little
chance there would be for an isolated plant with small
green flowers growing in a secluded location attracting
visitors! But a bright-colored flower in the same locality
would be much more likely to gain the attention of pollin-
ating insects.

On September 20, 1910, at 2:15 r.m., numerous black
bees were coming to the feeder for honey. At a distance
of three feet away I laid on the grass a bright yellow
flower of golden glow (Rudbeckia laciniata) two inches in
diameter. On the opposite side of the feeder three feet
from its base, I laid the end of a spike of Amarantus
retroflexus about three inches long. The small, pale
green flowers are thickly crowded in panicled spikes. An
ample supply of honey was placed on both. In the course
of fifteen minutes there were 18 visits to the flower of
golden glow and only 8 to the Amarantus cluster. If a
bee flew to either object, but did not alight because of the
large number of bees already there, this was counted as a
Visit.

The bees were again accustomed to visit the feeder.
In the preceding experiment one of the objects had been
placed on the east side of the feeder and the other on the
west. Both the flower of the golden glow and the spike
of Amarantus were now laid side by side on the grass in
the sunshine three feet to the north of the feeder. There
was honey on both. In less than ten minutes there were
fifteen visits to the golden glow and only three to the

102 THE AMERICAN NATURALIST [Von XLVI

spike of. Amarantus. At one time there were five bees
on the golden glow and only two on the spike of
Amarantus.

At 2:45 p.m., I repeated the preceding experiment, but
I placed the flower of the golden glow and the spike of
Amarantus on the south side of the feeder three feet
from its base, but only three inches apart. Honey was
put on both at the beginning of the experiment. In ten
minutes there were 18 visits to the golden glow and 5 to
the green spike of Amarantus. At one time there were
four bees on the flower of golden glow, but only one on
the spike of Amarantus. It often happens when a bee
comes to a flower on which one or more bees are already
at work that they will all fly up in the air and then all or
in part settle down again. Such flights were not counted.
Frequently a bee flew directly to the golden glow as
though it had been seen from a distance.

It will be remembered that Plateau put honey on the
green inflorescence of several species of Chenopodium,
besides other anemophilous flowers, and when it was
found by insects reasserted his oft-repeated conclusion
that winged pollinators are guided to flowers almost ex-
clusively by odor and that color contrast is of little value.
Plateau employed no control experiments, but it appears
from the experiments just described that though the odor
of the honey drew insects to the green inflorescence,
nevertheless it was at a disadvantage because of the
absence of bright coloration.

In several of the experiments of 1909 a blue slide was
used, prepared by placing the leaves of the perianth of
the bee larkspur (Delphinium elatum) between two glass
object slides tied firmly together with black silk. It
might perhaps be objected that the scent of these floral
leaves would escape through the narrow crack between
the two glass slides. While I think this improbable, and
that in any event it would be so slight as to bear no com-
parison with that of the honey placed upon the upper
glass slip and, therefore, would exert no influence on the

No. 542] COLOR SENSE OF THE HONEY-BEE 103

behavior of the bees, still it seemed desirable to test the
matter. For this purpose the following experiment, clos-
ing the series of 1910, was performed on September 23.
A blue slide was prepared as described and the edges
were sealed with several applications of gold size, the
odor of which is no doubt unpleasant to bees. The blue
slide, a dandelion leaf, and the base of the feeder formed
the angles of an equilateral triangle, each side being
three feet in length. As the weather was becoming
colder the bees were not flying freely. An ample supply
of honey was put on the blue slide and the leaf, which
were laid on the grass of the lawn at 9 r.m. At 9:20 the
honey on the feeder was entirely consumed. Presently
a bee hovered over the blue slide, but did not alight.
Another bee hovered over the blue slide for a long time
and finally alighted. A second, third and fourth visit was
made by bees at intervals. At 9:40 I discontinued the
experiment. No attention had been paid to the honey on
the leaf, though in the sunlight it could be seen for a long
distance. The hesitation of the bees at first may have
been caused by the repellent odor of the gold size. Bee-
keepers never paint their hives inside, as the scent of
paint is believed to be disliked by bees. The blue slide
and the leaf were left in position and when twenty minutes
later I examined them again all of the honey had been
removed from the slide, but that on the leaf appeared to
be untouched. Evidently the only factors which had in-
fluenced the bees in the previous experiments were the
honey and the color.

Of the series of experiments performed in 1911 only
three will bedescribed. A few observations were thought
desirable in which one or two bees were employed instead
of a larger number, in order that the behavior of an indi-
vidual bee might be followed when given the choice be-
tween a conspicuous and an inconspicuous object. A few
bees were accustomed to visit a glass slide for honey. |
While they were absent at the hive, the slide was re-
moved and a large rhubarb leaf was laid in its place. ©

104 THE AMERICAN NATURALIST [ Vou. XLVI

About two inches from the base of the leaf there was put
a quantity of amber-colored honey sufficiently large to
form an oval mass, which could be seen in the shade at a
distance of twenty feet. Twelve inches from the honey
and a few inches from the apex of the rhubarb leaf there
was placed a bright red flower of the Zanzibar balsam
(Impatiens sultani), an inch in diameter, on which there
was a small amount of honey.

A bee returning from the hive went directly to the red
flower, where it took up its load and flew away.

A bee came to the red flower. Two more bees came and
were impounded. The first bee left for the hive.

A bee returned to the flower. A second bee came, both
flew up in the air, and one of them went to the mass of
honey but soon returned to the flower. The first bee left
for the hive. I attempted to impound the second, but it
escaped.

A bee came to the flower, and after five minutes re-
turned to the hive.

The bee returned to the flower. A second bee came, and
hovered in the air for some time, but finally settled by
the bee on the flower. Both bees left for the hive. Both
bees returned to the flower, and when they again left I
discontinued the experiment. The rhubarb leaf was re-
moved and the bees were given honey on a glass slide.

It seems impossible to explain the behavior of the bees in
this experiment on the supposition that they were guided
chiefly by odor. In view of the large quantity of honey
and its easy accessibility there would have been no occa-
sion for surprise had the bees given it much greater
attention.

After carefully removing the honey from the rhubarb
leaf I placed near its apical end four flowers of the Zanzi-
bar balsam, forming a bright red square. On one petal of
each flower there was a small drop of honey. Ten inches
away near the base of the rhubarb leaf I put a single
petal of a balsam flower on which there was a large drop
of honey. While both bees were away I removed the

No. 542] COLOR SENSE OF THE HONEY-BEE 105 .

glass slide and substituted the rhubarb leaf, reversing
its position, however, so that the small object was where
the larger had been before.

Both bees returned to the cluster of four balsam
flowers. One of them presently flew over to the single
petal, but soon returned to the cluster; later it again
went to the petal and again returned. Both bees left
for the hive. :

A bee returned to the cluster, did not alight, but flew
over to the petal and sucked. When the second bee re-
turned it disturbed the bee on the red petal, and both
went to the cluster. One of the bees left for the hive.

A bee came from the hive to the cluster. One of the
bees then flew over to the petal but did not alight, re-
turned to the cluster. Both bees left for the hive.

Both bees returned to the cluster. One of them left for
the hive and on its return went to the petal. The bee on
the cluster left for the hive.

While the cluster of four red flowers received the
greater number of visits, as would be expected, more
attention was given to the drop of honey associated with
a red petal than was received by the larger oval mass of
honey alone in the preceding experiment.

In a series of interesting experiments with cotton
flowers where the visitors were chiefly a species of Melis-
sodes (M. bimaculata), recently described by Allard, it
was observed that when a flower was partially screened
by leaves the attention it received decreased ; and when
the petals were masked on both sides with sections of
green leaves the flower was ignored entirely.” This re-
sult was confirmed by the following experiment. On a
cloudy, windy day while a number of black bees were
visiting the feeder for honey, I placed on the grass two
red flowers of the Zanzibar balsam; each was five feet
from the base of the feeder and their distance apart was

3 Allard, H. A., ‘Some Experimental Observations concerning the behav-
ior of Various Bees in their Visits to Cotton Blossoms,’’ AMER. NAT., 45, :
615 and 672, 1911. -

106 THE AMERICAN NATURALIST [Vou. XLVI

two feet. There was a small quantity of honey on both.
One of these blossoms I screened with dandelion leaves
on the side toward the feeder, but it was visible in every
other direction. Some time after the honey was all con-
sumed on the feeder two bees flew over the unconcealed
flower but did not alight. A wasp (Vespula victua
Sauss.)24 found it and at the end of half an hour it was
visited by a bee. The partially concealed flower received
no attention. During this experiment the bees seldom in-
spected objects on the lawn though they frequently flew
to where I was sitting, ten feet away.

The conclusion derived from a study of the phylogeny,
ecology, distribution and fertilization of green flowers
that they are at a disadvantage in attracting insects be
cause of their color was fully sustained by a long series
of experiments, in which honey-bees were given the choice
between a green and a bright colored object placed on a
green background, or between a conspicuous and an in-
conspicuous object. In the experiments described both
black and Italian bees were employed, the number of
which varied from one to fifty. The observations ex-
tended over portions of three seasons. Conspicuous and
inconspicuous objects were in some instances placed dia-
metrically opposite to each other at varying distances,
in other cases side by side or a few feet apart. In six
experiments there were no visits to the inconspicuous
object; while in the other experiments the number of
visits to the conspicuous object was much larger than to
the inconspicuous object, usually twice or three times as
large. The theory that bees in gathering nectar are in-
fluenced only by the olfactory sense and not by color or
form does not afford a satisfactory explanation of the
facts presented. If, however, bees are guided by the
sense of vision as well as by that of smell, then their rela-
tions both past and present to green flowers are not
difficult to understand. To reject a natural and wholly

“For the determination of this species I am indebted to Mr. S. A.
Rohwer.

No. 542] COLOR SENSE OF THE HONEY-BEE 107

satisfactory explanation of their behavior in favor of an
improbable hypothesis has the appearance of shunning
the truth in a vain search for novelty.

CoNCLUSIONS

Green flowers are not well adapted to entomophily, and
many species, possibly all, have been derived by retro-
gression and degeneration from larger more highly de-
veloped entomophilous forms. They are usually small,
or even minute, and are often incomplete, while ane-
mophily and autogamy prevail. Entomophilous green
flowers are as a whole sparingly visited by insects belong-
ing to the less specialized families, and as a rule retain
the power of self-fertilization.

The fact that insects have been observed feeding on
over-ripe or decaying fruit, or the glandular secretions
of the vegetative organs of plants, or the excretions of
Aphidide on foliage, or greenish or brownish flowers, or
dull-colored receptacles which have contained sugar or
sweet liquids, affords no evidence that conspicuousness
is not an advantage to entomophilous flowers. Any sur-
face, whether it is bright or dull-colored, on which there
is nectar or honey, will be freely visited by bees for
stores after these liquids have once been discovered ; but
they will not be discovered as quickly on a surface which
does not contrast in hue with its surroundings as on one
which does so contrast.

The experiments and observations of Plateau on green
or greenish flowers in the absence of control or compara-
tive observations are fallacious, as pointed out by Knuth,
and do not prove that ‘‘all flowers might be as green as
their leaves without their pollination being compro-
mised.’’

When honey-bees are given the choice between a con-
spicuous and an inconspicuous object under similar con-
ditions, they exhibit a preference for the former. This
preference is sufficiently marked to account for the de-
velopment of color contrast in flowers. oo

SHORTER ARTICLES AND DISCUSSION

IS THE CHANGE IN THE SEX-RATIO OF THE FROG,
THAT IS AFFECTED BY EXTERNAL AGENTS,
DUE TO PARTIAL FERTILIZATION?

In a review in this journal (XLV, 1911) of certain experi-
ments by Kuschakewitsch!? on frogs’ eggs in which by delaying
fertilization for 89 hours he obtained 100 per cent. of males,
I pointed out that unless more than half of the eggs were fèr-
tilized the interpretation of the 100 per cent. ratio might be mis-
leading. For should the delay act more injuriously on one kind
of egg than on the other, assuming two kinds to exist, the result
might mean only selective destruction by an external agent
rather than a change in sex of the eggs. I found no explicit
statement in the section of Kuschakewitsch’s paper dealing with
these results to show whether or not all of the eggs had been
fertilized, but in a recent rejoinder? to my review Kuschake-
witsch points out that he had stated that practically all of the
eggs (‘‘so gut wie alle Eier’’) were fertilized and developed.
This information is given in an appendix which I had overlooked.
His statement completely sets aside the possibility of the sug-
gestions that I made, but leaves the explanation of his results as
obscure as before.

The details of the principal experiment and of some of the
others are of interest. A pair of copulating frogs were caught
at 12:00, midday, May 31. The female began to lay at once.
At 6:00 p.m. the male was removed. On the 4th of June at
8:00 p.m. the eggs that had remained in the uterus of the female
were artificially fertilized. They are recorded as 89 hours old
at this time. Practically all segmented, and only 5 died at the
gastrulation stage. From this lot, 434 eggs hatched. Only 12
deaths occurred later. Three hundred of the tad-poles were
examined at the time of or after metamorphosis. Of these, 299
were males, and one was a bilateral hermaphrodite.

There can be little doubt, therefore, that, in some way, delay in
fertilization has caused practically all the eggs to produce males;
and the evidence is the clearer since the eggs fresh laid, fertil-
ized by the same male, produced 55 males and 53 females.

It may seem futile, therefore, to attempt to explain this result
in any other way than as the result of the action of the environ-
ment on the sex of the egg. But how has the environment

* Festschrift, R. Hertwig, Bd. II, 1910.

* Anatom. Anzeiger, 1911.

108

No. 542] SHORTER ARTICLES AND DISCUSSION 109

acted? The evidence that sex is regulated by an internal mech-
anism has become so strong in recent years that until the action
of the environment is made clear one may well hesitate to accept
the case as showing that sex is actually changed or produced by
an external agent. Curiously enough, every one seems to have
overlooked still another possibility that may solve the difficulty.
The delay in the fertilization may cause the polar spindle to stick
to the surface of the egg so that it fails later to take part in the
development, in which case the sperm nucleus alone would pro-
duce the nuclei of the embryo. Or, on the other hand, the delay
may cause the early stages in the formation of the female pro-
nucleus to progress so far that after fertilization the sperm nu-
cleus may be excluded in part or entirely from the development.
In either case the presence of a single nucleus would be expected
to give rise to a male. It is significant in this connection that
the changes described by King that affect the sex-ratio of the
frogs’ eggs produce a higher percentage of males.

There is another curious fact in relation to sex-determination
in the frog. Pfliiger described a high percentage of hermaph-
rodites amongst the tadpoles. Kuschakewitsch has given a de-
tailed account of the development of the hermaphroditic glands.
Most or all of these organs are later transformed into testis.
In general it may be said that eggs from a pair give either equal
numbers of males and females; or a mixture of males, females,
and hermaphrodites; or all hermaphrodites (potentially males).
It is possible that the pseudo-hermaphroditic condition may be
connected with the failure of one of the two pronuclei to take
part in the development.

If the explanation that I have suggested is correct we might
expect to find evidence in its support from the number of chro-
mosomes in the tadpoles that develop from these late fertilized
eggs. This would be expected if it is the male pronucleus that
gives rise to the nuclei of the embryo. But if it is the female
pronucleus that is responsible for the result, the number of chro- —
mosomes in the cells of the embryo might be haploid or diploid —
depending on whether the second polar body was, or was not —
given off. At any rate, this suggestion should be put to the
test of observation before we conclude that sex may be deter-
mined by external agents. If the view here suggested prove
true, sex is still determined by an internal factor in the same

sense that the sex of the bee’s egg is determined by the presence

of ane or of two promot o a

NOTES AND LITERATURE
HEREDITY

A SUBJECT of vital importance to the theory of heredity is the
behavior of the chromosomes during the life history of the cell,
and especially during the process of cell division. This subject
has received an enormous amount of attention from investigators
but there is far from unanimity amongst cytologists as to the
actual phenomena of cell division, not to mention the signifi-
cance of these phenomena. Realizing that the heterotypic di-
vision in gametogenesis is the critical point in the life history of
the organism, so far as the theory of heredity is concerned, at-
tention has been concentrated very largely on this division.
This is in some respects unfortunate. A good many investigators
who have studied this division have attempted to interpret the
phenomena observed without full knowledge of the behavior of
the chromatic elements in ordinary somatic divisions, and have
attributed to phenomena observed in the heterotypic division
very special meaning for inheritance, when these same phenom-
ena are regular occurrences in all divisions, and hence are to be
interpreted in their relation to growth rather than to reproduc-
tion. :

The writer is not a cytologist, and realizes fully that his
opinions on cytological questions will not be regarded seriously,
especially by those who have worked at problems of this char-
acter until they have gotten fixed in mind certain theories as to -
the meaning of the phenomena observed. Nevertheless, he has
given careful attention to published results of investigations of
this character, and has been driven by study of these results to a
particular interpretation of the principal phenomena reported.
It has seemed to me for some years that the double spireme so
often reported in the heterotypie division, and so often inter-
preted as a conjugation of homologous chromatin elements, al-
though this double spireme occurs in somatie divisions, appar-
ently quite as generally as in the heterotypic, is nothing
more than the expression of a division of chromosomes which
really occurs at least as early as the resting stage following the
previous nuclear division. It appears that this division may
begin at an even earlier period.

110

No. 542] NOTES AND LITERATURE iit

Fraser and Snell, in their paper on ‘‘The Vegetative Divisions
in Vicia faba,” present important evidence on this point. They
show very clearly that the division begins in the teleophase of the
previous division. This beginning of chromosome division in
teleophase had earlier been noticed by Gregoire, and by Stromp,
as pointed out by Fraser and Snell, but these earlier investiga-
tors had not perceived the meaning of this phenomenon. Fraser
and Snell were able to follow the life history of the chromosomes
through their complete history from one teleophase to the next
(in root tips and in other somatie parts, as well as in the game-
tophyte stages), and they show clearly that the double nature
of the spireme is due to splitting which begins as the daughter
chromosomes congregate at the poles in the teleophase of the
previous nuclear division.

That this double spireme is not due to the approximation of
two elements, one representing maternal and the other paternal
chromatin, is further shown by the fact that in the pollen cell,
where the chromosome number is haploid, and hence where there
can be no question of union of elements from the two parents,
exactly the same phenomena occur.

Another very interesting fact shown by these investigations
(of Fraser and Snell) is that some of the elements which be-
have as single chromosomes, so far as their distribution on the
spindle and to the poles is concerned, are made up of segments
united end to end, as if two or more small chromosomes were
united more or less closely into a larger one. The authors point
out the possible significance of this fact for Mendelian coupling,
and suggest that it may also be of significance in connection with
the fact that in some species more Mendelian factors have been
observed than there appear to be chromosomes.

East and Hayes have recently published the results of ex-
tended investigations on inheritance in maize.” After discuss-
ing the taxonomy of the group and pointing out the adaptability
of maize to genetic investigations (or the lack of such adapta-
bility), the authors give an excellent résumé of former investiga-
tions with this interesting group of plants.

A brief account of their results follows. Amongst endosperm
characters they found that starchiness (S) is dominant to its

*Ann. of Bot., 1911 845-856.

7B. M. ong pe Hayes: Bul. No. 167, Conn. Exp. Sta.—“ Inher-
itance in Maize.’’

112 THE AMERICAN NATURALIST (Vou. XLVI

absence, non-starchiness (s). When S came from the ¢ parent
zenia appeared in all cases. All F, seeds showing zenia proved
to be heterozygous. No extracted recessives of the F, generation
ever proved to be heterozygous. ‘‘From this one may conclude
that the second male nucleus that fertilizes the endosperm nu-
cleus always bears the same characters as the first male nucleus
- that fertilizes the embryo nucleus, or egg.’’ A few seeds, all
heterozygous, were part starchy and part not; 7. e., one side was
starchy. The authors consider that this confirms Correns’s view
that, in such cases, the second male nucleus did not fuse with the
endosperm nucleus but that each developed separately.

One semi-starchy ear occurred, grown from a non-starchy seed.
The authors suggest two possible causes for this phenomenon.
Hither there is an incomplete segregation, resulting in contamin-
ation of a gene by its allelomorph, this contamination, by selec-
tion, being capable of accumulative effects, or the semi-starchy
ear arose as the result of a progressive variation. They point
out that the infrequency of this phenomenon is an argument
against the theory of partial or incomplete segregation, and
incline to the idea that it is a case of progressive variation. The
data presented certainly favor this interpretation of the case.

YELLOW AND Non-YELLOW ENDOSPERM

Two independent factors for yellow were found, each capable
of producing yellow endosperm. The colors produced by these
two factors appear to be the same. The pigments occur in
rhombic plates, and are insoluble in water, but soluble in ether,
chloroform, ete. They appear to be related to the anthochlorins.

Some crosses between yellow and white gave the ratio 3:1, due
to the presence of only one of the factors for yellow. Others.
gave the dihybrid ratio, Yellow appeared as xenia in the hybrid
seeds (seeds which produced the F, plants).

Yellow was dominant, but imperfectly so under certain con-
ditions, so that, in certain crosses in which the grains had soft
starch at the tip the heterozygotes could be distinguished from
the homozygotes. In other crosses yellow was completely
dominant.

The same original ear of some of the parent stocks had some
seeds containing both factors for yellow, while other seeds on the
same ear had only one of these factors.

The two yellow factors together generally gave darker yellow
seed than one factor alone.

ag
gates oa an

No. 542] NOTES AND LITERATURE 113

PURPLE AND NON-PURPLE ALEURONE CELLS

The experimental data relating to crosses between purple and
non-purple races indicate two factors, P and C, which, when to-
gether, produce purple color in the aleurone layer. In some of
the non-purples used one of these factors was missing, in others.
both. In certain combinations one of these factors alone pro-
duced faintly colored purple. In most of the crosses splashed
purple occurred, part, but not all, of the heterozygotes having
this peculiarity. It was not hereditary, but behaved in subse-
quent generations as pure purple.

In one family red aleurone occurred in F,. It appeared to be
due to the interaction of two factors R and C.

In one family of this cross (purple non-purple) the F, gen-
eration gave purples, reds, and non-purples in the ratio 12:1: 3.
The actual numbers were 1,843:188:545. The ratio 12:1:3 did
not occur in F,, but instead one fifth of the ears bearing F,
grains gave the ratio 9:3:4. Several possible hypotheses to
explain these anomalous results are discussed and discarded,
amongst them Bateson and Punnet’s hypothesis of the formation
of gametes in the ratio TAB : 1aB:1Ab: Tab. The data from F,
agree best with the assumption that the constitution of the two
parent types was pcR and PCR respectively, but the F, ratio
is not explained by this hypothesis. The authors leave the ques-
tion as to the real explanation an open one. The reds in this
family were different in color from the reds of the family pre-
viously mentioned. Not only that, but all the F, reds found in
this family proved to be homozygous, indicating that both par-
ents possessed the factor F.

Other families of this cross gave results that indicated the
presence in the non-purple parent of a factor which more or
less completely inhibited the development of purple. Some non-
conformable results were found, due probably to the presence of
other factors, one at least of which appeared to modify purple
by making it lighter.

In the above crosses xenia was found as follows: when non-
starchy races were fertilized by pollen from starchy races (no
xenia appeared in the reciprocal cross) ; when non-yellow endo-
sperm is crossed with yellow endosperm. In this case xenia al-
ways appeared when yellow was used as the male parent. It
also appeared in the reciprocal cross when the grains of the fe-
male parent had extensive development of soft starchy endo- —

sperm at the end of the grain, as in these cases the he

114 THE AMERICAN NATURALIST [ Vou. XLVI

yellow was lighter in color than the homozygous, and hence dis-
tinguishable from it. In crosses between yellow and non-yellow
endosperm, when the non-yellow endosperm was entirely soft
(not corneous), as in the so-called flour corns, xenia appeared in
all cases, for reasons just stated. When the endosperm of the
non-yellow parent was entirely corneous, as in the popcorns, xenia
usually occurred only when yellow was used as the male parent,
though in a few instances it was perceptible when the cross was
made the other way. Xenia also occurred when purple or red
aleurone was crossed with non-purple, or non-red, when the pur-
ple or red was used as the male. The only other case in which
xenia was observed was in crosses between white and red (or
purple) when the white (male) parent carried an inhibiting
factor for purple and red. Sometimes the reciprocal cross shows
xenia, since the inhibition of red or purple is not always com-
plete. The following law regarding xenia is formulated by the
authors: when two races differ in a single visible endosperm
character in which dominance is complete, xenia occurs only.
when the dominant character is the male; when they differ in a
single visible endosperm character in which dominance is in-
complete or in two characters both of which are necessary for the
development of the visible difference, xenia occurs when either
parent is used as the male.

Correns’s conclusion that where xenia occurs the seeds showing
it are always hybrid is confirmed. This shows that Mendelian
segregation must occur previous to the division of the pollen nu-
cleus. The authors found no case in which a seed showing no
xenia where it is to be expected proved to be a hybrid; i. e., the
hybrid in which xenia is to be expected always showed xenia
though like Webber and Correns they found seeds showing xenia
on only one side. This is interpreted as the result of the inde-
pendent development of the endosperm nucleus and the second
male nucleus.

In crosses between podded maize (maize having each grain
covered by husks) with common maize, the pod character proved
to be a dominant Mendelian factor, which segregated perfectly
in

Red sap colors appear in maize in the pericarp, the cob, the
husks, the silks, the glumes, and in the anthers.

Red pericarp (R) without red on the cob or in the silk, crossed
with white pericarp gave three reds to one white in F,, the segre-
pakn being perfect.

No. 542] NOTES AND LITERATURE 115

‘An ear of corn was found where only white corn had been
planted, one side of which produced grains with red pericarp,
the other white or striped with red. The red here seemed to be
due to the same factor as in the case just noticed (R). Red
grains from this ear gave red and white ears in equal numbers.
The white and striped grains gave white ears and ears with a few
striped red seeds in equal numbers. A selfed red ear in this
generation gave three reds to one’ white in the next. The orig-
inal red and white (or striped) ear is accounted for as a somatic
variation, part of the ear varying from white to red, the remain-
der from white to striped. In this family red cob is perfectly
correlated with red pericarp.

Two other red pericarp colors, apparently independent of the
above, were found. One is a dark red occurring in stripes which
radiate from the point of the attachment of the silk to the grain.
The other is a dirty red, more abundant at the base of the grain,
and nearly wanting at the tip. It occurs in Palmer’s Red
Nosed Yellow variety. It is completely coupled with red silks.

Two other red pericarp factors were found. They are very
similar, but not allelomorphiec to each other. They give a rose
red pericarp, but do not develop except in sunlight. They are
barely perceptible on ears covered by heavy husks.

Red cob color proved to be dominant to white and the cross
segregated in a 3:1 ratio. Red cob appears not to be correlated
with any of the red pericarp colors.’

Red silk color presented some difficulties, and the data are not
analyzed. In some instances a 3:1 F, ratio was obtained, in
others a third type with red hairs on a greenish-white silk oc-
curred, the F, numbers being 198 reds, 29 greenish-whites with
red hairs, and 94 greenish-whites. Red silks may occur with no
other red on the plant.

Red glumes were always accompanied by red in other parts of
the plant, though in one race the only other red was in the silks.

The question whetherall these reds are due to one or to dif-
ferent genes is discussed most interestingly (pp. 109-10), but
the discussion is too long to quote here.

It is gratifying to note that these authors are not afraid to
mention the chromosomes in connection with Mendelian factors.
For several years an inexplicable obsession seems to have pos-
sessed biologists in this matter. Apparently every one regards

? But see reference to Emerson’s results below. 3 ae

116 THE AMERICAN NATURALIST [ Vou. XLVI

it as highly probable that these cell organs are in some way
responsible for Mendelian phenomena, yet a large number of
biologists seem to be afraid to refer to them in this connection.

Crosses between flint and dent varieties indicated that the dif-
ference between these two classes relates to two factors in some
ease, especially when the dent parent has considerable corneous
endosperm, and to two or more factors, especially when the dent
parent has little corneous starch. This seems to be another case
where several similar factors exist, as found in earlier inves-
tigations by East, by Nilsson-Ehle, and by Shull.

Crosses between races having different modal numbers of rows
of grain on the cob indicate clearly that several similar factors
are here concerned. The evidence that segregation occurs is,
the authors believe, conclusive. The data could not be definitely -
analyzed because of the fluctuating variation of the various bio-
types, and the small differences between adjacent biotypes.
Height of stalk and length of ear behave similarly. Apparently
size of grain does the same.

Irregular rows of grain occur both as fluctuations not inherited
and as a hereditary characteristic. Should the percentage of
irregular rows be higher than about 4 per cent. the authors think
the irregularity is probably hereditary.

This paper closes with a discussion of the inheritance of var-
ious abnormalities found during the progress of the work.

The Journal of Genetics for August, 1911, contains some arti-
cles of unusual interest. R. N. Salaman presénts the results of
an investigation on the inheritance of the peculiar physiognomy
known as the Jewish face. He shows that this is a simple Men-
delian character. It is distinctly recessive to the ordinary Eu-
ropean (Gentile) physiognomy, though the hybrids sometimes
show traces of the Jewish face, especially late in life. On the
other hand, this peculiarity is dominant over the Pseudo-Gentile
face sometimes seen amongst the Jews; also to the Gentile phy-
siognomy of the Moors and certain ithe Mediterranean peoples.
We may explain this in terms of the presence-absence hypothesis
by saying that the distinctive type of face seen amongst the
Hebrews is due to the presence of a gene, while in the peoples of
northern Europe there is a gene which inhibits the Jew face char-
acter. The data presented, while not extensive, seem to be quite
conclusive that the character segregates as a so-called unit char-
acter. Salaman’s paper is an exceedingly clear presentation of
data, and is written in a style that is attractive and readable. |

No. 542] NOTES AND LITERATURE 117

Bateson and Punnett* give the results of their study of the
inheritance of the peculiar black pigmentation in the skin, perios-
tium, and other tissues of the silky fowl. While some excep-
tions occur, the results on the whole are in agreement with the
assumption of a pigment factor, P, and an inhibiting factor J,
the latter exhibiting the phenomenon of spurious allelomorphism
with the female sex factor. The authors suggest that the excep-
tions found may be due to failure of the repulsion supposed to
exist normally between the female sex factor and the inhibiting
factor,

While a large part of the work on which Mendel’s principles
of heredity depend has been done with pigments, very few in-
vestigations have been undertaken in order to determine the
connection between the phenomena of inheritance of these pig-
ments and the chemical reactions which underlie these phenom-
ena. This is quite natural, since few of those who have con-
ducted the investigations relating to Mendelian inheritance have
had the training, and hence the opportunity, to study the chem-
ical side of the question. Likewise, those relatively few individ-
uals who have become well versed in the highly complex and
difficult subject of physiological chemistry have seldom had any
direct interest in the phenomena of inheritance. The wisdom of
an endowment for an all-sided research of heredity such as the
Carnegie Institution of Washington has provided at Cold Spring
Harbor is manifest in the fact that Dr. Davenport has been able
to institute research on both sides of the question. The results
secured by Mr. Gortner in his study of the origin of melanin,
and its relation to the phenomena of Mendelian inheritance will
be eagerly read by students of Mendelism. In the May number
of the Journal of Biological Chemistry Mr. Gortner gives some
exceedingly interesting results of his work. He shows that the
body filling of the meal worm (Tenebrio moliter) contains two
oxidases, a laccase-like enzyme, and a powerful tyrosinase. Also
that there is a chromogen present in the larva which, when acted
upon by the tyrosinase, gives a series of colors ending in a black
melanin-like body. Larve killed by ether developed pigment
when left exposed to air, but when the air was excluded by car-
bonic acid or nitrogen no pigment developed. When the larvæ
were heated (in water) sufficiently to destroy the activity of the
tyrosinase but not that of the laccase, no pigment formed.

The oxidase was evidently present in relatively large amounts, oe

* Jour. of Gen., Aug., 1911, pp. 185-203.

118 THE AMERICAN NATURALIST [ Vou. XLVI

but the chromogen only in small quantities. The results indi-
cate that the chromogen is formed slowly and used as formed.

In the September number of the same journal® the same author
deals with the nature of dominant and recessive white. He
shows that, in so far as the presence or absence of pigment is
concerned, these two types of white, in certain mammals, are as
indistinguishable to the chemist as they are to the breeder. He
accepts the view that pigment is formed by the action of an
oxidase on a chromogen, and points out that dominant white
arises from the presence of a third body which prevents the reac-
tion between the oxidase and the chromogen. This might occur
in three ways: (1) the third substance, such as orcin, resorcin,
phloroglucin, or other substances of similar nature, may act on
the chromogen and thus prevent its oxidation; (2) it may itself
be oxidized by the tyrosinase, thus preventing action on the
chromogen; (3) it may act as a true anti-oxidase, and in some
manner inhibit the action of tyrosinase.

The author gives abundant data to show that alternatives 1
and 2 are excluded in the cases with which he worked. Hence
the action must be of the third type—an inhibitory action. He
shows that dominant whites contain no pigment lacking in re-
cessives. He also shows that ‘‘ aromatic compounds which carry
two hydroxyl groups in the meta position to each other are
capable of inhibiting the action of tyrosinase on tyrosin.” If,
then, such a substance should occur in the animal body a domi-
nant white would result. Recessive white is presumably due to
the absence of either the chromogen or the oxidase, while at the
same time no inhibiting factor is present. From this it would
appear that an albino might be dominant to color if it carried the
inhibiting factor, yet it would require considerable work to dis-
tinguish between such an albino and a true dominant white,
i. e., by its genetic behavior.

To explain the occurrence of recessive whites which are not
albinic, and which are dominant in some crosses, we have the
following considerations. Orcin inhibits the action of tyrosinase
on tyrosine, but is itself oxidized by a specific enzyme. If we
represent tyrosine by C, tyrosinase by T, orcin by O, and the
specifie enzyme which oxidizes orcin by P, then an individual
containing only C and T would be colored, CTO would be white,

R. A. Gortner: ‘Studies on Melanin: IIT. The Inhibitory Action of

Certain Phenolie Substances upon Tyrosinase.’’ Jour. of Biol. Chem., Vol.
X, No. 2, Sept., 1911.

No. 542] NOTES AND LITERATURE 119

the cross CTO X CT would be white (i. e., white would be dom-
inant), the cross CTO X CTP would be colored (white reces-
sive), while the cross CTO TP (both white) would be colored.

It appears, therefore, that Gortner has been able to construct a
theory that renders intelligible the behavior of all kinds of white
color in inheritance, and to give what seems to be a very plaus-
ible chemical explanation of all these cases. We, of course,
already understand why the cross of two whites of types Ct and
cT should give color and why both of the latter types of white are
recessive to color.

Readers of the Narurauist are familiar with my contention
that in order to show that the chromosomes, as a whole, are not
responsible for Mendelian characters, it must be shown that more
independent dominant characters can be put into a single indi-
vidual than there are pairs of chromosomes. The fact has been
repeatedly cited that more Mendelian characters have been found
in Pisum than there are pairs of chromosomes, and the claim is
made that this disproves the chromosome theory of Mendelism.
I have repeatedly shown that this is not the case. There might
be a thousand Mendelian characters demonstrated for Pisum,
but until it is shown that more than six of them are genetically
independent, the chromosome theory is not affected thereby. It
has heretofore been assumed that two factors, each of which
when crossed with its absence behaves as a so-called unit char-
acter, are genetically independent of each other. This assump-
tion appears to be involved in the ‘‘presence-absence’’ hypo-
thesis. This hypothesis, as usually applied, seems to imply the
presence or absence of a particular body in the cell, which body,
when present once, passes into only one gamete, and that there
are as many such genetically independent bodies as these are
Mendelian factors. When correlation of Mendelian factors
occurs, it is assumed to be due to the adhering together of twe
of these bodies; likewise when the so-called spurious allelomor-
phism occurs, it is because two of these bodies repel each other.

I wish here to repeat what I have often said before, that the
fact that two factors each behaves as an allelomorph to its ab-
sence does not prove them to be genetically distinct. The occa-
sion for this repetition is some pertinent evidence which has just
been presented by some exceedingly interesting work of Profes-
sor Ronse ’s.° He found that red cob (with white pericarp)

R. A. Emerson: ‘‘Genetie Correlation and Spurious Allelomorphism in
States 24th An, Soha ik er et Ma, p

120 THE AMERICAN NATURALIST [ Vou. XLVI

X absence of red gives the usual Mendelian phenomena of 3 red
cob: 1 white in F,. Likewise, the cross red pericarp (with white
cob) X absence of red gives 3 red pericarp: 1 white. Thus each
of these types (factors) behaves as an allelomorph to its absence.
Hence they should be due to genetically independent genes. But
such is not the case; they are allelomorphic to each other.

Some students would say that this is because they repel each
other. But this explanation does not satisfy in this case, for it
can hardly be doubted that red cob in the one case and red peri-
carp in the other are due to the same cause, acting differently
in the two cases. I am of opinion that many similar cases of
factors behaving as allelomorphs to their absence will be found
to be also allelomorphie to each other. Such cases have usually
not been looked for. Quite a number of them have been reported,
and I hope some time to be able to bring them all together for
reference. Emerson’ gives a case in beans, in which a variety
with green leaves and green pods was crossed with another hav-
ing green leaves and yellow pods. F, consisted of three of the
former to one of the latter. Here green was allelomorphic to
absence of green. Later he crossed two varieties, one with green
leaves and green pods, the other yellow leaves and pods. F, com
sisted of three of the former and one of the latter. Here green
was again allelomorphic to its absence. These two crosses ap-
parently show that yellow pods with green leaves and yellow
pods with yellow leaves are not genetically distinct. Yet if yel-
low had been dominant in both these cases, it would have been
the usual custom to consider that the two different yellows were
independent genetically, because each was allelomorphic to its
absence. It would be interesting to know how the two yellows
would behave if crossed. It would not be at all surprising to
find these two ‘‘absences’’ exhibiting the phenomenon of spur-
ious allelomorphism. The case would be still more interesting if
a variety could be found with yellow leaves and green pods.
W. J. SPILLMAN ©

(To be continued)
Th a |

VOL. XLVI, NO. 543”

The American Naturalist

MSS. intended for publication and books, etc., perseae for review should be
sent to the Editor of THE AMERICAN NATURALIST, Garrison-on-Hudson, New York.
Articles containing research work bearing on the problems of a evolu-
tion = a welcome, and will be given in publicatio
hundrea reprints of contributions are supplied to authors free of charge.
Fistor rs reprints will be vidide at cost.
advertisements should be sent to the publishers. The
subscription price is four dollars a year. Foreign postage is fifty cents and
Canadian postage twenty-five cents additional. The charge for =~ copies is
thirty-five 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! tter, April 2, 1908, at the Post Office at L ster, Pa., under the Act of
Congress of Mareh 3, 1879.
THE BULLETIN—For bargains in Ethnol | TENTH EDITION.

=

{eal and Prehistoric Specimens, Books on Matoral | THE MICROSCOPE,
ane a — ve eto. See THE rang eae a "H swied GAGE, of Cornell Tuiversity. Over
: vO
lanj antayo pripe Sat

above 250 figures in this new
4 Duke St, Adelphi —London—England EEE

COMSTOCK PUBLISHING CO., Ithaca, N. Y-

BI RDS’ EGGS | | University of Colorado as
Mountain Laboratory —

W.F- OBENBERE, oo | Fourth summer session June 24 to At
l Hil, LE Song ie e Courses in Au

THE
AMERICAN NATURALIST

Vou. XLVI March, 1912 o ge eae

PROBLEMS OF EVOLUTION AND PRESENT
METHODS OF ATTACKING THEM?

PROFESSOR EDWIN G. CONKLIN
PRINCETON UNIVERSITY

Tue problems of evolution have been much the same
from Darwin’s day to this, but the present methods of
attacking them are in many respects different from those
which prevailed a generation ago. One great problem
with which the earlier naturalists were concerned, viz.,
the fact of evolution, is by common consent, no longer a
problem; if it has not been demonstrated that the living
world arose through evolution, it has at least been ren-
dered so probable that demonstration could add little to
our certainty. And yet we should all like to see the
demonstration of evolution on a large scale, such as must
have been operative in the past history of living things,
but we have little reason to hope that such a demonstra-
tion will soon be made. .

The enduring problems of evolution concern the means
or factors of transmutation. Here also the old method of
attack, viz., observation and induction, led to no certainty
but only to probabilities of a lower order than those
which speak for the truth of evolution. For the past
twenty years the futility of the old theories and discus-
sions has been generally recognized, and the desire for
more exact knowledge has been keenly felt. Consequently

1 Introductory address in the annual discussion before the American
Society’ of Naturalists, Princeton, N. J., December 28, 1911.

121

122 s THE AMERICAN NATURALIST [Vor. XLVI

analytical and experimental work on the problem of the
factors of evolution was begun some twenty years ago
and has been continued with ever-increasing interest to
the present time.

In*the first enthusiasm over the experimental method
of attack it seemed to many students of this problem that
at last a path had been found which would lead straight
to the goal, that the causes of all evolution were about to
be revealed, and that the practical control of evolution,
with all that this implies, was almost within reach. About
that time a young physiologist said to the Director of
the Zoological Station at Naples, ‘‘Why do you spend so
much money publishing these beautiful monographs on
the Fauna and Flora of the Gulf of Naples?’’ Dr. Dohrn
replied, ‘‘ You are the first person who has ever asked me
such a question; many have asked how and where I got
the money, but no one has asked why. What do you
mean?” ‘‘Only this,’’ said the physiologist, ‘‘that
within twenty-five years we shall be making experiment-
ally an indefinite number of faunas and floras, and the
present one will then be only one of many.’’ In the
opinion of many investigators at that time, experimental
evolution was soon to give us a new world of living
things, and it was about to reveal conclusively the causes
of evolution. We have now had one or two decades of
this experimental evolution, and it may be worth while
to inquire, What has been the net result? If the answer
should seem to be somewhat discouraging I would beg to
remind you that is so chiefly because the problems have
been found to be much greater than was at first supposed.
The experimental method as applied to the evolution
problem has justified itself; it has set the problem in a
clear light and it has brought forth facts of the greatest
significance, but it has not enabled man to do in twenty-
five years what it took nature twenty-five million years
to do.
I

The most significant work on geneties since the time
of Darwin is that which is identified with the name of

No. 543] PROBLEMS OF~EVOLUTION ` 123

Mendel. According to the doctrine of Mendel and his
followers each organism is composed of a multitude of
unit characters, which do not blend nor lose their identity
when mixed with others as a result of sexual reproduc-
tion, but which may be expected to come out in the end,
practically as they went in at the beginning. This con-
clusion has modified in a striking manner the entire con-
ception of evolution and heredity. We no longer discuss
the origin of species, but rather the origin of characters;
we no longer rely upon chance to bring out certain hered-
itary characters, but are enabled at will to make many
analyses and syntheses of these characters. These dis-
coveries probably mark the greatest advance ever made
in the study of heredity; they have made it probable that
evolution proceeds by the evolution of individual charac-
ters; but have they shed any light on the method and
manner of this evolution? Permutations of Mendelian
characters we may have without number, of new combina-
tions of these there may be no end, but, so far as known,
no new characters are formed by such temporary com-
binations, there is no ‘‘creative synthesis,’’ no lasting
change. Evolution depends upon the appearance of new
characters; the discoveries of Mendel show us how to fol-
low old characters through many combinations and
through many generations, but they do not show us how
new characters arise. These discoveries have given us
an invaluable method of sorting and combining hered-
itary qualities, but Mendelian inheritance, as commonly
expounded, does not furnish the materials for evolution.
Many modifications of Mendelian inheritance have been
described, many alterations of dominance, or blending of
characters, the causes of which are not yet well under-
Stood. Perhaps in these ‘‘unexplored remainders’’ may
be found the causes of new characters. It is not yet cer-
tain that the unit characters, or rather their determiners
in the germ, are beyond the reach of environmental in-
fluence ; it is not certain that in their mixture with others
they never combine or influence each other in such man-
ner as to form new unit characters. Indeed, it is difficult

124 THE AMERICAN NATURALIST [ Vou. XLVI.

to understand how new characters could ever appear
except under one or the other of these conditions. We
particularly need at this time more knowledge of the
mechanism underlying the gross phenomena of Men-
delian inheritance, and then perhaps we may learn under
what conditions this mechanism may be altered.

As a result of the work of Mendel and his followers we
know much more about heredity than was known before,
we have learned how to separate and to combine hered-
itary characters, we have learned to look for evolution in
the appearance of new characters, but we have not learned
how to produce new characters.

Il

Practically all who have ever thought or written on
evolution have found the principal causes of the trans-
mutation of old characters into new ones in the action of
extrinsic, or environmental, forces on the organism. As
the result of years of labor on this subject Darwin con-
cluded that ‘‘variability of every sort is due to changed
conditions of life.” It is well known that environmental
changes produce many kinds of modifications in organ-
isms, and in general these modifications are the more pro-
found the earlier they occur in ontogeny; it is known that
slight alterations of the germ cells may produce great
modifications of adult structure, and it seems reasonable
to suppose that environmental changes of the right sort
applied to the germ cells at the right stage would lead to
a permanent modification of the substance of heredity
and hence to the appearance of new characters of evolu-
tionary value. And yet one of the most striking results
of recent work is to show the small effect of environ-
mental changes of all sorts on racial characters. Marked
individual modifications may be produced which do not
become racial. Usually not one of thousands of varia-
tions which oceur have any evolutionary value. These
variations come with changing environment and with
changing environment they disappear. Just as in the
individual, so also in the race there seems to be a power

No. 543] PROBLEMS OF EVOLUTION 125

of regulation which causes a return to the type, when
once this has been departed from.

In several instances recent investigators have found,
or have thought they have found, experimental evidence of
the inheritance of characters acquired through environ-
mental changes. But these evidences are by no means
conclusive. In a few cases it is known that the effects
of changed environment last through two or three genera-
tions and then disappear. In such cases racial, or speci-
fic, regulation is slow; in most cases this regulation takes
place in the first generation after the environmental
change disappears. Perhaps in this lingering effect of
a changed environment we have the first indication of the
appearance and fixation of a new character. Here, un-
doubtedly, much work of value remains to be done.

Very rarely a sudden variation, or mutation, arises
which is perpetuated by heredity and which forms the
basis of a new race. In most cases which have been care-
fully studied such mutations consist in the dropping out
of some old character rather than in the addition of a new
one, but at least they represent modifications of the hered-
itary characters, and as such they furnish material for
evolution. Whence and how they appear we do not know,
for like the kingdom of heaven, they come without obser-
vation. Their infrequency, amidst the multitude of
somatic variations, indicates the wonderful stability of
racial types and teaches respect for Weismann’s doċtrine
of a germplasm, relatively stable, independent and con-
tinuous. :

This distinction between somatic and germinal varia-
tions, between those which concern only the individual
and those which are inherited and furnish material for
evolution, marks the greatest advance in the study of
evolution since the work of Darwin. And just as these
germinal variations are the only ones of importance An
the process of evolution, so the question of their origin 1s
the greatest evolutionary problem of the present day.
How are such germinal variations produced? Do they
occur as the result of extrinsic or of intrinsic causes? By

126 THE AMERICAN NATURALIST [ Vou. XLVI

instinct we are all Lamarckians and are inclined to fol-
low Darwin in ascribing variability of every sort, germi-
nal as well as somatic, to changed conditions of life. But
this is by no means a necessary conclusion. It is con-
ceivable that germinal variations result from combina-
tions of different germplasms, as Weismann supposed,
that the determiners of Mendelian characters do not
always preserve their individuality, but sometimes unite
in such way as to modify the unit characters; but as yet
we have no evidence that new characters are formed in
this way, and the study of Mendelian inheritance has
made this possibility less probable than it once was.
Again it is possible that germinal variations, and new
hereditary characters, may result from intrinsic changes
in the germplasm, comparable to the spontaneous changes
which occur in radium, for instance; such a view of
transmutation through intrinsic, spontaneous, changes
has points of resemblance to the doctrine of orthogenesis,
but of its truth or falsity we have no sufficient evidence.

If changes in the germplasm may be induced by ex-
trinsic conditions, then a real experimental evolution will
be possible; if they can not be so induced we can only look
on while the evolutionary processes proceed, selecting
here and there a product which nature gives us, but
unable to initiate or control these processes.

IMI

Darwin’s theory that selection is the most important
factor in preserving and building up evolutionary char-
acters remains to this day a theory. The brilliant re-
searches of our distinguished guest, Professor Johann-
sen, and of our President, Professor Jennings, have
shown that the selection of fluctuations, or somatic varia-
tions, have no permanent effect in modifying a race; but
selection or elimination of germinal variations may be
an important factor in evolution, though it has little or
nothing to do with the formation of new characters, and
serves merely as a sieve, as De Vries has expressed it, to
sort the characters which are supplied to it.

No. 543] PROBLEMS OF EVOLUTION 127

On the other hand, selection of favored races and elimi-
nation of the unfit is still the only natural explanation of
fitness, of adaptation, in organisms. As a species-form-
ing factor selection is probably of less importance than
Darwin supposed; as a possible explanation of the won-
derful adaptations which allliving things exhibit it seems
to be all important; but extensive experimental investi-
gations of the causes of adaptation are greatly needed.

IV

The microscopic study of the germ cells during the past
twenty-five years—their growth, maturation, union in
fertilization, and their subsequent development—has fur-
nished material of the greatest importance for the com-
prehension of the mechanism of heredity and evolution,
and yet almost everything in this field remains to be done.
The parts played by the different constituents of the cell
in assimilation, regulation and heredity are still in doubt,
and in spite of many alluring hypotheses we know prac-
tically nothing about the way in which hereditary char-
acteristics arise from the germ. The study of the cellu-
lar basis of heredity has to a large extent been guided
and influenced by our knowledge of the gross phenomena
of heredity, and this must always be the case; but the
brilliant discoveries of the last few years as to the cellu-
lar basis of sex show the great assistance which the study
of cytology may render to the science of genetics. Many
interesting experiments have been made upon the germ
cells in the attempt to shift dominance, to modify inher-
itance, to create new characters; in a few instances it has
been shown that certain modifications of the embryo or
adult organism follow certain modifications of the germ,
but in no instance has it been shown that such modifica-
tions are inherited and are consequently of evolutionary
value.

Not merely the constitution of the germ and the ways
in which this may be modified, but also the precise man-
ner in which the structures of the germ become trans-

128 THE AMERICAN NATURALIST [ Vou. XLVI

muted into the structures of the adult are evolutionary
problems of the greatest importance. It is an amazing
fact that the great problems of ontogeny—viz. the under-
lying causes and mechanism of differentiation—are
to-day, after more than a century of scientific observa-
tion and experiment, almost as complete a mystery as
ever. If we are as yet unable to determine the precise
manner in which the structure of the germ evolves into
the structure of the adult in the common, ever-present
phenomena of reproduction, it is small wonder that we
have been unable to determine in detail the way in which
one race is transmuted into another.

In conclusion I think it must be admitted that the ex-
perimental study of genetics has been a little disappoint-
ing. We had supposed that organisms would be more
tractable, more willing to evolve, than we find them. The
older view that organisms were plastic and could be
moulded ‘‘while you wait’’ now reminds one of the view
of certain childless theorists, that children are plastic
clay in the hands of parents or teachers; both of these
views neglect the fact that the living organism, delicate
and responsive beyond compare, is still wonderfully
strong, stable and stubborn. So far as the factors of evo-
lution are concerned experimental study has thus far
been a weeding-out process, and at times it seems that
nothing will be left.

The problems of evolution are as much problems to-
day as they ever were, and though some of these prob-
lems may soon be solved, we may rest assured that there
will always be the evolution problem. The path which
we thought led straight to the goal has had to be retraced
with much labor; the hilltop from which we confidently
expected to see the spires of Eldorado has only served to
show us how great are the difficulties before us. But this
is the order of nature, the common experience of all
search for truth, and we would not have it otherwise.
‘‘For to travel hopefully is a better thing than to arrive,
and the true success is to labor.”

LIGHT THROWN BY THE EXPERIMENTAL
STUDY OF HEREDITY UPON THE FAC-
TORS AND METHODS OF EVOLUTION’

DR. C. B. DAVENPORT

COLD Spring HARBOR, N. Y.

THe most important contribution of modern studies in
heredity to the topic of evolution has been a new formu-
lation of the problem. Until a decade ago the problem
of organic evolution was regarded as synonymous with
that of the origin of species. We were, however, not
agreed as to the definition of species; on the contrary,
we realized that our notion of the term was exceedingly
hazy. And, doubtless, the reason why we made so little
progress in getting at the methods of evolution was be-
cause of this bad formulation.

To-day all this is changed. We think less of the origin
of species and more of characteristics; of their nature,
their origin and their distribution. The concrete ques-
tion of the origin of a given species has become broken
up into the questions of the origin of its differential
characters. Thus, the problem of the origin of man has
been broken up into the problems of loss of the tail, of
the hairy coat, of skin pigmentation, of melanic iris pig-
mentation, the acquisition of a more complicated brain
structure, the reduction of the lower part of the face,
the acquisition of the ability to learn to count and
talk, to wear clothes, to be honest, truthful, regardful of
the property rights of others, and to exercise self-control
in the sex sphere. So long as we formulated our problem
as the explanation of the ‘‘origin of the human species,”
as though the human species were an indivisible unit, so
of Natu-

Read (with slight alterations) before the American Society
ralists at Princeton, December 28, 1911.
129

130 THE AMERICAN NATURALIST [ Vou. XLVI

long we floundered helplessly in the quicksand of un-
clearness and complexity; now that we recognize our
study to be the history of any inheritable trait we move
on surer ground. And so, in general, progress is to be
made in the future by careful attention to the evolution
of characters.

The second change in the formulation of the problem
that is due to the modern study of heredity is that we no
longer consider even the character as the ultimate unit
of evolution, but regard it rather as a product of such a
unit. For the character is, in some way or another, de-
termined by the conditions in or the constitution of the
germ-plasm and these conditions and this constitution,
though very different from the adult characters, are
their germinal representatives; and these germinal rep-
resentatives are the real units to be studied. Thus we
do not to-day formulate our problem as the evolution of
man, or of a blue-eyed man, but ask how the determiner
for brown-eye became lost from the germ-plasm. So, in
general, the problem of evolution is formulated as that
of the history of the germinal determiners of characters.

Besides formulating more precisely the problem of
evolution, modern studies have discovered certain
methods of evolution which were not appreciated a
decade ago. First, we have come to realize that, though
not uniformly, yet to a surprising degree, characteristics
are independent of one another and, hence, that their
determiners in the germ-plasm are commonly not bound
together. The evidence for this is found in the breeding
experiments that have been performed on scores of
species of animals and plants, both feral and domesti-
cated. Breeders have taken advantage of this independ-
ence to create almost any desired combination of known
characters. Thus, in poultry the single pea or rose
comb may be combined with a black, white or game
plumage, with or without unfeathered shanks. The
shepherd’s purse that grows by the roadside may be
made with either of two forms of capsules combined with

No. 543] FACTORS AND METHODS OF EVOLUTION 131

four forms of leaf in the rosette stage; fruit-flies (Droso-
phila) may have any one of several colors of eye com-
bined with either short or long wings, and so on. The
characteristics that are associated in an individual are,
for the most part, not necessarily associated. The group
of characteristics that distinguishes individuals of one
‘‘ species”? from those of another is largely an accidental
one; and it is, therefore, not surprising that we so often
find individuals which in one, two or several characters
differ from the conventional description of their species,
and these have in the past caused great difficulty to the
species maker.

The fact that most characteristics are not necessarily
associated—that they may occur in various combinations
—certainly accounts for the multiplicity of ‘‘varieties’’
in domesticated species; and for much of the variation in
feral species. Moreover, it probably accounts for the
presence of many ‘‘species’’ in a genus. I may repeat
here what I wrote in 1909.

Dr. Ezra Brainerd has shown how many wild “species” of Viola
have arisen by hybridization, as may be proved by extracting from them
combinations of characters that are found in the species that are
undoubtedly ancestral to them. In such highly variable animals as
Helix nemoralis and Helix hortensis it is very probable that individuals
with dissimilar characters regularly mate in nature and transmit diverse
combinations of characters to their progeny. Indeed, if one examines a
table of species of a genus or of varieties of a species one is struck by
the paucity of distinctive characters. The way in which species, as
found in nature, are made up of different combinations of the same
characters is illustrated by the following example, taken almost at
random. Among the earwigs is the genus Opisthocosmia, of whieh the
five species known from Sumatra alone may be considered. They differ,
among other qualities, chiefly in the following characters (Bormans and
Kraus, 1900) :

Size: A, large; a, small

Wing-seale: B, brown; b, yellow.

Antennal joints: C, unlike in color; c, uniform.

Forceps at base: D, separated; d, not separated. —

Edge of forceps: E, toothed; e, not toothed. 3
Fourth and fifth abdominal segments: F, granular; f, not granular.

132 THE AMERICAN NATURALIST [ Vou. XLVI

The combinations of these characters that are found are as follows:
Opisthocosmia ornata: AbcDEF.

insignis: ABcDEf.

longipes: AbCDEf.

tenella: AbCdef.

minuscula: aBCDEf.

Other species occur, in other countries, showing a different combina-
tion of characters, and there are characters not contained in this list,
which is purposely reduced to a simple form; but the same principles
apply generally.

The bearing upon evolution of the fact that species are varying com-
binations of relatively few characters is most important. Combined
with the fact of hybridization it indicates that the main problem of
evolution is that of the origin of specific characteristics. A character,
once arisen in an individual, may become a part of any species with
which that individual can hybridize. Given the successive origin of the
characters A, B, C, D, E, F, in various individuals capable of inter-
generating with the mass of the species, it is clear that such characters
would in time become similarly combined on many individuals; and the
similar individuals, taken together, would constitute a new species. The
adjustment of the species would be perfected by the elimination of such
combinations as were disadvantageous.

Second, modern studies have taught us that we have
regarded the steps of progress in evolution in too crude
a way. One school adhered to the view that characters,
as we know them in the adult, arose gradually in phy-
logeny as in ontogeny, i. e., that the germ-plasm under-
goes a development as the child does. Another school
proclaimed for discontinuity in phylogeny; i. e., that the
conditions in or the constitution of the germ-plasm un-
dergoes from time to time more or less abrupt changes.
Such abrupt changes are not altogether unknown in
ontogeny ; for the sundering of a chromosome or the per-
foration of a membrane involves essentially abrupt or
discontinuous processes. The new era of experimental
breeding is leading us to a position that is in some
respects intermediate between the views of these two
schools. We have discovered a hitherto unsuspected
multiplicity of inheritable units, indicating a vaster com-

» plexity of the system of determiners in the germ-plasm

No. 543] FACTORS AND METHODS OF EVOLUTION 133

than we had dreamed. Sometimes a prominent charac-
ter is represented by a single determiner like (perhaps)
roseness of the comb of the fowl; but in most cases there
is a multiplicity of factors, as in human hair and skin
pigments, in the yellow of mice, in shank feathering of
fowls and in seed coat-color of oats. In consequence of
the fact of this multiplicity of factors and of the fact
that a variable number may be present in different cases
the adult character appears in numerous grades of de-
velopment.

Indeed, the gradation of characters is, in these cases,
such that one has to recognize that discontinuous varia-
tion passes over into continuous variation, in the sense
that 40, 41, 42 form a continuous series, if not in the
sense that x, x + dx, «+ 2dz, etc., do. If a desire for
uniformity leads us to conclude that all variations in the
germ-plasm are discontinuous at least we see in many of
these variations sufficient justification for the continuity
hypothesis of the old-fashioned selectionist. The new
light that has been thrown on the subject is the certainty
of discontinuity in most cases and apparent continuity
only in the limiting case. The reason why the old con-
tinuity hypothesis was for so long a time accepted was
that we had underestimated the fineness and the multi-
plicity of the units of inheritance.

Third, experimental work has thrown a new light on
the process of selection. It is clear that Darwin confused
under this term two ideas that we now sharply separate ;
namely, the selection of the most favorable individuals
and the selection of the most favorable blood, race, strain
or pure line (biotype, Johannsen). In so far as not the
soma but the germ-plasm is the proper basis of selection
it is clear that the favorable biotype is what we should
seek for to make most rapid advance. By this means
Pearl has increased the fecundity of his poultry; thus,
probably Castle has extended step by step the color
pattern of rats; this poultry fanciers have improved the
color pattern of Barred Plymouth Rocks; thus I have
gained a syndactyl race of fowl.

134 THE AMERICAN NATURALIST [ Vou. XLVI

The method of personal selection has been widely used
by the less philosophical breeders. I do not think it fair
to Darwin to designate it as the exclusively Darwinian
selection. Whether advance can be made by personal
selection seems to me still an open question. Granting
our inability to reason about genotypical constitution
from the phenotypical, still, other things being equal,
and in the long run and with great numbers of individuals
an extremely high variant is more apt to belong to a
genotype with a high mean than to one with a low or
intermediate mean. Thus a breeder who selects merely
the very best somas of a large number will be apt to
select any superior biotype that may occur in his mate-
rial. This is doubtless the reason why breeders who con-
sider only the somas of their breeding stock nevertheless
sometimes make progress; for they are occasionally for-
tunate enough to stumble upon a new biotype.

Fourth, the results of experiments have thrown light
on the long-discussed question of the discontinuity be-
tween species, of the swamping effects of intercrossing
new varieties with the parent species and the necessity
for isolation to permit new varieties to become established
as distinct species. We now realize that the danger of
swamping which formerly seemed so logically necessary
is, from our new point of view, not really to be seriously
considered. Characters are rarely, if ever, swamped.
Apparent swamping by intercrossing occurs when the |
new character depends on many determiners. But it is
not, even in this case, really swamped; for no true blend
occurs but, on the contrary, a segregation of the original
extreme conditions takes place. This is well illustrated
by the case of human skin color. When the germ cell
that carries white skin color unites with the germ cell
that carries black skin color the ‘‘white’’ character seems
swamped in the offspring; but the swamping is only
apparent. Two mulatto parents have children of various
tints and, occasionally, one with a cleat white skin, as well
as one with a black skin like the original negro ancestor.

No. 543] FACTORS AND METHODS OF EVOLUTION 186

Neither white nor black is truly swamped. The extreme
white or black conditions are rather rare, as is to be ex-
pected where a multiplicity of factors is involved. Thus,
if there were two (2) factors P’, P” involved in the negro
skin then in F, one in sixteen should be negro-black, one
in sixteen pure white, and half of the remainder should
be light mulatto and half dark mulatto. Although studies
on this subject are not sufficient to warrant exact quanti-
tative conclusions, it is certain that more than two and
probably more than three factors are involved in the
pigmentation of the negro. If a number of mulattoes
inhabited (as sole occupants) an oceanic island, and bred
there, in the course of generations both the white and the
black types of skin color would be found again—the two
extreme types are not swamped. Consequently from our
present point of view, isolation is much less essential
than was formerly thought to be the case. Practically
important as it may be to keep races pure and ensure the
absence of intergrades or hybrids, it is not essential to
the survival of new traits that have arisen in the midst
of the old stock.

Fifth, the experimental study of heredity has thrown
light upon the question of the origin of new determiners.
Every critical experiment that has been tried demon-
strates again that the somatic condition exercises little
or no influence upon the determiners in the germ-plasm.
The first crucial experiment on this subject of which I
know was that of Francis Galton, who infused into tt gil-
ver-gray” does the blood of either yellow, black and
white, or common agouti rabbits. In one case an angora
buck and yellow doe had their carotid arteries So con
nected that for over half an hour the blood of each
flowed into the body of the other; so that about one half
of the blood of each was alienized. Yet, when the rab-
bits that had been operated upon were used as parents,
the offspring indicated that their germ-cells had under-
gone no modification ìn consequence of the foreign blood.
Recently the question has been revived by Guthrie, who

136 THE AMERICAN NATURALIST [ Vou. XLVI

made the experiment of engrafting foreign ovaries into
foster mothers very unlike the original females whence
they were taken. He concluded that the offspring were
modified in such a way as to prove that the transplanted
germ-plasm had received something from the foster
mother. Unfortunately Guthrie erred here, as my repeti-
tion of his experiments showed. For unquestionably, the
hens that were operated upon regenerated their proper-
ovaries and produced no eggs from the engrafted ovaries.

Dr. Phillips, working with Castle, engrafted black-bear-
ing eggs from one female guinea pig into albino guinea
pigs and then mated the females that had been operated
upon with an albino male. All offspring were entirely
black, proving, first, that the engrafted ovaries were
functional and, second, that the determiners of the en-
grafted germ-plasm were not modified by the soma of the
albino mother. On the other hand, the experiments of
Standfuss, Tower and Kammerer on animals and Mac-
Dougal on plants apparently indicate that under the
influence of various conditions of moisture, temperature
and chemical action the germ-plasm may be changed.
These results, probable as they are, await confirmation.
If fully confirmed they will afford a picture of one way
in which new determiners may originate.

Finally, some light has been thrown by modern experi-
mental studies on the subject of adaptation—for Darwin
the corner stone of organic evolution. But here, it must
be confessed, the contribution has not been great. That
there is such a thing as selective elimination is plainer
than ever. That some characteristics are compatible
with the environment and some incompatible is incon-
testibly true. Two cases in poultry illustrate this. I
have a lot of rumpless fowl; the cocks are sexually active
and the hens lay numerous eggs; but every egg is sterile,
for the reason that the erection of the tail feathers in the
hen is essential to the clean exposure of the cloacal open-
ing for the transfer of the sperm. Hence, since in the
rumpless hens the cloacal opening is not accessible to the

No. 543] FACTORS AND METHODS OF EVOLUTION 137

sperm, such a sport must, in nature, be eliminated.
Under domestication it is continued by trimming away
the feathers that cover the vent. Similarly, winglessness
in male fowl renders copulation difficult because the
wings serve the cock as balancers while treading the hen.
These then are examples of characteristics that must be
eliminated in nature. In the case of certain striking
colors in poultry there is evidence that they are selected
against; their possession gives their owner a handicap.

On the other hand certain new characteristics of fowl
may be preserved because apparently they offer no handi-
cap. Thus in the rumpless fowl the oil gland is absent
and the birds seem to be none the worse on that account;
their plumage’is bright and quite as resistant to a wet-
ting as that of birds with an oil gland at the base of the
tail. The striking fact that our experimental work yields
is the great number of new characters that seem to bear «
no relation to fitness or unfitness, but are truly neutral.
Thus I can not find that polydactylism, shank-feathering
or its absence, and the lower grades of single, pea and
rose comb have any adaptive significance for poultry.
One can invent adaptive explanations for them or their
absence in birds, but there is no reason for thinking that
the explanations are significant. On the other hand,
there is accumulating considerable experimental support
for Darwin’s theory of sexual selection; but of this it is
early to speak. On the whole, I think it may be fairly
said that experimental work supports the principle of
selective elimination but finds many characters that are
wholly neutral. i

To sum up, modern experimental study of heredity has
given-a new formulation to the problem of evolution and
has given definite data on the method of evolution. It
formulates the problem of evolution as the problem of the
nature and origin of the germinal determiners of char-
acters. It has shown that, for the most part, the new
determiners arise one at a time and are independent ef
one another, may occur in any combination and may be

138 THE AMERICAN NATURALIST [Vou. XLVI

transferred from one strain or species to another. It
has been shown that the unit characters are much more
numerous and finer things than we had thought and,
therefore, that the steps of evolution are frequently very
small ones and are taking place in many directions. It
has shown the relative unimportance of the isolation fac-
tor, since true blends of characters rarely, if ever, occur.
It has demonstrated the lack of influence by soma upon
germ-plasm; but has rendered it probable that external
conditions may directly modify the determiners of the
germ-plasm. It brings support for the view of selective
elimination of undesirable traits but finds that many, if
not most, characters that arise are neutral in respect to
any adaptive significance. Finally, it looks forward with
a justifiable expectancy to the completer experimental
test of the factors of evolution and their eventual com-
plete elucidation.
LITERATURE CITED
MEE A. de, and EE H. 1900. Forficulidæ and Hemimeridæ. Das
eich, 11 Lief. Berlin.

dastia W. E., and Pe J. ©. 1911. On Germinal Transplantation in

Vertebrates. Carnegie Institution of Washington, Publication No. 144.
Davenport, ©. B. 1909. Inheritance of Characteristics in Domestic: Fowl.

Carnegie Institution of Washington, Publication No.

Davenport, C. 1911. The Transplantation of Cvaries in Chickens.
Jeurnal of iirohotoas, 22: 111-122.

BIOTYPES AND PHYLOGENY

Dr. HUBERT LYMAN CLARK
MUSEUM or Comparative ZOOLOGY, CAMBRIDGE, Mass.

[Tue substance of this paper was presented to the
American Society of Naturalists at the Princeton meet-
ing under the title ‘‘Pure Lines and Phylogeny.” Dr.
Johannsen entered an emphatic protest against the use
of ‘‘pure line’’ in the sense of a group of individuals
characterized by an identical combination of the same
determinants. Subsequent conversation with Dr. Jo-
hannsen, and the recent clear exposition by Shull (Sci-
ence, Jan. 5, 1912, pp. 27-29) satisfied me that what I had
considered ‘‘pure lines’’ (such as those distinguished by
Jennings in Paramecium) are the pure strains called
biotypes by Johannsen. I have modified my paper ac-
cordingly and have avoided using the term ‘‘pure line.’’
I have also abandoned the very convenient term ‘‘pheno-
type’’ because my use of it as a contrast to biotype is
not strictly in accord with Johannsen’s usage of it as a
contrast to ‘‘genotype.’’ At Princeton, I protested
against Johannsen’s use of the word genotype, because
the word is preémpted for a totally different usage.
_I suggested a substitute, but this failed to meet
with Dr. Johannsen’s approval. Since I have seen
Shull’s definition of ‘genotype’? (to which Dr. Johann-
sen himself referred me), I think the objection to the
word is greater than before because ‘‘type’’ implies a
single definite thing or model and Johannsen’s ‘*gen-
otype’’ is not that but is ‘‘the fundamental hereditary

. combination of genes of an organism.” In other
words it is not a concrete thing but the intangible char-
acter of that thing. It seems to me the termination
“‘plast?? (mAaords, moulded, formed, i. e., formed from

139

140 THE AMERICAN NATURALIST [Vor. XLVI

the genes) expresses the idea better than “type”? (TÚTOS,
a figure, impression, model) and ‘‘oenoplast’’ is quite
as euphonious as ‘‘genotype.’’ The adjective form is
equally satisfactory, while the use of this term will not
require the abandonment of ‘‘gene.’’ In the following
pages therefore I have used ‘‘genoplast’’ and ‘‘geno-
plastic’’ in place of genotype and genotypical and I do
not believe any misunderstanding will be possible. I
have no desire to insist on these words, however. The
whole matter is a very trivial one and I would very much
prefer that Dr. Johannsen should himself choose a substi-
tute for ‘‘genotype.’’ I can not, however, agree with him
that genetics and systematic SERE are so far apart
that no confusion can result from using identical terms
in totally different senses. I believe that so far as pos-
sible workers in any branch of biology ought to keep in
touch with as much of the whole field as may be possible,
and that we should all endeavor to avoid ambiguity and
unintelligibility in the use of such technical terms as are
necessary.—H. L. C.]

Systematic zoology and botany deal primarily with
species and varieties, and can not therefore be expected
to throw light upon the existence of genoplastic groups.
Indeed, only those systematists who deal with organisms
which reproduce asexually or parthenogenetically are
likely to have any personal contact with them or even to
meet with direct evidence for or against their occurrence.
Since, however, the existence of such pure strains (bio-
types) seems to have been definitely proved! the question
of their relationship to the phylogenetic problems with
which the systematist has to deal becomes one of some
interest.

‘The problems of phylogeny are those of complicated
polygenoplastic groups—so complicated indeed that the
most complex of chemical compounds is simple in com-
parison. The study of these problems makes for caution

* JENNINGS, H. 8., 1911, AMERICAN NATURALIST, Vol. 45, pp. 79-89.

No. 543] BIOTYPES AND PHYLOGENY ' 141

in affirming that any one theory or hypothesis contains
all the truth. Thus we are coming to realize that neither
the Darwinian nor the De Vriesian theory of the nature
of the material upon which selection works is altogether
complete in itself and that neither when properly under-
stood wholly debars the other. If we accept the current
Mendelian and genoplast theories of heredity, must we
not admit that all variation is fundamentally discontin-
uous and that what has been called continuity is not
really such? It may be convenient to use such terms as
‘‘continuity’’ and ‘‘discontinuity’’ but are they not sub-
jective ideas rather than objective realities of impor-
tance? So, too, is it necessary to claim that the geno-
plast theory of heredity contains all the truth and that
the transmission or ‘‘phenotype’’ theory is wholly false?
It is easy to see how in pure line breeding ‘‘ancestral in-
fluence’’ is, as Johannsen says, ‘‘a mystical expression
for a fiction” but in the complicated polygenoplastic
groups of the higher Metazoa it is hard to see why the
history of the formation of a gamete may not be of im-
portance. Is this not virtually admitted by Johannsen
when. he grants the existence of ‘‘perturbations by in-
fection or contamination’’? And if this be granted, why
is there any necessary antagonism between the genoplast
theory of heredity and the belief that ‘‘discrete particles
of the chromosomes’’ may be ‘‘bearers of special parts
of the whole inheritance ’’?

However this may be, none of us has any doubt that the
discovery of biotypes has been a real stimulus to experi-*
mental work, and there is no reason why it may not also
be a stimulus to the investigation of phylogenetic prob-
lems even though it does not assist greatly in their im-
mediate solution. Among the difficulties of the system-
atist perhaps none is better known than that which we
may call the problem of large genera—genera made up
of dozens, in some cases indeed of hundreds, of species,
many of which are poorly defined and more or less inter-
grading. Some of these genera, as Crategus, Unio and

142 THE AMERICAN NATURALIST [ Vou. XLVI

Salmo, have become notorious and are not infrequently
referred to as proof of the futility of systematic work.
Does the discovery of biotypes afford any help in explain-
ing the existence of such genera?

I think that it does, particularly when considered in
connection with the broadest interpretation of Mendel’s
law. If we compare one of these inclusive genera with
one which contains few and well-defined species, we see
that the essential difference lies in the latter having the
characters sharply defined, with little diversity and no
blending, while in the former the same or similar char-
acters show so much diversity and such a tendency to
blend that the resulting recombinations are most perplex-
ing. It has occurred to me that we have here a condi-
tion of affairs analogous to what we find in the develop-
ment of the individual. Certain individuals with unlike
parents show what seems to be a blending of the parental
characters, while in numerous other cases the characters
of the individual can be referred unhesitatingly to one or
the other parent. Thus, as the well-known investigations
of Castle have shown, if lop-eared rabbits are crossed
with rabbits having ordinary ears, the character of the
ears in the offspring can not be referred to one parent
rather than to the other, while if pigmented and albino
rabbits are crossed, the color-character of the offspring
in succeeding generations can be so referred without
difficulty. This difference has been interpreted by Dav-
enport and others as due to the potencies of the deter-
minants, the apparent blending being associated with
equipotency or an approach thereto, while the distinct
characters result from allelopotency. Now may it not
be that a similar inequality of potency occurs among
the biotypes which go to make up a species? And so
when reproduction takes place we find some species in
which well-defined characters are dominant and the re-
sulting individuals form easily recognized groups, while
in other species there is a lack of definiteness and a blend-

No. 543] BIOTYPES AND PHYLOGENY 143

ing of characters which make the resulting forms most
confusing.

Jennings has shown that there are inherent difficulties,
which have so far been prohibitive, in securing crosses
between biotypes of Paramecium under experimental
conditions, yet it is obvious that such crossing must
occur constantly in nature; otherwise the whole geno-
plast theory becomes reduced to an absurdity. Granting
then the natural crossing of biotypes, let us consider the
case of a species, which for simplicity’s sake we will
suppose is made up of three biotypes (1, 2 and 3), each
of which is distinguished by certain character-combina-
tions, designated a, b and c, respectively. If the union
of 1 and 2 is readily effected, while that of 1 and 3 or
that of 2 and 3 rarely occurs, it is evident that ab will
far more commonly characterize the species than ac or
be which will indeed seldom appear. The species will
therefore approach identity with one of its biotypes,
which may thus be considered the dominant strain. The
inequipotency of the biotypes and the resulting definite-
ness of character in the species are obvious. If, however,
the union of 1 and 3, and of 2 and 3 are as readily effected
as that of 1 and 2, ac and be will occur as frequently in
. the species as ab. In such a case the biotypes are equi-
potent and the resulting species may be correspondingly
ill-defined.

The hypothesis here suggested of the ‘‘inequipotency
of biotypes’’ may thus be the explanation of the existence
of the well-defined species so generally known, while the
occurrence of large heterogeneous assemblages of either
Species or varieties may be interpreted as due to an un-
usual equipotency. The experimental determination of
the existence of this hypothetical difference in the po-
tency of the biotypes within a species would well be worth
while, if it should ever prove to be possible. The study
of large heterogeneous groups may suggest some other
lines of investigation into the nature and even the origin
of biotypes. For example, such groups occur chiefly, if

144 THE AMERICAN NATURALIST [Vou. XLVI

not wholly, in the more specialized portion of any stock
and in some cases at least appear to be associated with
the fading-out or senescence of that particular branch.
This suggests the possibility that the potency of a biotype
ultimately alters, even though there is no visible or tan-
gible evidence of change.

A second problem which puzzles the systematist is the
variability in the value of a character for distinguishing
species, genera and even higher groups. Color is a fa-
miliar example of this. It is of real value among birds
and in numerous other cases, but is almost worthless
among many invertebrates. Does the knowledge of the
existence of biotypes help us to understand why this is?
At first thought one might say that here again the inequi-
potency of the biotypes was the explanation of the phe-
nomenon, but further consideration will show that this is
not the case, for of course the potency of a biotype will
involve all of its characteristic determinants and not
merely that or those associated with the character in
question. It is clear then that the value of any char-
acter for distinguishing species from each other—in
other words, its value for systematic work—depends on
the actual determinants in the genoplastic groups com-
posing those species. The variability in systematic value
shown by a given character is due then, not to the
potency, but to the composition of the biotypes involved.
Thus if all the biotypes contain identical color determi-
nants, then color will be an absolutely constant char-
acter in that species, but the greater the diversity in the
color determinants of the biotypes the more variable will
the color of the species be and the less useful the color
be as a distinguishing character. Conversely, we may
say that the value of color in systematic work will depend
on the degree of identity in color-determinants among
the biotypes composing the species concerned. If this is
so, the study of systematic characters and the measur-
ing of their diversity may suggest some characteristics of
biotypes as yet unsuspected. Thus biometrical work

No. 543] BIOTYPES AND PHYLOGENY 145

even in a polygenoplastic population receives an added
indorsement.

A third problem of the systematist (and for this occa-
sion the last) is found in the fact that diversity of mor-
phological characters in any given species is not hap-
hazard or indiscriminate, but is generally restricted to
such definite lines as to indicate more or less distinct
stages in the phylogenesis of that species. The belief that
diversity is significant and that its meaning may be dis-
covered has received extraordinary confirmation in Jack-
son’s just published, magnificent monograph on Echini?
in which the subject is very fully discussed. An illustra-
tion taken from his work will help to make clear the de-
sired point. In any regular sea-urchin, such as Arbacia
or Strongylocentrotus, a group of ten plates surrounds
the periproct, five of which are radial in position and are
called oculars while the other five are interradial and are
called genitals. Now in some echini all of these ten plates
are in contact with the periproct and thus form a simple
continuous ring but in most of the Recent species, the
oculars are much smaller than the genitals and some or
all of them are separated from the periproct by the meet-
ing of adjoining genitals. In other words, some of the
oculars may be excluded from the periproct and such are
said to be exsert, while those which separate adjoining
genitals and reach the periproct are called insert. Now
Jackson has demonstrated conclusively, contrary to the
widely held belief that the insertness of oculars is a mat-
ter of age and size, that for each species of sea-urchin
there is a characteristic arrangement of the genito-ocular
ring and that this arrangement is oftentimes a very con-
Stant character. Thus in 2,100 Arbacias from Woods
Hole, 87 per cent. have all the oculars exsert and in more
than 20,000 Strongylocentroti from Maine 95 per cent.
have the two posterior oculars insert. |

Having demonstrated the constancy of this character,
Jackson has gone on to an analysis of the variations from

* Jackson, R. T., 1912. Mem. Boston Soc. Nat. Hist., Vol. VII.

146 THE AMERICAN NATURALIST [ Vou. XLVI

the normal arrangement, occurring in large series of
adult specimens. And he has clearly shown that these
variations are nearly always significant. There are 32
possible arrangements of the plates of the genito-ocular
ring and there is no mechanical or structural reason why
any one of them should not occur. If variation were
perfectly haphazard every one would occur and there is
no obvious reason why they might not occur, with equal
frequency. Yet in fifty thousand specimens examined by
Jackson, representing 137 different species of Mesozoic.
and Recent Echini, ten of these possible arrangements
never occurred, and of the remaining 22 fourteen are so
rare that altogether they aggregated less than 14 per
cent. of the specimens. As a very large proportion of
these were individuals abnormal in some other particular,
it is fair to say that of 32 possible arrangements of the
genital and ocular plates only eight (or at most ten) occur
normally. Even more striking are the following facts:

When only a single ocular plate is insert, it is one of the
posterior pair; this is the case in 994 per cent. of the
specimens having one ocular insert.

When two oculars are insert, they are the posterior
pair in more than 99 per cent. of the cases and in every
case one of them belongs to that pair.

When three oculars are insert, they are the two poste-
rior and usually the left, but sometimes the right ante-
rior; this is demonstrated by almost 99 per cent. of the
cases.

When four oculars are insert, the one exsert is invar-
iably either the mid-anterior or right anterior.

These figures show how surprisingly definite variation
is in a character which, so far as we can see, might vary
with equal ease in any one of 32 ways. Yet it is only
when we examine a particular case that the significance
of this definiteness appears.

Jackson’s work is full of such cases, but as most of
us are familiar with Strongylocentrotus, we will consider
an illustration from that genus, which, in the old, broad

No. 543] BIOTYPES AND PHYLOGENY 147

sense, accepted by Jackson, includes more than twenty
species. Of these some have the ambulacra relatively
simple, the compound plates being made up of only four
or five elements each, while in the more specialized spe-
cies there may be as many as ten elements in each com-
pound plate. The various species can be arranged
roughly in a series beginning with the simplest and end-
ing with the most specialized? and Jackson shows that
the species with the simplest ambulacra (S. lividus)
has ‘‘no oculars insert’’ as the species character, with
‘‘right posterior ocular insert’? as a common variant,
while those with the most complex ambulacra (S. fran-
ciscanus and purpuratus) have two and often three
oculars insert. Now in our common Strongylocentro-
tus from Maine, while practically 95 per cent. have two
oculars insert, nearly 3 per cent. have only one insert, as
in the common variant of S. lividus, while about 2 per
cent. have three insert as in the usual variants of S. pur-
puratus. Jackson calls these arrested and progressive
variants, respectively, according to whether they resem-
ble a more simple or a more complex allied species.
Whether the terminology be accepted or not, the signifi-
cance of such facts can not be ignored. Are we any
better prepared, with our present knowledge of the ex-
istence of biotypes, to understand the reason for this
significance of variation?

If we compare a polygenoplastic group with a highly
complex chemical compound, an analogy is suggested
which warrants our answering this question affirma-
tively. In building up such a compound synthetically,
the specific properties of the constituents result in the
formation of certain definite compounds. These sub-
Stances are necessary for the further combinations
without which the ultimate compound could not be
formed. In other words, the formation of the desired
product is possible only because the chemical reactions

*There are some interesting exceptions, but as they do not affect the
subsequent argument, they need not be d here.

148 THE AMERICAN NATURALIST [ Vou. XLVI

will take place in an orderly sequence, in consequence
of the fixed specific properties of the elements involved.
Now any existing species of plant or animal is a
similar union of diverse elements and the possibilities of
its development would seem to be limited by the same
conditions which limit the formation of the chemical
compound, namely, the nature of the elements and
the orderly sequence of the reactions. (In either
case external conditions, the environment, would make
a profound difference, but for simplicity’s sake we may
omit reference to that influence.) As long as one be-
lieves that the elements composing a species are poten-
tially variable in all directions, it is evident that only the
pressure of external conditions can prevent an indefinite
and unmeaning variety in the product. Such a belief
results in making natural selection through the environ-
ment the supreme directive agent in evolutionary prog-
ress, and really puts more responsibility upon that im-
portant factor than it can reasonably be expected to
bear. But as soon as it is shown that the elements in-
volved are persistently unchanging to a remarkable de-
gree, it becomes clear that an orderly sequence in their
successive interactigns will follow just as in the forma-
tion of a chemical compound. Now biotypes are the
biological elements which enter into the formation of a
species, and the discovery of their existence and apparent
persistency makes the existence of an orderly sequence
in development quite comprehensible and indicates
clearly why diversity is so rarely haphazard. As in the
chemical synthesis, used as an illustration, the final re-
action follows necessary antecedent reactions, so in the
development of the species the last step necessarily de-
pends on the preceding, and the evolution of a group is
therefore bound to be strictly linear and in definite direc-
tions. Now just as in a chemical synthesis without add-
ing to or subtracting from its original constituent ele-
ments the process may be stopped, altered or accelerated,
either by addition or removal of some substance or by

No. 543] BIOTYPES AND PHYLOGENY 149

change in the external conditions, so the process of de-
velopment of a species or of any of its component indi-
viduals may be arrested, altered or accelerated by similar
means. Thus variants arise, individual or racial, some-
times slight, sometimes marked, but necessarily within the
limits laid down by the specific properties of the biotypes
involved. This may be well illustrated by one of Jack-
son’s discoveries about variation in the genito-ocular |
ring of the common tropical sea-urchin, T'ripneustes.
In specimens from Florida and the West Indies, 36 per
cent. have only the two posterior oculars insert, 38 per
cent. have three (the left anterior plus the posterior) and
18 per cent. have four (right and left anterior plus the
posterior). Evidently then individual variants both ar-
rested and progressive are common but within very re-
stricted limits. Further than this it appears that in
Bermuda a racial variant can be distinguished, for in
specimens from that locality 61 per cent. have only two
oculars insert, 35 per cent. have three and only 2 per cent.
have four. Now it matters not at all whether the Ber-
muda race is considered an arrested variant or the West
Indian form a progressive variant; the important fact
is the evidently marked but definitely limited racial diver-
sity. The study of such variants, however little light it
may throw on the immediate cause of their appearance,
is bound to help make clear the normal line of develop-
ment of the species to which they belong, and emphasizes
the definiteness and the significance of their diversity.
But if biotypes are really the fixed and unchanging ele-
ments which compose a species, the problem as to why
this diversity is so commonly definite and significant is
apparently simplified not a little by our knowledge of
their existence.

It is unnecessary to suggest any other phylogenetic
problems and the bearing of the study of genetics on
them, for if in these which I have suggested it has not
been shown that such study is helping us to understand
these problems better and is even indicating solutions,

150 THE AMERICAN NATURALIST [ Vou. XLVI

multiplication of such cases will not help the matter.
Personally, I believe that the experimental work now so
extensively carried on in the study of genetics is throw-
ing a flood of light on all biological questions and that
systematists not only may but must make use of the
demonstrated results of such study, in attacking their
own special problems, if they are really in earnest in the
_ purpose to solve them.

SHORTER ARTICLES AND DISCUSSION
A LITERARY NOTE ON MENDEL’S LAW

THE ever-increasing extension of the doctrine of Mendelism
brings it year by year to the attention of a widening circle
of general readers. Its application in the field of eugenics has
aroused a popular desire for a further knowledge of the now
famous principles of Mendel. Owing to the relative inacces-
sibility of Mendel’s original publication the exact terms in which
he formulated his conclusions have not been readily available.
To meet in some measure this lack of ready reference the fol-
lowing brief, synoptic statement of the fundamental principles
of Mendel is here presented in his own words, together with
certain collateral notes that may be of value to students of this
important law.

The general term ‘‘Mendel’s Law”? is usually applied to sev-
eral complex principles discovered by Gregor Mendel while
studying inheritance in certain plant hybrids. Various other
designations, however, appear in the literature, e. g., Mendel’s
“Law of Heredity,’ ‘‘Law of Inheritance,’ ‘‘Laws of Alter-
native Inheritance,” and ‘‘Law of Varietal Hybrids.’’* These
principles were enunciated by Mendel in a paper entitled, “‘ Ver-
suche über Pflanzenhybriden’’ (‘‘Researches on Plant Hy-
brids’’), which appeared in the Verhandlungen der naturfor-
schenden Vereines in Briinn, Vol. 4, 1865, Abhandlungen, pp.
3-47,5 but escaped the attention of biologists until the year
1900, when by De Vries, Correns” and von Tschermak,* they

*Castle, W. E., Proceedings American Academy - of Arts and Sciences,
Vol. 38, 1903, p. 535.

7 Biffin, R. n. Journal Agricultural Science, Vol. 1, 1905, p. 1.

x Weldon, W. F, R., Biometrika, Vol. 1, 1901, p. 228.

“De Vries, H., ‘‘ Species and Varieties,’’? Chicago, 1905, p. 716.

* Reprinted in Fora: Vol. 89, 1901, pp. 364-403.

*De Vries, H., ‘‘Das Spaltungsgesetz der Bastarde,’’ Berichte der
deutschen Sitaki Gesellschaft, Vol. 18, 1900, pp. 83-90.

"Correns, C., ‘‘@. Mendels Regel über das Verhalten der Nachkom-
menschaft der Kaaa, ? Berichte der deutschen botanischen Gesell-
schaft, Vol. 18, 1900, pp. 158-168.

151-

152 THE AMERICAN NATURALIST [ Vou. XLVI

were independently and almost simultaneously rediscovered.?

Mendel’s principles have been rephrased by later writers, and
they are now usually referred to as the Law of Dominance, the
Law of Segregation, and the Law of Recombination, respectively.

I. THe Law or DOMINANCE!

This so-called law is derived from the following principle of
Mendel (‘‘ Versuche,’’ etc., pp. 10-11

In der weiteren Besprechung werden jene Merkmale, welche ganz
oder fast unverändert in die Hybride-Verbindung übergehen, somit
selbst die Hybriden-Merkmale repriisentiren, als dominierende, und
jene, welche in der Verbindung latent werden, als recessive bezeichnet.

Translated this principle reads:

Those characters which pass entirely or almost entirely unchanged
into the hybrid combination and consequently in themselves represent
the characters of the hybrid, are designated as dominant, and those
which become latent in the combination are termed recessive.

Since dominance is rarely absolute this principle is not general
and should not be termed a law; indeed Mendel did not claim
it as a law. Recent statements of the ‘‘Law of Dominance’’
may be thus summarized:

When the two parents differ in respect of two contrasted characters,
only one, the dominant character, will appear in the hybrid. Domi-
nance, however, is seldom perfect, so that the dominant character in a
hybrid seldom reaches as full expression as in the dominant parent.

II. THE Law or SEGREGATION.

Mendel’s second principle (‘‘Versuche,’’ ete., p. 17) is thus
stated :

* Von Tschermak, E., ‘‘ Ueber künstliche Kreuzung bei Pisum sativum,’’
Zeitschrift fiir z landutetashutdions Versuchswesen in Oesterreich, Vol.
3, 1900, pp. 465-

*De Vries’s Bas was received for publication on March 14, 1900, and
that of Correns on April 24, 1900. Tschermak, in a postseript to his com-
T, says: ‘‘Die giciehucitios Entdeckung Mendels durch Correns,
de Vries (Berichte der deutschen piange Gesellschaft, 1900) und mich
sila mir besonders erfreulich,’

w<‘ Man hat dieses die pare aa! genannt,’’ Correns, C., ‘‘ Uber
Vererbungsgesetze,’’ Verhandlungen der Gesellschaft deutscher Naturforscher
und Arzte, 77 Versammlung, 1905, Part I, Leipzig 1906, p. 207

No. 543] SHORTER ARTICLES AND DISCUSSION 153

Da die Glieder der ersten Generation unmittelbar aus den Samen der
_ Hybriden hervorgehen, wird es nun ersichtlich, dass die Hybriden je
zweier differirender Merkmale Samen bilden, von denen die eine Hälfte
wieder die Hybridform entwickelt, während die andere Pflanzen gibt,
welche constant bleiben, und zu gleichen Theilen den dominirenden und
recessiven Character erhalten.

This may be translated as follows:

Since the members of the first generation” arise directly from the
seeds of the hybrids, it is now evident that hybrids, for each pair of
differentiating characters, form seeds, one half of which develops again
the hybrid form, while the other half produces plants which remain
constant and in equal proportions receive the dominant and recessive
characters.

Various terms have been applied to this law by different
authors, e. g., ‘‘Law of Disjunction,’’!* ‘‘Law of Purity of Germ
Cells,” “Law of Separation of Characters in Crosses,’’* and
“Law of Dichotomy.’ Generalized, the law may be stated in-
the following form:

In self-fertilized species an individual which is a hybrid with refer-
ence to a particular pair of characters tends to produce progeny one
fourth of which is of pure race like one of the parents of the hybrid,

another fourth of pure race like the other parent, while the remaining
half is hybrid, like the original hybrid itself,”™ that is, “ from the inbred
heterozygote come dominants and recessives in the proportion of 3:1,
and only one dominant in three is pure, the other two being hetero-
zygote.” ee

The above statement is purely objective ;** it states the results

* This is He F, generation of current literature.

“De Vries, H., Comptes rendus de l’académie des sciences, Paris, Vol.
130, 1900, ny 845-847,

* Castle, W. E., Proceedings American Academy of Arts and Sciences,

Vol. 38, 1903, p. 5 537.

“De Vries, H., Journal Royal Horticultural Society, Vol. 25, 1901,
p. 243.
“ Davenport, C. B., Biological Bulletin, Vol. 2, 1901, p. 307.

* Spillman, W. J., " $6 Application of Some of ‘te Principles of Heredity
to Plant Breeding,’ ? eya No. 165, Bureau of Plant Industry, U. 8.
Dept. Agriculture, 1909, p. -.

* Punnett, R. C., eee 1 Cambridge, 1907, p. 23.
"Por this paragraph. and the one immediately iy eee the writer is
indebted to Professor W. J. Spillman.

154 THE AMERICAN NATURALIST [ Vou. XLVI

of a cause, but gives no hint of that cause. It is not strange,
therefore, that the modern statement of this law should have
gravitated backward toward the fundamental cause underlying
the law. The more usual statement of it at the present time is
an inference from the facts observed, and may be stated as
follows:

When an individual is heterozygous for a given character it produces
two kinds of gametes with reference to that character, one like those
of one of its parents and the other like those of the other parent.

Mendel himself gives the corresponding statement of the law
of recombination ; that is, he states the inference about the kinds
of gametes a hybrid must produce, as inferred by the types of
the resulting progeny.

The principle of segregation, closely approximated long prior
to Mendel both by Goss? and by Sageret, was clearly enun-
ciated by Naudin,”* but these writers did not formulate their

*<<Tn the summer of 1820, I deprived some blossoms of the Prolific blue
of their stamens, and the next day applied the pollen of a dwarf Pea, and
of which impregnation I obtained three pods of seeds. In the following
spring, when these were opened, in order to sow the seed, I found, to my
surprise, that the colour of the Peas, instead of being a deep blue, like their
female parent was of a yellowish white, like the male. Towards the end
of the summer I was equally surprised to find that these white seeds had
produced some pods with all blue, some with all white, and many with both ~
blue and white Peas in the same pod.

‘*Last spring, I separated ‘al the blue Peas from the white, and sowed

each colour in separate rows; and I now find that the blue produce only
i while the white seeds yield some pods with all white, and some with
both blue and white Peas intermixed.’’—Goss, John, ‘‘On the Variation in
the Colour of Peas, occasioned by Cross Tuiprepnation. ?? In a letter to the
Secretary. Read October 15, 1822. Transactions of the Horticultural So-
-_ of London, Vol. 5, 1824, pp. 235.

T.r Ainsi done, en definitive il m’a paru qu’en général la ressemblance
de l’hybride à ses deux ascendans consistait, non dans une fusion intime des
divers caractèrs propres à chacun d’eux en particulier, mais bien plutôt dans
une guitars: soit égale, soit inégale, de ces mêmes caractères: je dis
égale u inégale, parce qu’elle est bien loin d’être la même dans tous les
individs hybrides provenant d’un même origine; et il y a entre eux, une
rès grande diversité. ’’_Sageret, A., ‘‘Considérations sur la praduction des
hybrides, des variantes et des variétés, en général, et sur celles de la fa
des Cucurbitacées en particulier.”? Annales des sciences naturelles, Vol. 8,
1826, p. 302.

aí Tous ces faits vont s’expliquer naturellement par la disjonction de
deux essences specifiques dans le pollen et les ovules de 1’hybride.’’

No. 543] SHORTER ARTICLES AND DISCUSSION 155

results in terms of numerical ratios as did Mendel. Knight?
and Gärtner”? as well as Wichura,?* Godron®® and Vilmorin?’

‘í Supposons, dans la Linaire hybride (L. vulgaris X L. purpurea) de
première génération, que la disjonction se soit faite A la fois dans l’anthere
et dans le contenu de l'ovaire; que des grains de pollen appartiennent totale-
ment à 1’espéce du père, d’autres totalement a 1’espéce de la mère; que dans
d’autres grains la disjonction soit nulle ou seulement commencée; admet-
tons encore que les ovules soient, au méme degré, disjoints dans le sens du
père et dans le sens de la mère; qu’arriverat-il lorsque les tubes polliniques
descendront dans l’ovaire et vont chercher les ovules pour les féconder?
Si le tube d’un grain de pollen, revenue à 1’espéce du père, recontre un
ovule disjoint dans le méme sens, il se produira une fécondation parfaite-
ment légitime, dont le resultat sera une plante entièrement retournée A
l’espéce paternelle.’’

Naudin, Ch., ‘‘ Nouvelles recherches sur ]’hybridité dans les végétaux,’”
Nouvelles archives du muséum d’histoire naturelle de Paris, Vol. 1, 1865,
pp. 150, 153

*<<Blossoms of a small white garden pea, in which the males had pre-
viously been destroyed, were impregnated with the farina of a large clay-
coloured kind with purple blossoms. The produce of the seeds thus ob-

were increased from seven to eight, to eight or nine, and not unfrequently,
in one variety to ten. The newly made gray kinds I found were easily made
white again by impregnating their blossoms with the farina of another white
kind. In this experiment some of the seeds in the same pod would produce
gray, = others white offspring, as occurs frequently in animals, which
bring many young ones at birth, when the breeds of the male and female
are of teat colours.’
ight, T: A, “A Treatise on the Culture of the Apple and Pear, and on
the Manufacture of Cider and Perry.’’ Ludlow, 1801, pp. 37-88, footnote..
26t Het allerduidelijkst evenwel is de invloed van leet vreemde stuifmeel
op de eitjes in dit opzigt bij de Leguminosen, wanneer b. v. de stempel van
Pisum sativum luteum met het stuifmeel van Pisum sativum viride bestoven
wordt, zoo ontstaan in deszelfs peulen zaadkorrels welke aan die der moeder
gelijk zijn, doch andere van eene groene en nog andere van eene gevlekte,
dat is, van eene groene en gele kleur. Deze zaadkorrels uit de eerste voort-
teling, zoowel de groene als de gele en de gevlkte, geven in de tweede voort-
teling pe, die hetzelfde verschil, als die uit de eerste voortteling,
opleveren,
Gärtner, C. F., ‘‘Beantwoording der Prysvraag over bastardeering,’’
N atuurkundige Verhandelingen van de le, Maatschappy der Weten-
pen te Haarlem, Vol. 24, 1838, p.
_ “ Wichura, Max, ‘‘ Die Baster dbežruchtung $ im Pflanzenreich, erläutert an
den Bastarden der Weiden,’’ Breslau,

156 THE AMERICAN NATURALIST [ Vou, XLVI

also seem to have come very near to the discovery of Mendel’s
principles.

III. THe Law or RECOMBINATION”*

This law is based pee the following Mendelian principle
(‘‘Versuche,’’ ete., p.

Die Nachkommen der Hybriden, in welchen mehrere wesentlich ver-
schiedene Merkmale vereinigt sind, stellen die Glieder einer Combina-
tionsreihe vor, in welchen die Entwicklungsreihen für je zwei differi-
rende Merkmale verbunden sind.

This principle may be thus translated :

The progeny of hybrids, in which several- essentially different char-
acters are combined represent the terms of a series of combinations, in
which the development series for each pair of differentiating characters
are united.

Spillman in his recent paper entitled ‘‘The Application of
Some of the Principles of Heredity to Plant Breeding,’’** thus
concisely states this law:

In the second generation of a hybrid there tends to occur every pos-
sible combination of the original parent characters.

This law was also discovered by Spillman independently 1m in
1901 and announced provisionally by him in a paper read before
the Association of American Agricultural College and Experi-
ment Stations in November of that year.?°

The clearest and most comprehensive account of Mendel’s work
extant is probably that of Bateson?’ in which may be found a
full discussion of the doctrine of Mendelism together with a
translation in English of Mendel’s original papers. The French

*Godron, D. A., ‘‘De l'espèce et des races dans les êtres organisés,”’
2d ed., Vol. 1, Paris, 1872, pp. 180-266.

*Vilmorin, E, “Note sur une expérience relative à 1’étude de 1’hérédité
dans les végétanx,? ’ Paris, 1879, pp. 1-11. Extrait des mémoires de la
société nationale d’agriculture de France—Année 1877.

"cí Gesetz der Selbständigkeit der Merkmale,’’ Correns, l. c., p. 208.

* Bulletin No. 165, Bureau of Plant Industry, U. S. Dept. Agriculture,
1909, p. 22.

* Bulletin No. 115, Office of Experiment Stations, U. S. Dept. Agri-
culture, "s
” Bat

» 88.
a, W, ‘*Mendel’s Principles of Heredity,’’ Cambridge, 1909.

No. 543] SHORTER ARTICLES AND DISCUSSION 157

translation by Chappellier,** and also the recent works by Baur,**
aecker,** and Goldschmidt,** will be found very useful to the
general student of Mendelism.
W. W. STOCKBERGER
U. S. DEPARTMENT OF AGRICULTURE

THE RANGE OF SIZE IN THE VERTEBRATES

A SMALL shrew, Microsorex winnemana, recently described
from Virginia by Preble, is said, if conclusions from but two
specimens may be drawn, to be the smallest species of shrew and
therefore the smallest mammal yet known from America. These
specimens measured 78 and 86 mm., respectively, in total length.
Microsorex hoyi (Baird) of the eastern and central states aver-
ages from 82 to 88 mm. and Sorex fontinalis Hollister and Sorex
personatus Geoffroy are but slightly larger. Blarina parva (Say)
a short-tailed shrew, averaging in total length about 75 to 80 mm.,
is in this respect smaller than M. winnemana, but much larger
in breadth, cranial and other characters. The smallest existing
mammal is probably a minute Crocidura of the subgenus
_ Pachyura from Madagascar.

The Insectivora, essentially an archaic and primitive group,
reached its highest development in point of numerous and di-
verse adaptations in the Middle Eocene, from which there has
been a gradual and steady decline. Sorex is known from the
Middle Oligocene, Talpa, the mole from the Upper Oligocene
and Erinaceus, the hedgehog from the Lower Miocene. These
are some of the oldest of existing genera of mammals. The
Malayan genus Gymnura, an Erinaceid, was said by Huxley’ to
possess the most generalized structure of all placental mammals.
The persistence of the group is without doubt due in a large part
to the small size and relative inconspicuousness of its members.

* Chappellier, A., ‘‘Recherches sur des hybrides végétaux (Traduction
francaise). Bulletin Scientifique de la France et de la Belgique, Vol. 41,
1907, pp. 371-420,

* Baur, E., ‘‘Einfiihrung in die experimentelle Vererbungslehre,’’ Berlin,
1911,

= Haecker, V., ‘‘ Allgemeine Vererbungslehre,’’ Braunschweig, 1911.

* Goldschmidt, R, ‘‘ Bintührung in die Vererbungswissenschaft,’’ Leip-
zig, 1911.

* Proc. Biol. Soc. Wash., Vol. 23, P. 101, 1910.

? Proc. Zool. Soc. London, p. 657, 1

158 THE AMERICAN NATURALIST [ Vou. XLVI

The separation of Madagascar from Africa has permitted the
continuance of the relatively large Centetide, the tenrecs, of
which Ericulus, Centetes and Hemicentetes have developed a
spiny coat and Limnogale has become aquatic. Similarly, the
rare and interesting Zalambdodont, Solenodon, has been able to
continue an existence by isolation in Cuba and Hayti. Even
thus isolated, Solenodon, judging from its extreme rarity, barely
maintains a foothold. The aberrant Tupaiide, or Oriental tree
shrews, are, as indicated by their name, arboreal and the African
Macroscelididæ, or elephant shrews seek refuge by leaping and
by skulking. The South African Chrysochloride, or golden
moles, and the familiar Talpide are subterranean. A subter-
ranean habitat implies a restricted stature. Erinaceus, the well-
known hedgehog, and the African Potamogale, the only rela-
tively large non-insular insectivores, are well protected, the
former by its spiny coat and the latter by its aquatic habits.
The Soricide containing the smallest members of the order are
largely nocturnal. During the long stretch of time since the
Eocene culmination of the group and the gradual evolution of
more modern mammalia, the Insectivora have become extinct
with the exception of those especially protected, insular, sub-
terranean or of insignificant size. Smallness here seems an
attendant trait of archaism. The earliest American mammals,
the Triassic Protodonts, Dromatherium and Microconodon and
among recent mammals certain Murine rodents closely simulate
this diminutive stature.

The massive Rorqual whale, Balenoptera sibbaldii Gray, of
the North Atlantic, sometimes reaching a length of eighty-five
feet, is the bulkiest vertebrate which has ever existed. The
Cetacea are likewise a primitive and probably degenerate group.
Other aquatic mammals, such as the Sirenia, Pinnipedia, etc.,
similarly reach immense proportions, due, very likely, to the lack
of a compensatory element in the environment. The tallness of
the giraffe, which is an adaptation to arboreal grazing, produced
by an elongation of the cervical vertebræ, coordinated of course
with limb structure, has independently arisen in at least one
other family. The giraffe-camels of the genera Oxydactylus and
Alticamelus, respectively, of the American Oligocene and Mio-
cene, parallel the existing giraffes.

The smallest known bird is Calypte helene (Gundlach) of

No. 543] SHORTER ARTICLES AND DISCUSSION 159

Cuba, measuring in total length but 57 mm.* Several other hum-
ming birds, notably Mellisuga minima (65 mm.) of Jamaica, are
but slightly larger. The Trochilide comprise about 600 known
forms, most of which are excessively small. Patagona gigas of
the higher Andes, the largest of the group measures about 215
mm. in length. The size of the members of this family is an
adaptation to the physical requirements of a highly active life,
which is essentially that of securing food while hovering before
blossoms. The Nectariniide or sun-birds of the Ethiopian and
Indian Regions parallel the Trochilide in size and brilliancy.
The Dicæidæ or ‘‘ Flower-peckers ’’ of India and Australia, and
the Troglodytide and Regulide, wrens and kinglets of this
country, likewise contain very small forms.

Spherodactylus sputator Gray (45 mm. plus), occurring in
the West Indies, is one of the very smallest of American rep-
tiles. S. notatus Baird (50 mm.) of Florida and the West Indies
and a number of other West Indian lizards are hardly larger.
The largest of existing lizards are members of the Varanide of
Africa, Asia and Australia, the Malayan Varanus salvator
(Gray) reaches a length of eight feet. Of the adaptive radia-
tions which the Lacertilians have undergone, that branch con-
taining the Amphisbenidx, minute, legless, burrowing forms of
tropical America and Africa, is of peculiar interest. This family
presents a most remarkable illustration of the principle of
heterology of Cope, of parallelism and convergence of other
authors. Through the dominance of essentially similar environ-
mental factors, certain species have come to resemble the earth-
worm with such fidelity, that the very chickens which follow
the plow are said to seem unable to differentiate and indiserim-
inately pick them up.’

These remarks may apply equally well to the Typhlopide
and Glauconiide, still widely distributed in tropical countries,
archaic and degenerate, sightless, burrowing snakes, relicts of a
once cosmopolitan assemblage, which contain the very smallest
known Ophidians. Helminthophis petersi Boulenger (110 mm.)
of Ecuador, Typhlops anchiete Bocage (119 mm.) of Angola,
Africa, Glauconia dissimilis (Bocage) (104 mm.) of the White
Nile and Glauconia bilineata (Schlegel) (110 mm.) of the West

* Ridgway, Rep. U. S. Nat. Mus., 1890, p. 295 (1892).

* Boulenger, ‘‘Cat. Lizards Brit. Mus.,’’ 2d ed., Vol. 1, p. 219, 1885.

* Eigenmann, Biol. Bull., Vol. 8, no. 2, p. 60, 1905.

160 THE AMERICAN NATURALIST [ Vou. XLVI

Indies, are some of the smallest species. The largest existing
reptile is the Ganges crocodile, Gavialis gangeticus (Gmelin) of
northern India, which is said to attain a length of thirty feet or
more. This, however, is insignificant in comparison with that
attained by the sauropodous dinosaurs of the Mesozoic. The
American genera Atlantosaurus, Brontosaurus, Camarasaurus
and Diplodocus were immense creatures. Atlantosaurus im-
manis Marsh of the Wyoming Upper Jurassic, supposedly terres-
trial from mechanical considerations, is one of the largest land
vertebrates which has ever existed, probably upwards of one
hundred feet in length. The uncompensated extravagance of
energy in the maintenance of such immensity, coupled with the
small and primitive brain, were without doubt to a great extent
factors in the extinction of these gigantic vertebrates. These
animals are without living descendants. _

The familiar ‘‘spring peeper,’’? Hyla pickeringii (Storer),
about 20-28 mm., of eastern North America is the smallest
American Hyla and thus one of our smallest batrachians. The
Hylidx, comprising about 150 species, have a practically cosmo-
politan distribution with the exception of Africa and the Malay
Archipelago. Essentially arboreal, in the dense, steaming
tropical forests of South America they attain the highest diver-
sity of generic and specific types. H. pickeringii does not ordi-
narily ascend into the trees until early in autumn. A number
of true frogs are as small or smaller, thus Arthroleptis sechel-
lensis Boettger of the Seychelles, interesting from its habit of
carrying its eight or nine tadpoles affixed by their ventral sur-
faces to its back, is but about 20 mm. in length of head and
body. The smallest salamander of the eastern United States is
the red-back, Plethodon cinereus erythronotus Green, an elon- ©
gate form of about 75 mm. The largest salamander and the
largest existing batrachian is Megalobatrachus japonicus (Tem-
minck), the giant salamander of Japan, which reaches a length
of over five feet. At no time have the Batrachia been the domi-
nant type of vertebrate life, either in size or variety of forms.

It is among the fishes that we find the smallest known verte-
brates. Thus, Mistichthys luzonensis H. M. Smith,® an extraor-
dinarily minute goby from Lake Buhi, Luzon, P. I., is accorded |
this distinction. The average length of this species is 12.9 mm.
The average for egg-bearing females which exceeds the average

* Science, N. S., Vol. 15, p. 30, Jan. 3, 1902.

No.543] SHORTER ARTICLES AND DISCUSSION . 161

for males by one millimeter is 13.5 mm., the maximum is 15 mm.,
and the minimum is under 12 mm. This species is said to be a
food fish of considerable importance as it is seined by the natives
in large quantities, dried in cakes or mixed with spices, and is
eagerly sought for. The great majority of the six hundred
known species of Gobiide are less than 75 mm. in length.
Philypnus dormitor (Lacépède) of Central America and the
West Indies, the largest of the family, reaches a length of
600 mm.

The Etheostomine or darters contain some of the smallest
spiny-rayed fishes. Microperca punctulata Putnam (25-38 mm.)
of the Central States is next to Elassoma evergladei Jordan
(20-33 mm.) a minute percoidean of our southeastern swamps,
the smallest American spiny-rayed fish and a number of other
darters are but slightly larger. Heterandria formosa Agassiz,
a diminutive viviparous Peeciliid, occurring in swamps and
ditches from South Carolina to Florida, was for a long while
considered the least of American vertebrates. This species has
an average length of 19 mm. for males and 25 mm. for females.
Acanthophacelus bifurcus of the ponds and creeks of British
Guiana, recently described by Eigenmann,’ is still smaller; the
average for both sexes is 21.5 mm. and the largest of 74 speci-
mens is 29 mm. long. A number of pregnant females are but
20 mm. in total length and one of these, preserved in alcohol, in
the Indiana University collections weighs but .076 gram. The
weight of living specimens is probably somewhat greater. The
exact measurements of this specimen (I. U. No. 11,765) are as
follows: total length 20 mm., length to base of caudal 16 mm.,
depth 5 mm. and breadth or thickness 3 mm. The smallest male
specimen here, one in full nuptial coloration, measures 19 mm.
over all. In all probability, the smallest new world vertebrate is
Heterandria minor Garman,’ from Villa Bella, Brazil. The
average total length of this species is 18 mm. for males and 20.5
mm. for females. Females of but 19 mm. ‘‘contain fully devel-
oped embryos.’? With this may be contrasted the bulk ot
Arapaima gigas (Cuvier) of neighboring streams of Brazil and
Guiana, said to reach a length of fifteen feet and to be the largest
Strictly fresh-water fish known,

" Annals Carnegie Museum, Vol. 6, p. 52, 1909.

` Memoirs Mus. Comp. Zool., Vol. 19, no. 1, p. 92, 1895.

162 THE AMERICAN NATURALIST [ Vou. XLVI

The Peeciliide or killifishes, found in most warm portions of
the globe, comprise about two hundred very small species, the
largest Fundulus, Anableps, ete., seldom exceed 300 mm. In
the inviting streams of Central America where the majority of
species occur, adaptive radiation, as in every large family of
fishes, has taken place, from the central type Fundulus, result-
ing in the depressed catfish-like Rivulus, the beaked garpike-like
Belonesox, the sunfish-like Goodea, the earp-like Cyprinodon, ete.

The origin of these small forms may probably be explained
by the selective migration and the successive adaptation of the
species occupying the deeper reaches of the streams. The
Peeciliide usually feed at the surface and thus may tend to
disseminate throughout the full extent of the shallower waters.
Acanthophacelus has evidently been derived from Pecilia, from
which it differs chiefly in the acquirement of two rows of retrorse
hooklets on the modified anal fin of the male, while Heterandria
is of a different type with conical, carnivorous dentition and
shortened alimentary tract. Further exploration may reveal
still smaller species of these interesting little fishes, which are
likely, however, to pass unobserved by all but the trained
naturalist.

ARTHUR W. HENN

INDIANA UNIVERSITY

BLOOMINGTON, INDIANA:

NOTES AND LITERATURE
HEREDITY

F. E. Lutz contributes a very interesting paper* in which he
presents evidence bearing on the effect of selection in a ‘‘pure”’
strain. Amongst wild flies of this species (Drosophila ampel-
ophila) about one third of one per cent. present abnormalities in
wing venation. By selection and inbreeding he was able to pro-
duce strains with practically 100 per cent. abnormality. In a
good many crosses the abnormal venation behaved as a Mendelian
recessive in which dominance of normal was not perfect. But
in other cases the phenomena observed departed widely from
the typical Mendelian behavior. The author suggests that the
abnormality is due to a factor which has high fluctuating vari-
ability, and that these fluctuations are to a certain extent in-
herited. It is shown that Galton’s law of ancestral inheritance
does not apply to individual families, even when these are large
(100 to 200 individuals).

If we knew the real cause of this abnormality we might pos-
sibly be able to reason about its peculiar behavior with some
hope of elucidating it. Any attempt at explanation must
necessarily be largely hypothetical. The writer would suggest
that most of Lutz’s data point to the presence of two factors for
abnormality, one much stronger than the other, and one or both
of them highly variable (fluctuating variability). Davenport
found such a case (two factors, one stronger than the other).
for the hood in certain fowls.? The writer has often thought
there ought to be cases of Mendelian factors which just barely
give rise to visible expression in the soma. Since all characters
exhibit fluctuating variability, such a factor should behave much
as the character which Lutz studied.

Lutz concludes that he has produced effects by selection in ‘‘ pure
lines.”? This is much to be doubted. The results he secured by
Selection fit very well the curves shown in my paper on ‘Some
of the Principles of Heredity Applied to Plant Breeding’ ”™” for

3‘ Experiments with Drosophila Ampelophila Concerning Evolution, ’’
F. E. Lutz, Carnegie Institution of Wash., 1911. `

* Proc. Wash. Acad. Sci., ‘Lecture on Heredity, ”” by C. B. Davenport.

* Bulletin Bureau of Plant Industry, No. 165.

| 163

164 THE AMERICAN NATURALIST [ Von. XLVI

selection in dihybrids. Lutz’s results seem to be further compli-
cated by the fact, to which he calls attention, that the great
variability of the character in question gave apparent normals
that were really abnormals genetically, and this very fact ren-
dered the establishment of really pure lines a matter of great
difficulty, if indeed not an impossibility. A line is not necessar-
ily pure (genetically) because it has descended through many
generations of the closest possible inbreeding from a single orig-
inal pair. It is only pure in a Mendelian sense when all its alle-
lomorphs are duplex and the members of each pair of them are
identical in character. An individual might be pure in this
sense even though both its parents were complexly bred mon-
grels. On the other hand, if Lutz’s original pair were not thus
pure, their descendants would only by chance constitute a pure
line. It is very doubtful if Lutz has really proved the effect of
selection in modifying a pure line, for it is not established that
he worked with a pure line, in the sense in which that term
is now used.

Professor E. B. Wilson has recently published an excellent
review of cytological investigations relating to sex inheritance."
There appears now to be no doubt of the relation of sex to cer-
tain chromosomes in dicecious organisms. This relation is also
interpretable from the Mendelian viewpoint, so that we may say
that sex inheritance belongs in the category of Mendelian phe-
nomena. In nearly all organisms thus far studied the female is
to be regarded as homozygous and the male heterozygous for
sex. Hence the male produces two kinds of spermatozoa, while
the females produce only one kind of egg. When an egg is fer- _
tilized by the one type of spermatozoa the zygote is a female; if
by the other, the zygote is a male. In general, the female dip-
loid nuclei contain two of the sex chromosomes or two groups
of them when the sex chromatin element is compound. The
nuclei of the male contain only one of these elements, with or
without a synaptic mate different from it in character. In sea
urchins it has been shown that it is the female that is heterozy-
gous. But in this case the female possesses one of the sex ele-
ments and the male none. In both cases the female is charac-
terized by an excess in the number of the sex elements as com-
pared with the male.

A large number of ordinary Mendelian characters have been

“**The Sex Chromosomes,’? Arch. f. Mikrosko. Anat., Bd. 77, 1911,
pp. 249-271,

No. 543] NOTES AND LITERATURE 165

shown to be sex limited in inheritance in such a manner that it
seems certain they are in some way connected either with the
sex-determining chromatic element or with the synaptic mate
of this element (in the heterozygous sex). This links these
characters with the chromosomes, and further strengthens the
chromosome theory of Mendelian inheritance.

The writer is fully aware of the danger, in dealing with sta-
tistical data, of drawing conclusions from insufficient data or
from data that have not been submitted to a full mathematical
analysis, such as the determination of means, norms, standard
deviations, and coefficients of correlation. Although I have no
very full or accurate data on the subject and have not actually
determined statistically the coefficients of correlation between
the characters about to be mentioned, I am of opinion, after
more or less superficial observation extending now over nearly
half a century, that there is a noticeable degree of correlation
between positiveness of statement, and inaccuracy of statement.
An illustration of this correlation will be found in a paper by Dr.
J. Arthur Harris, published in the November (1911) number of
this journal. It is now fairly well established that the norms
of a group of related genotypes can, in some cases at least, be
arranged in a frequency curve. In my review of de Vries’s
“The Mutation Theory’’? I cite this fact to show that the type
of variation represented by these genotypes comes under the head
of what de Vries there defines as ‘‘continuous variation,” as
opposed to the ‘‘discontinuous variation,’’ in which the norms of
the variants can not be thus arranged. In Dr. Harris’s paper he
represents me as having cited the fact that these genotype norms
form a frequency curve as a proof of the genotype hypothesis.
I have not been able to find the time to look up other similar
citations in this paper to see whether the same inaccuracy applies
to them. The type of correlation we are considering is further
illustrated in this paper in Dr. Harris’s definition of genotype,
from which he omits the word ‘‘homozygous.’’ The definition
is further inaccurate in including clonal varieties under the
definition of genotype. The author of the term genotype, in a
recent lecture in this city, defined it as ‘‘the descendants of a
single homozygous individual propagated by self-fertilization.’’

= AMER. NAT., Dee., 1910. ;

* Since the above was written it has been pointed out by several writers
that the word genotype should be replaced, im the above sense, by ‘‘ biotype”?
or ‘pure line.’’

166 THE AMERICAN NATURALIST [Vou. XLVI

THE DISTRIBUTION OF PLANTS IN NORTH
| A

AME

A COMPREHENSIVE account of the phytogeography of North
America has long been a desideratum, and it will be welcome
news to botanists that this need has at last been filled. The
work of Dr. Harshberger,' just issued, is an imposing volume,
which will be quite indispensable to all students of plant dis-
tribution, and evidently represents an enormous amount of
labor on the part of the author. A fair criticism that might be
made is that it perhaps is too comprehensive. The book in the
writer’s opinion could have been very considerably reduced in
volume without lessening its value, by drastic cutting of the
first three parts, which are really introductory, and yet take
up nearly half the book. The really essential facts concerning
the geology, geography and climatology of North America could
certainly have been presented in much less space. There is much
needless repetition, and the arrangement of the topics is at times
very confusing.

In addition to Dr. Harshberger’s own work there is a summary
of the book, sixty-three pages in extent, written in German by
the editor of the series, Professor O. Drude.

It is the fourth section of Dr. Harshberger’s book which will
undoubtedly be of most value to the average botanical student.
This deals with the North American phytogeographical regions,
formations and associations, and comprises an account of the
floras not only of the whole North American continent, but of
the West Indies as well. |

Nearly one hundred pages are devoted to the history of work
published on the floras of various parts of North America, the
bibliography alone comprising forty-six pages. This makes up
Part One of the work. The second part—Geographic, Climatic
and Floristic Survey—ocecupies seventy-seven pages. In this
section are given very much in detail an immense body of facts,
many of which could have been omitted without materially less-
ening the value of the summary. Nevertheless it contains very
much that the student will find exceedingly useful.

This section makes very clear the great range of conditions on
the North American continent, which explains the richness and
variety of its flora. Very full statistics are given in this chap-

*««Phytogeographie Survey of North America,’’ John W. Harshberger,
‘*Die Vegetation der Erde,’’ Vol. XIII.

No. 543] NOTES AND LITERATURE 167

ter, including many tables, temperature, rainfall, ete. From
the eternal ice and snow of Greenland to the equatorial condi-
tions on the Isthmus of Panama, practically the whole gamut
of climate depending upon latitude is covered. Moreover, the
great variety in topography in North America results in pre-
cipitations ranging from practically nothing in the extreme
deserts of the arid southwestern United States and Mexico, to a
rainfall of two hundred inches or more in some tropical regions
like parts of Jamaica.

The presence of lofty mountains permitting a rich Alpine flora
in the northwestern parts of the United States and Canada, is
also an interesting feature of the vegetation of North America.
In his division of the phytogeographical regions Harshberger
recognizes four main divisions: Northern, Central, Southern and
West Indian. The United States lies almost entirely within the
second of these, only the region of the Great Lakes and a small
area in the Gulf region encroaching respectively upon the
northern and southern zones. The natural phytogeographical
areas, however, are by no means determined by latitude alone,
this being especially the case in the central zone, where there
are several quite unrelated fioristic regions lying in practically
the same latitude. Such, for example, are the eastern forest area
and the western Cordilleran region.

The pronounced continental climate of the great plains, with
its yearly range of temperature from arctic cold to tropic heat,
may be contrasted with the uniform mild climate of the Pacific
Coast, where longitude plays quite as important a rôle in deter-
mining the distribution of plants as does latitude. For example,
Victoria in latitude 48 has a mean temperature for January of
41.9° F., while St. Louis, almost 10° further south, is 12 degrees
colder. This illustrates the differences in the conditions between
the Pacific Coast region and the pronounced continental climate
of the Middle West. i

The southern area, including Mexico and Central America, 1s
for the most part tropical in its climate, except for the temperate
uplands of Mexico. Within the United States, Florida and the
Southernmost tier of states adjacent to Mexico, really belong to
the southern area rather than to the central one.

In the third part of his volume Dr. Harshberger treats at
great length the geologic history of North America, and its bear-
ing upon the origin and distribution of the present flora. In

168 THE AMERICAN NATURALIST [Vou. XLVI

the course of this section he expounds his own views as to the
origin of the North American flora, and the factors which have
brought about its present distribution. While this section con-
tains much important and interesting material, it can not be
said that its arrangement is all that could be desired. After a
perusal of the 174 pages contained in it, there is left in one’s
mind a very confused impression of a conglomeration of geology,
phytogeography, and the origin of floras, together with other
more or less disconnected topics. A large part of this section
deals again with the phytogeographical areas already discussed
at length in the previous section, and repeated in the fourth
part of the book, but deals with them from the standpoint of
centers of distribution rather than merely from latitude. It is
to be regretted that the subject of the phytogeographical areas
has not been treated as a connected whole, and also materially
reduced, so as to bring out the salient points, instead of present-
ing a mass of details of very varying importance.

While, as might be expected, the plants treated are for the
most part the flowering plants, still the lower plants are not en-
tirely neglected, the Algæ especially coming in for a fuller
treatment than is usually accorded them in works on plant dis-
tribution.

As the flowering plants, especially Angiosperms, are the pre-
dominant plants of the existing land floras, the geological history
of the country before the Cretaceous is of relatively small impor-
tance as bearing upon the distribution of the existing vegetation,
since it is not until the Cretaceous, or at the earliest the Sub-Cre-
taceous is reached, that recognizable remains of Angiosperms
are met with. Professor Harshberger gives an excellent ac-
count, with maps, of the distribution of land and water areas
during the Cretaceous and their significance as affecting the
future distribution of North American plants.

During the lower Cretaceous a solid land mass occupied ap-
proximately the present area of North America, but was sepa-
rated from South America by a wide gap through what is now
Central America, allowing free communication between the At-
lantic and the Pacific. At that time also there was still a broad
land connection with Western Europe. Throughout the north-
ern area both of the old and new world, it is evident that a
_ very uniform vegetation flourished, with Conifers as the predom-
inant trees.

No. 543] NOTES AND LITERATURE 169

During the later Cretaceous extensive changes took place.
The great western mountain chains were thrown up and com-
munication with South America was established; but there was
a complete separation of the eastern and western land masses.
A broad body of water extended without interruption from the
Arctic Ocean to the Gulf of Mexico, and the whole of the great
plains region to the east of the Rocky Mountains was then sub-
merged. Professor Harshberger believes that this was the per-
iod which determined the separation of the Eastern and Western
coniferous floras, and the foundation of the great differences in
the general floral characters of Atlantic and Pacifie North
America. :

The Cretaceous is notable in the history of the vegetable king-
dom as the time when the highest types of plants first appeared,
or at any rate when they are first recognizable. There are no
certain evidences of Angiosperms previous to the Sub-Creta-
ceous. Apparently the Cretaceous was a period of great devel-
opment of new plant forms, perhaps conditioned by the great
changes in the physical character of the continent, these changes
undoubtedly causing great changes of climate as well. Harsh-
berger thinks that the Cretaceous was probably characterized by
periods of mutation—used in the sense of De Vries. Whatever
may have been the reason, during the Cretaceous most of the
modern angiospermous types, and many of the living genera
appeared. The great feature of the later Cretaceous and Ter-
tiary forests was the preponderance of dicotyledonous trees and
shrubs whose remains are found in great numbers. It is certain
that many existing herbaceous genera must also have been
present, but with rare exceptions these have not been preserved
as recognizable fossils, no doubt owing to their perishable tis-
sues which were not fitted to leave fossil remains except under
exceptionally favorable conditions.

he high northern extension of the Cretaceous and Tertiary
forests, especially in the Miocene, shown by the fossil deposits
as far north as Greenland and Spitzbergen, indicates a relatively
warm climate for these higher latitudes of the northern hemi-
Sphere. Moreover this northern forest belt seems to have been
much the same throughout the whole extent of the northern
hemisphere, very many species occurring both in the eastern
and western hemisphere. Many of the most characteristic of
the Miocene genera still exist in America, especially in the great

170 THE AMERICAN NATURALIST [ Vou. XLVI

deciduous forest region of the southern Appalachians. Among
the widespread Miocene genera which still exist in America
may be mentioned Magnolia, Liriodendron, Vitis, Sassafras,
Aralia, Nyssa. Indeed it is supposed that some of the Miocene
species still survive. Among these are the southern Cypress
(Taxodium distichum), Balsam Poplar (Populus balsamifera),
Sweet Gum (Liquidambar styraciflua), Black Walnut (Juglans
nigra) and many others.

With the redistribution of the forest flora due to the glacial
epoch, many of these types disappeared from western America
and Europe but have survived in eastern America and eastern
Asia. It is clear also that genera now confined to the Old World
also existed at that time in North America. A long list of the
Cretaceous and Tertiary genera of North America shows, for
example such extra-American genera as Banksia (Australia),
Casuarina (Australia and Malaysia), Encephalartos (South
Africa), Eucalyptus (Australia), Ginkgo (China), Hedera (Eu-
rope, Asia), Sterculia (tropics of the Old World and Austral-
asia).

During Tertiary times the eastern and western parts of tle
continent were again united, and there seems to have been a
fairly uniform northern flora throughout the Atlantic and Pa-
cific regions, although the separation of the Atlantic and Pacific
floras was already indicated, the difference in climate no doubt
being influenced by the changes in elevation caused by the up-
lifting of the great western mountain masses, inducing a drier
climate in the interior of the continent. It seems probable that
to the north of the great forest belt extending across the con-
tinent was an arctic flora, remnants of which are found in the
Alpine regions of the higher mountains where they presumably
took refuge after the migrations subsequent to the great glacial
epoch.

It is pretty generally agreed that the present distribution of
plants in America is mainly the result of the advance of the
great glacial sheet from the north. In the southward retreat
of the great forest belt there was a division which, it has been
suggested, was caused by the presence of the dry ‘plains of the
interior, which were unfitted for forest growth, and thus acted
as a wedge separating the western forest, predominantly of
coniferous trees, from the eastern forest mainly deciduous in
character. Whether or not this is the true explanation, the fact

No. 543] NOTES AND LITERATURE 171

remains that at present the western forest is composed in the
main of a great variety of coniferous trees adapted to dry sum-
mer conditions, while the eastern forest is distinguished by its
great richness in deciduous trees adapted to a humid summer
climate. Close to the retreating line of the forests, there pre-
sumably followed a belt of arctic vegetation clinging to the edge
of the advancing glaciers.

With the retreat of the glaciers the plants advanced north-
ward again, and the present distribution was finally established.
Owing to the much greater altitude of the western mountains, it
is especially upon these that we find a true Alpine flora, the more
or less changed remnants of the arctic vegetation forced south-
ward by the advancing ice sheet. It is true that a few arctic
plants, like the Greenland sandwort, occur upon the highest
peaks of the Appalachian system, but the number of these is com-
paratively insignificant when compared with the rich and beauti-
ful Alpine flora of the Canadian Rockies and the Cascades.

At the close of the glacial epoch the distribution of the plants
was practically as at present, the most marked features being the
eastern and western forest areas and the great open region of
the great plains.

Part four deals with the floral regions as they now exist in
North America. Probably most botanists will feel that the
weakest feature of this account of the flora of North America is
its lack of proportion. While the relatively uniform flora of the
Atlantic half of the continent receives over 150 pages, the Rocky
Mountain flora is dismissed with 23 pages, and the extraordi-
narily varied and interesting flora of the Pacific Coast with less
than fifty. This no doubt is to be explained by the author’s
very intimate acquaintance with the flora of the Atlantic states,
and his evidently very casual observations at first hand upon
that of the western third of the continent; but he has not shown
the best judgment in the selection of topics for discussion, deal-
ing at great length with relatively very unimportant regions.
Nearly as much space is given to a discussion of the pine-barren
flora of the Atlantic Coast states as to the whole of the Pacific
Coast from Alaska to Mexico. The multiplication of details
in the discussion of the floras of small and unimportant areas,
results in a general sense of confusion, and one is often at a loss
to see the bearing of many of the detailed descriptions of local
floras upon the larger problems. Itis a pity that the abundant

172 THE AMERICAN NATURALIST [VoL. XLVI

materials had not been better digested, and much might just as
well have been omitted. The contrast between the over-elabora-
tion of many topics dealing with the Eastern flora, and the very
casual treatment of the far more important and divergent condi-
tions of the Rocky Mountain and Pacific region, is both striking
and regrettable. However, with all these serious faults, the
patient reader will be able to get a fairly comprehensive view of
the extremely diverse features making up the flora of the North
American continent.

With the last retreat of the great ice sheet, there moved into
the territory thus uncovered plant migrants from various sources.
Keeping close to the retreating ice sheet are the various strictly
arctic or glacial types, and south of this northernmost flora is
the extensive development of bog and tundra formations, which
form so important a feature in the arctic and subarctic regions
of Alaska, Canada and Greenland.

This northern area in America shows three sections, ranging
from east to west. Greenland, separated from the mainland by
the deep water of Baffin’s Bay and Davis Strait, belongs really
with Europe, its scanty flora being composed almost exclusively
of European types like those of Scandinavia and Lapland. The
middle region extending from Labrador to the Mackenzie River:
is marked by a larger preponderance of purely American types.
Westward, including Alaska, the aretie flora is again charac-
terized by a preponderance of Old World species. Out of 364
species of Western Arctic America no less than 320 occur in
Northwestern Asia, many of them extending southward to the
Altai and Himalaya Mountains. Of the 379 species belonging to
the middle region 73 species are strictly American, the other 306
occur also in the arctice regions of the Old World.

The characteristic ‘‘ tundras ’’ are plains whose subsoil is per-
manently frozen, but whose surface soil is covered with a dense
growth of mosses or lichens, among which grow in places dwarf
Alders, Willows, Birches, and various ericaceous shrubs, e. g.,
Cassiope, Andromeda, Vaccinium, Ledum, ete., as well as many
herbaceous plants, some with showy flowers like the Iceland
poppy. It is evident that the arctic flora is remarkably uniform _
throughout the whole northern hemisphere.

In connection with a study of the land plants some interest-
ing features may be noted in regard to the arctice American
Alge. There is a remarkable difference between the Algæ of the

No. 543] NOTES AND LITERATURE 173

arctic regions of the Atlantic and the Pacific, many striking
genera including the giant Kelps being quite unrepresented on
the Atlantic side of the continent. Harshberger thinks that
this points to these two regions as being distinct centers of de-
velopment and distribution, and it may also show that some of
the extraordinary Kelps of the Pacific Coast may be relatively
recent developments.

South of the arctic zone is a sub-arctic forest zone, also of
very similar composition throughout its whole extent from Labra-
dor to Alaska. This forest is largely composed of Conifers, the.
predominant trees being White and Black Spruce, Balsam-Fir,
Tamarack and Scrub pine.

With the Conifers are associated Aspens, Balsam Poplars, Wil-
lows, Alders and Birches. Swamps and peat bogs abound and
in these as well as in the forest itself grow many attractive
shrubs and herbaceous plants. Ericaceous shrubs are especially
abundant, Rhododendron, Kalmia, Ledum, Vaccinium, etc., as
well as wild Roses, species of Rubus and Ribes, ete. In the
Sphagnum bogs, especially in the eastern portions, are found
Pitcher plants, Sundew, and several beautiful Orchids. A con-
siderable number of species, e. g., Pyrola rotundifolia, Linnea
borealis, Cornus Suecica, are identical with Old World species,
and most of the genera, although not the species of trees, are the
same.

To the southward of this uniform forest zone a greater diver-
gence in the floras is noted, and within the United States the
character. of the vegetation changes radically as one proceeds.
from the Atlantic seaboard westward.

The whole of the eastern third of the United States may be
said to comprise one great phytogeographical area, although of
course there are certain regions where the flora has a more or less
marked type of its own. Nowhere within this area are the
mountains of sufficient elevation to form barriers against the
ready migration of plants, and the whole region is one of abun-
dant rainfall. The whole area being continuous, the result is a
very uniform flora, when contrasted with the Pacific side of the
continent. Many species of trees, e. g., Elms, Oaks, Walnuts,
Hickories, ete., occur throughout the range and the same is true
of very many herbaceous plants. It is true that the number
of species is much greater in some parts than others, owing to
more favorable conditions of soil and temperature, but there are

174 THE AMERICAN NATURALIST [ Vou. XLVI

no such radical differences within the area as those, for instance,
between the Southern California desert and the Redwood belt
of the coast region of the same state. The most important fea-
ture of this Eastern area is the great development of the dicoty-
ledonous forest trees. Probably nowhere outside the tropics is
to be found a richer forest fiora, and the trees, both by their
size and variety, constitute a truly magnificent forest. This
forest reaches its finest development in the Southern Alleghenies
of western North Carolina and in parts of the Ohio and Missis-
. sippi Valleys. Unlike the northern forest belt and the forests
of most parts of Europe, this forest is composed of a great many
species intermingled. It is stated by Harshberger that probably
the richest forest vegetation in the north temperate zone is found
in the lower Wabash Valley in Illinois, where one hundred and
seven species of trees occur. In an area of less than a square
mile in extent seventy-five species of trees were counted.

While many of these trees belong to European genera, e. g.,
Poplar, Beech, Birch, Oak, Elm, Ash, etc., many genera are
absent from the European forests. Such, for example, are the
gums (Liquidambar and Nyssa), Magnolias, Tulip trees (Lirio-
dendron), Hickories, Locusts (Gleditschia and Robinia), the Cy-
press (Taxodium), ete. Some of these trees reach really gigan-
tic size, in regard to which Harshberger gives some striking fig-
ures. Thus a Red Oak is cited which was seven feet in diameter,
with a height to the first branch of ninety-four feet, and a total
height of one hundred and eighty-one feet. A White Oak and
Bur-Oak of nearly equal size.are mentioned, while a Cotton-
wood (Populus monilifera) was eight feet in diameter and one
hundred and ninety feet in height. Thus it will be seen that
some of the Eastern deciduous trees almost rival the Pacific
Coast Conifers in size. Other giants of the eastern forest are
the White Pine, the Sycamore and Tulip tree. With these are
associated many shrubs and climbing plants, as well as a charac-
teristic flora composed of herbaceous plants, usually flowering in
the early spring before the leaves appear upon the trees.

In the Mississippi Valley the forest is almost exclusively com-
posed of deciduous trees, but in the Appalachian forest there is
a mixture of coniferous trees, especially Pines and Hemlocks,
while in the north the proportion of these Conifers increases
and at the same time many of the deciduous trees disappear, and
the forest then consists of a comparatively small number of spe-

No. 543] NOTES AND LITERATURE 175

cies. In many parts of the northern states there is a prepon-
derance of Beech and Maple, with a comparatively small number
of other trees growing with them, but nowhere in the real forest
area do we find the pure forests of Beech or Oak which are
common in northern Europe.

In marked contrast with the preponderance of deciduous for-
ests over most of the eastern states, are the areas known as Pine-
barrens, characterized in certain districts by an almost pure for-
est of Pines, usually of a single species. Professor Harshberger
gives a very exhaustive account of the Pine-barren flora, which
he has evidently studied with great care.

At the north the prevalent Pine is Pinus rigida, with which
are found associated some scrub oaks, Nyssa, Liquidambar,
Sassafras and others, as well as a rich growth of showy ericace-
ous shrubs. The poor sandy soil is especially adapted to Erica-
cee, and Huckleberries, Azaleas, Kalmias, Rhododendrons, ete.,
are conspicuous features of the undergrowth. There are also
extensive cranberry bogs where many beautiful Orchids occur,
as well as Pitcher plants, Sundews and other interesting bog
plants. In the coastal bogs of North Carolina grows the unique
Venus’s fly trap (Dioncea) and the long trumpets of the South-
ern Pitcher plants.

Over extensive regions of the Gulf states reaching into Texas
are extensive forests of Pines, whose value as timber trees is only
too well appreciated. The most important of these southern
Pines is the long-leaf pine (P. palustris).

In the southern portion of the Atlantic states there is an infu-
sion of tropical types; the most striking of which are the Palms,
the common Palmetto reaching as far north as the coast regions
of North Carolina. Certain genera of unmistakable tropical
affinity reach even to the northern states. A conspicuous ex-
ample is the Paw-paw (Asimina), a representative of the charac-
teristic tropical family of Custard-apples. In the Gulf states
there are many examples of plants of tropical affinities, includ-
ing certain Orchids and members of the Pineapple family, of
which the most familiar example is the Spanish moss, Tillandsia.

The southern part of Florida is very different in its flora from
the northern portion, which is essentially the same as that of the
other Gulf states. The southern extremity of Florida together
with the Florida Keys really belongs to the West Indian floristic
region. This is the only part of the United States which has a

176 THE AMERICAN NATURALIST [ Vou. XLVI

real tropical flora. Along the coast there are extensive Man-
grove swamps and the strand flora includes many West Indian
types. Among the strand plants is the ubiquitous Ipomæa pes-
capre found on every tropic beach throughout the world. An-
other very widespread tropical plant occurring in the swamps
along the shore is the stately fern Acrostichum aureum. This
is also a common and conspicuous plant in the Nipa swamps of
the East Indies.

A feature of southern Florida is the presence in the grassy
lands of the everglades and in the dry sandy pine forests, of
areas covered with hardwood trees. These wooded areas are
known as ‘‘hammocks,’’ and the vegetation, both the trees and
lower growing plants, are mainly tropical types, for the most
part West Indian species. There are many epiphytic Orchids
and Bromeliads as well as tropical species of ferns, Palms, Figs,
Mahogany and other West Indian types. In this region is found
the only Cycad (Zamia floridana) occurring in the United

tates.

Proceeding westward, the rich forest vegetation of the Atlan-
tic Gulf states and the eastern Mississippi Valley begins to thin
out with the falling off in the rainfall toward the interior. In
western Michigan, Indiana and Illinois there are already en-
countered open prairies, and the forest is much less luxuriant.
Open groves of Oaks, Hickories and Walnuts, the so-called ‘‘oak ©
openings,’’ are characteristic formations of this region. Fur-
ther west these disappear, and the whole country forms a con-
tinuous open prairie. How far the formation of the prairie is
due to climatic causes, i. e., insufficient rain and cutting winds,
and how much is due to compact soil (loess) which often under-
lies them, and which seems unfavorable for tree growth, is not
quite clear; but throughout the prairie region trees are practi-
cally confined to the beds of the streams and the sheltered gullies
between the hills.

Still further westward the prairie insensibly merges into the
semi-arid plains, which end abruptly at the foot of the grent bar-
rier of the Rocky Mountains.

Compared with the eastern third of the continent the flora of
the great central plains is scanty. Grasses predominate, but
there are a good many showy flowers which adorn the prairie in
the spring and summer. Conspicuous among these are numer-
ous Compositæ, Rosin Weed, Rudbeckia, Sunflowers, Asters,
Golden-rods, and many others.

No. 543] NOTES AND LITERATURE An

The western third of the continent is much more varied in its
topography than the Atlantic side, and shows a correspondingly
greater richness and diversity in its flora, since the climatic dif-
ferences, due to the extremely varied topography, are much
greater than is the case in the eastern states.

Rising abruptly from the plains, the Rocky Mountains form
the beginning of the great complex of mountains and deserts
that stretches to the Pacific Coast. Unlike the worndown Appa-
lachian system, the western mountains are extremely rugged,
and their much greater elevation, rising above the level of per-
petual snow, permits of a true Alpine flora. The region bor-
dering on the great plains is arid and semi-arid, and has a pro-
nounced continental climate, with great extremes of heat and
cold and a scanty rainfall.

The lower elevations are mostly destitute of forests except
along the water courses and in sheltered cafions. Higher up the
precipitation increases, and at elevations of from five to ten
thousand feet a fairly heavy forest is found, in many places
composed mainly of coniferous trees. Of these the most im-
portant are the Rocky Mountain Yellow-Pine, the Douglas Fir,
the Lodge-pole Pine and Engelmann Spruce. At the lower ele-
vations where trees are found they are mostly of Eastern species,
some of which, like the White Elm and Cottonwood, are found
sparingly as far west as the base of the Rockies.

For the most part, however, the Rocky Mountain flora is allied
to that of the Pacifie Coast and the arid southwestern region
whose plants are mainly of Mexican origin. A marked feature
of the eastern Rocky Mountains is the formation of beautiful
glacial parks, which are especially well developed in Colorado.
The floors of these parks are covered with a rich carpet of grasses
and showy flowers, and the mountainsides support a fairly dense
growth of Pines and Spruces with a few deciduous trees like the
Aspens and Birches. To the west of the main range of the
Rockies lies the elevated arid plateau of the Great Basin, inclu-
ding most of Nevada, Utah, northern Arizona and New Mexico.

This region especially the southwestern portion is one of the

most interesting botanical areas of the continent. It is continu-

ous with the Mexican plateau which seems to have been the center
of development for the many interesting and peculiar xerophytic

types which are so remarkably developed in this region. — Amg

these the Cacti take first place, an almost exclusively American

fe Sp i

178 THE AMERICAN NATURALIST [Vou. XLVI

family. Other striking and characteristic forms are Yucca,
Agave, Mesquit (Prosopis), the Creosote-bush (Larrea), the
Ocotilla (Fouquiera) and several other characteristic American
genera. This peculiar xerophytic flora is especially well devel-
oped in southern Arizona, and visitors to Tucson have an excel-
lent opportunity of seeing it in great perfection. i

In the northern part of the Great Basin region the Cacti are
almost entirely absent and the desolate country is covered by
Sagebrush, Greasewood, and many scrubby Composite, produc-
ing a landscape that is monotonous in the extreme.

In the northwestern area, through Canada and the northern
tier of states, the true desert is practically absent, and the moun-
tains are covered with a heavy growth of timber mostly made up
of western Conifers. The other plants are largely subarctic and
northern types, but mingled with these are a good many western
genera, such as Pentstemon, Fritillaria, Gilia, and others.

On the western slope different conditions prevail and the ame-
liorating effects of the west winds from the Pacific make them-
selves felt. This is, however, not fully seen until the last of the
mountain ranges is reached, the great Cordillera, which stretches
practically without a break from Alaska to Patagonia. It. is
this great barrier, averaging in California some ten thousand feet
in height, that largely determines the character of the Pacific
Coast climate. Protected entirely from the sudden changes of
the great interior continental area, with prevailing westerly
winds from the vast Pacific whose temperature scarcely changes
from one year’s end to the other, the whole western coast enjoys
an extraordinarily equable and temperate climate.

On the Pacific Coast topography is a far more important fac-
tor than latitude in determining climate. Close to the coast the
climate is uniformly cool, and seldom either hot or cold. San
Francisco in latitude 37°, with an annual mean temperature of
56° F., shows a range of scarcely ten degrees between the hottest
and the coldest months. Sitka, 20° farther north, has a winter
temperature about the same as that of New York City. Natur-
ally the character of the climate has affected the vegetation pro-
foundly, and with the great differences in elevation and precipi-
tation found along the Pacific Coast, the variety of the vegeta-
tion is much greater than in Atlantic North America.

The coast of Alaska, as far north as about sixty degrees, has
a winter temperature about the same as that of the middle

No. 543] NOTES AND LITERATURE 179

Atlantic states, but with a very heavy precipitation, so that it
supports an extremely dense forest of large coniferous trees, of
which the Tideland or Sitka Spruce (Picea Sitchensis) and the
Hemlock (Tsuga Mertensiana) are the predominant trees. There
is an impenetrable jungle of undershrubs and herbaceous plants
of almost tropical luxuriance.

Further south the Sitka Spruce is largely replaced by the
Douglas Fir, the characteristic tree of the coast region about
Puget Sound and northern Oregon. This tree rivals the Cali-
fornia Redwoods in height though not in bulk. At the lower
levels of the coast in Washington and northern Oregon the
forest over large areas is composed almost exclusively of this
species, but back from the coast at the higher levels this is to
considerable extent replaced by Pines, Firs and Hemlocks. This
change in the forest is very clearly shown on the slopes of Mt.
Rainier.

The undergrowth of this northern forest is very much like
that in Alaska. Among the characteristic species are Rubus Nut-
kanus, Rubus speciosus, several species of Elder, the Vine-leaved
Maple (Acer circinnatum) and the Devil’s-club (Echinopanar
horridum). At the lower levels Alders and Poplars of large size
are common along the streams, and with these are found the Big-
leaved Maple (Acer macrophyllum). Where the forest has been
destroyed, especially in burnt-over areas, the ground is covered
in mid-summer with a sheet of pink flowers of the Fire-weed
(Epilobium angustifolium), Bracken ferns of gigantic size and
the striking Aroid, Lysichiton Kamtchatcense, give a tropical
aspect to the dense undergrowth. Most of the herbaceous plants
are northern types, Linnza, Oxalis, Smilacina, Clintonia, Cornus
canadensis, ete. i

At an elevation of about five thousand feet there begins an
Alpine flora which is especially well developed on the higher
snow-capped mountains of the Cascades and the Canadian
Rockies. On Mt. Rainier, where the line of perpetual snow
descends to but little over six thousand feet, the close turf, left
uncovered for a couple of months after mid-summer, is covered
with masses of brilliant flowers of great variety and beauty.
Among the most striking of these are two heathlike plants,
Bryanthus and Phyllodice, a very large and beautiful Erythro-
nium (E. montanum), a splendid crimson Castilleia, bright blue
Lupins, Veronica, white Valerian, several species of Pedicularis,
Anemone, Mimulus, Gentian, and many others. o

180 THE AMERICAN NATURALIST [ Vou. XLVI

Passing southward from Oregon into California, the outer
coast mountains from the Oregon line to Santa Cruz, some fifty
miles south of San Francisco, are dominated by the Redwood (Se-
quoia sempervirens), the tallest of all trees, which reaches its
greatest development in the northern coast counties of Mendocino
and Humboldt. These giant trees occur only in the narrow belt

within reach of the heavy summer fogs which prevail along the
coast. In most parts of the Redwood belt there is a greater or
less mixture of other trees, e. g., Madroño (Arbutus Menziesi),
the Tan bark Oak (Quercus densiflora) the Douglas Fir, wild
Nutmeg (Torreya californica), the wild Bay tree (Umbellularia)
and the big-leaved Maple.

South of San Francisco there is a rapid diminution in the rain-
fall and the Redwoods gradually disappear and are replaced
by more xerophytie types. Some of these, like the Monterey-Pine
and Cypress and the Torrey Pine of Southern California, are of
very restricted range, forming scattered forests along the coast,
and extending down to the very edge of the ocean.

Many of the plants of the Redwood belt are northern types
which thrive in the cool forests of the outer coast range. Repre-
senting familiar northern and southern genera may be men-
tioned species of violets, Trillium, Clintonia, Oxalis, Smilacina,
Asarum, Aquilegia, Delphinium.

Of characteristic shrubs of this region there may be cited spe-
cies of Ribes, Ceanothus, Gaultheria, Rhododendron, Azalea,
Vaccinium, Rubus and Arctostaphylos. In middle California
these northern plants follow the shady cafions almost to the
level of the valleys, where they mingle more or less with the
valley flora which is comprised mostly of strictly Western types.

In many places along the Coast there is an extensive devel-
opment of sand dunes, which support a very rich and interesting
flora. Characteristic plants of the sand dunes are species of
Abronia, Lupinus, Erigeron, Œnothera, Eschscholtzia, Fragaria,
Mesembryanthemum.

In middle California three ranges of mountains parallel with
the coast are found. The Outer Coast-range skirts the ocean,
and a break in this at San Francisco forms the Golden Gafe,
opening into the Bay of San Francisco, which receives the united
waters of the two rivers draining the great central valley. The :

_ latter is separated from several smaller valleys to the west, the: =

most important being the Santa Clara Valley, by the inner Coast

No. 543] NOTES AND LITERATURE 181

Range which farther south joins the outer Coast Range. The
rainfall in these valleys shut off from the ocean by the inter-
vening mountains is relatively small, and the floors of the
valleys are usually open plains with only a scattered growth of
spreading trees, mostly Oaks. In the great central valley much
of the land is too dry to support any tree growth and there are
extensive grassy plains recalling the prairies of the Middle West.

The open valleys and the foothills are the home of an extra-
ordinary variety of beautiful flowers, which in spring spread a
gorgeous carpet over the landscape. These flowers are largely
annuals which start into growth at the first autumn rains, and
flower in the early spring, ripening their seeds and dying with
the oncoming of the summer drought. There are also many
bulbous plants, some of great beauty like the exquisite Mariposa
lilies and Brodiwas. While some of these valley plants like the
California Buttereup and species of Clover are allied to Eastern
types, the great majority are western or belong to western and
southwestern genera. Representative genera are Lupinus, Or-
thoecarpus, Eschscholtzia, Nemophila, Gilia, Platystemon, Go-
detia, Clarkia, and an extraordinary variety of showy Com-
posite.

A characteristic feature of the hillsides and dry mountain
slopes is the ‘‘Chaparral,’’ a dense and often impenetrable
growth of shrubs of many kinds, scrub Oaks, Chinquapin,
Poison-oak (Rhus), Buckeye, Ceanothus, Ribes, Heteromeles,
Adenostoma. Most of these are evergreen, but some like the
Buckeye and Poison-oak, are deciduous. i
- To the east of the great valley lies the Sierra Nevada, rising to
an altitude of nearly fifteen thousand feet with a flora of great
variety and beauty. The lower slopes covered with chaparral
and with scattered Oaks and Digger Pines (P. Sabiniana) give
way at higher altitudes to perhaps the finest coniferous forests
in the world. At an altitude of about four to six thousand feet —
is the great forest belt where grow the scattered groves of the
Giant Sequoia, associated with hardly less imposing Sugar Pines,
Yellow Pines, Firs and Incense Cedar. As an undergrowth
there are found Dogwoods, Oaks, Aspens, and many beautiful
flowering shrubs like the Azaleas, Ceanothus, Philadelphus. The
low meadows near the water courses are full of showy flowers,
Lupins, Castilleia, Lilies, Veratrum, Monk’s-hood, Larkspur, and
many others. ; oes |

182 THE AMERICAN NATURALIST [ Vou. XLVI

In the chaparral may sometimes be seen the magnificent Wash-
ington-lily and the stately Humboldt-lily, while under the shade
of the great Conifers the vivid crimson Snow-plant (Sareodes)
may often be found.

Southern California is largely a desert region, the transverse
Tehachapi range of mountains shutting it off from the great
central valley to the north. The Mojave desert, south of the
Tehachapi, is an arid plateau whose scanty vegetation is of a
marked xerophytic type. Cacti are abundant and the most
striking plants are the tree Yuccas (Y. arborescens). Other
species of Yucca with magnificent panicles of white flowers also
occur, flowering in the late spring and early summer. Farther
south there are rich valleys opening toward the sea, and less
arid than the tableland to the north, and the Colorado desert to
the southeast. The latter region, which includes Death Valley,
is a desert of the most pronounced type, some of it lying below
sea level, and extremely hot and dry. Opening into the Colo-
rado desert, however, there are small cafions with permanent
streams coming down from the lofty mountains, and supporting
a more or less mesophytic growth of Cottonwoods, Sycamores
and Willows, and in some of the cañons there are imposing
groves of the stately Washington Palm (Washingtonia filifera).
This southern region is really part of the great Sonoran region
of Mexico and its vegetation is essentially the same.

The remarkable range in conditions often within very short
distances, e. g., the eastern and western slopes of the Coast
ranges, results in an extraordinary variation in the vegetation
within a relatively small distance, and perhaps no equal area in
the world, outside the tropics, can show a greater number of spe-
cies than California. Moreover, a very large number of these
species are endemic and there are also very many peculiar genera.
Some of the genera are characterized by a great number of spe-
cies. Thus the genus Ribes (see Harshberger, p. 273) is credited
with forty species against thirteen for the whole region covered
by Britton and Brown’s Manual. Other large genera are Calo-
chortus, Orthocarpus, Pentstemon, Lupinus, Trifolium, Gilia.

Professor L. R. Abrams has kindly furnished the writer with
some data showing the remarkable endemism of many California
genera. Of the forty known species of Ceanothus thirty occur
in California, and twenty-five of these are endemic. Sidalcea
with a total of nineteen species is represented in California by —

No. 543] NOTES AND LITERATURE 183

seventeen, of which thirteen are peculiar. California has some
sixty species of Lupinus, out of a total of about one hundred
described species, and of these forty-five are endemic. More
than half of the known species of Gilia occur in California and
more than half of these (thirty-four out of fifty-seven) are en-
demic. Many other cases could be cited but these will suffice to
show the extraordinary degree of endemism exhibited by the
flora of California.

The remarkable development of coniferous trees in California
is of course famous. Within the state are nearly twice as many
species of coniferous trees as in the whole region east of the
Rocky Mountains. Moreover, a considerable number of these
species are quite peculiar to the state, and often of very limited
range, as for example the Big Tree and the Monterey Pine and
Cypress.

While the flora of Oregon and Washington has much in com-
mon with the Rocky Mountain region, and even with the Atlantic
states, the northern elements are much less prominent in Cali-
fornia, although owing to the southward trend of the mountains,
many northern types reach far southward, especially in the cool,
moist forests of the outer Coast range, where such northern
genera as Trillium, Oxalis, etc., are abundant. There is a cer-
tain number of types which seem to be of Asiatic affinity, e. g.,
Fritillaria, Lysichiton, and Tan-bark Oak, but in the drier re-
gions the plants are for the most part allied to Mexican and
South American forms. Of the latter there are certain species
belonging to Chili and Argentina which also occur in California,
but are absent from the intermediate territory.

An interesting feature of the Pacific flora is the occurrence
of certain genera belonging to the Mediterranean region, e. 9.,
Arbutus, Cupressus, Lavatera, and there are also a few Old
World Ferns and Liverworts which are absent from Eastern
North America. Of the latter we may cite Woodwardia rad-
icans, Equisetum telmateia, Targionia hypophylla. The sur-
vival upon the Pacific Coast of these latter types, which are prob-
ably old ones, is presumably due to climatic reasons, and perhaps
this explanation may apply to the occurrence of the phenogamic
genera as well, a case comparable to the many correspondences
between the flora of eastern Asia and that of Atlantic North
America.

Only a small part of North America lies within the tropics,

*

184 THE AMERICAN NATURALIST [ Vou. XLVI

and much of this is so elevated that the flora is not really a
tropical one. The whole of the interior of Mexico is an elevated
tableland, with a flora closely resembling that of the southern
and southwestern United States.

The eastern coast of Mexico, together with Central America:
and Panama, possesses a tropical flora belonging with that of
South America. In this region Palms, Aroids, Cannaceæ, Bom-
bacæeæ, arborescent Leguminose, and other tropical types
abound, and there is a very rich development of epiphytic
Orchids and Bromeliads.

The extreme southern part of Florida, as we have seen, is the
only part of the United States which properly belongs to this
tropical area, although a good many Southern types extend
much farther north than this. Such Southern types are the
Palmettos, Tillandsia, and other Gulf-state species.

While the eastern coast of Mexico has a humid hot climate with
the typical tropical flora, the western coast is extremely arid
with a scanty xerophytie vegetation.

The West Indian region is treated by Dr. Harshberger as a
distinct province. There is developed here a rich and interesting
flora, allied, as might be expected, to that of the American main-
land, but comprising many peculiar species. The larger islands
are very mountainous and great differences in the vegetation
conditioned by topography may occur within very short dis-
tances. This is well illustrated in Jamaica, where the vegeta-
tion on the north and south sides of the island is very different,
although only forty miles apart. This difference is caused by
the presence of a range of mountains over seven thousand feet
high in the middle of the island. Jamaica is remarkable for the
great development of Ferns, being perhaps the richest area of
its size in the world. It is said that some five hundred species
occur in the island, which has an area of only about four thou-
sand square miles.

Dr. Harshberger’s handsome volume is illustrated with a pro-

fusion of excellent figures, many of them original photographs, — Ms

which add greatly to the value of the book, which must remain

for a long time to come the standard work on the distribution of

the plants of North America.

Dovetas HouGHTON CAMPBELL |

STANFORD UNIVERSITY,
December, 1911

VOL. XLVI, NO. 544

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.
Articles containing research work bearing on the problems of organi evolu-
tion os “ea ne! welcome, and will be given preference in publica
hundrea reprints of contributions are supplied to authors re of charge.
Further siero will be supplied at cost.
Subscriptions and advertisements aee be sent to the gaei a
ou a

yea Foreign postage
Canadian pae Pia aiis cents additional. The charge for dagien copies is

thirty-five ce The advertising rates are Four Dollars for a
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 of

Congress of March 3, 1879.

ical and Pre-historic Specimens. Books on Natural | THE MI CROSCOPE,
History, Sei Mic Methods and to i
History, Science, Travel, Voyages, ete. See THE ae eee Gaon of ¢ san gr

BULLETIN post free for 3 cent stamp. 350 octa and 250 figures in this new
DOr Aa : Š and S recent oditi: Price $2.00, postpaid.
P AA ee Tanini Ena COMSTOCK PUBLISHING CO., Ithaca, N. Y. —

University of Colorado <
Mountain Laboratory

Fourth summer session June 24 to Au-
| gust 2, 1912. Courses in Field DE.
(animal and plant), Ecology, § 2
Botany, Biology e Pam pas Streams.
| Suitable work for undergraduates and grad-
| uates. Daily Wield es

THE
AMERICAN NATURALIST

Vor. XLVI April, 1912 No. 544

THE CONTINUOUS ORIGIN OF CERTAIN UNIT
CHARACTERS AS OBSERVED BY A
PALEONTOLOGIST'

DR. HENRY FAIRFIELD OSBORN

RESEARCH PROFESSOR OF ZOOLOGY, COLUMBIA UNIVERSITY, CURATOR EMERITUS
OF VERTEBRATE PALEONTOLOGY IN THE AMERICAN MUSEUM OF
ATURAL HISTORY, VERTEBRATE PALEONTOLOGIST
UNITED STATES GEOLOGICAL SURVEY

One method of ascertaining the height of a mountain
is with a single instrument, the barometer; another
method is by triangulation with several instruments.
Thus we may differ from Johannsen in his remark that
morphology as a science of great collections in museums
is of no value in genetics. The brilliant progress in
heredity of the last nine years, beginning in 1903 with
the rediscovery of Mendel’s law, should not blind us to
the four broad inductions from paleontology,’ that trans-
formation is a matter of thousands or hundreds of thou-
sands of years, that to the living observer all living
things may be delusively stationary, that invisible tides
of genetic change may be setting in one direction or
another observable only over very long periods of time,

*One of. the Harvey Lectures of 1912, delivered before the Harvey
Society, January 20, 1912, Abstract presented in the Annual Symposium
of the Society of American Naturalists, Princeton, December 28, 1911.

* Osborn, Henry F., ‘ ‘Darwin and Paleontology.’’ One of the addresses
in ‘*Fifty Years of Darwinism.’’ 8vo. Henry Holt & Co., New York,
May 1, 1909. ee

185

186 THE AMERICAN NATURALIST [Vou. XLVI

that discontinuous mutations or saltations may be mere
ripples on the surface of these tides.

Whatever the truth as to this thought, by a strange
paradox it is certain that some stationary characters,
some apparently dead things in the eyes of the zool-
ogist and botanist, become movable and alive in the eyes
of the paleontologist. Thus a paleontologist comes be-
fore the Harvey Society of Physiologists and Physicians
with the conviction that his vision is of a different angle
from that of the experimentalist, and that by the tri-
angulation of experiment, of anatomy and of paleontol-
ogy the truth may at least be more nearly approached.

Is there more evidence of discontinuity and of law-
lessness, or of continuity and of law, in the origin of new
characters? Perhaps no more appropriate question
could be chosen as the subject of a lecture in memory of
William Harvey, the author of the doctrine of epi-
genesis, for the essence of this doctrine is that of
‘‘suecessive differentiation of a relatively homogeneous
rudiment into the parts and structures which are char-
acteristic of the adult.’”’ Paleontology is at one with
embryology in the belief that differentiation is in the
main gradual and continuous.

Yet to our question the answer prevailing among ex-
perimentalists and Mendelians at the present time is that
there is little evidence either for continuity or for law;
this despite the fact that a large part of the evidence for
discontinuity in the origin of characters is most un-
sound. In fact, our first purpose in this Harvey Lecture
is to show how surprisingly unsound this evidence 1s
when we consider that discontinuity has become prac-
tically a dogma among a very large number of zoologists
and botanists. It is true that the evidence for discon-
tinuity in the heredity of characters is as convincing as
that for discontinuity in the genesis of characters is de-
batable. Our second purpose in this Harvey Lecture 1S
to show that the evidence for continuity in the genesis
of certain characters in man and other mammals is very

No. 544] ORIGIN OF UNIT CHARACTERS 187

strong indeed, further, that some of these characters,
while apparently continuous in origin, certainly become
discontinuous in heredity; from which it follows that
discontinuity in heredity constitutes no proof of discon-
tinuity in origin.

The question is, how do these manifold characters of
which the body is made up arise, continuously or discon-
tinuously? A conservative opinion from what may be
gathered in the whole field of observation at the present
time is that while the greater part of evolution is continu-
ous, especially in the origins of certain parts and in the
development of certain proportions, there must also be
a discontinuity, especially in numerical or meristic struc-
tures, such as vertebre and teeth, and in chemical com-
ponents and reactions which are essentially antithetic
or discontinuous. In other words, there is both continu-
ity and ‘discontinuity, and one problem is to show what
is continuous and what is discontinuous.

The history of ‘‘characters’’ is our quest, and now
that attention is concentrated all along the line of obser-
vation in plants and animals, living and fossil, on the
genesis and behavior of single characters, we have laid
the train for substantial progress. A vast gain is that
which relegates the problem of species to a side issue, or
rather to an incidental result of the accumulation of a
greater or less number of units.

A “‘character’’ may be racial shape of head or length
of limb, it may be a cusplet on a grinding tooth, color of
hair, or a sportive white lock of hair, it may be the brown
or blue color of the eye, it may be the speed of a horse,
or the obstinacy of a mule, in short, any structure or
function, simple or extremely complex, which is stable
and distinct in heredity. A ‘‘new character’’ is some-
thing which is unknown before, it may be a new unit,
like the horns of cattle, it may be a new form or propor-
tion of such a unit. ‘‘I understand by the term unit char-
acter,” observes Morgan, ‘‘any particular structure or
function that may appear in heredity independent of

188 THE AMERICAN NATURALIST [ Vou. XLVI

other characters. Such unit characters may in them-
selves be extremely complex and include the possibility
of further splitting up.’’ The point where Mendelism
bears on the problem is, therefore, the continuous or dis-
continuous origin of the thousands of characters which
display this more or less complete discontinuity in hered-
ity.
I. EVIDENCES ror DISCONTINUITY

Darwin has been widely misunderstood of late as be-
lieving in continuity,? whereas he chiefly believed in dis-
continuity. In his original (1859) and final (1872)
opinion evolution is due chiefly to the selection of herit-
able ‘‘individual differences’’; these have been under-
stood by some as ‘‘fluctuations.’’ His true meaning as
to these individual differences is to be found in the cases
he cited, which may be collected from hundreds of ob-
servations in the ‘‘Origin of Species’’ and ‘‘ Variation
of Animals and Plants under Domestication,’’ to the
effect that such individual differences or new characters
were in the nature of minor saltations, structural or
functional, and always hereditary. If we note some of
the observations which he assembled in commenting on
the genesis of the race horse and the grayhound, breeds
which he used by way of illustration of the genesis of
new forms in nature, we find they include such suddenly
appearing new characters as horn rudiments, tailless-
ness, curliness of the hair, characters which are discon-
tinuous in Bateson’s sense, or mutations in that of De
Vries ;> intermingled with these new characters he cited

* Cf. Poulton, E. B., ‘‘ Darwin and the ‘Origin,’ ’’ 1909, pp. 49-50. —
observation and study of Nature led him to the conviction that large varia-
tions, although abundant, were rarely selected, but that evolution proceeded
gradually and by small steps,—that it was ‘continuous’ and not ‘discon-
tinuous.’ °? In answer to this opinion of the most eminent British exponent
of pure Darwinism it may be said that small steps are discontinuities.
(H. F. O. :

*Osborn, H. F., ‘‘Darwin’s Theory of Evolution by the Selection of
Minor Saltations,’’ AMER. NATURALIST, Vol. XLVI, No. 542, 1912, pp- 1e

* In 1909 L. Plate showed clearly that the ‘‘mutations’’ of De Vries are
practically identical with the ‘‘individual differences’’ of Darwin. See
‘*Darwinismus und Landwirthschaft,’’ Berlin, 1909.

No. 544] ORIGIN OF UNIT CHARACTERS 189

others which are obviously reversional. That he be-
lieved in the adding up of minor saltations there can be
no question; but on the admirable ground that no evi-
dence had been adduced in nature of evolution by major
saltations, he rejected St. Hilaire’s hypothesis of the
natural appearance of entirely new types of animals
and plants, or of new or profoundly modified organs;
there was no evidence in 1872 and there is none to-day
of the sudden appearance in nature of such a breed as
the short-legged Ancon sheep. Morgan remarks,’ ‘‘Dar-
win undoubtedly supposed that by the continuous selec-
tion of minor saltations a character could be slowly
shifted in the direction of Selection. This also appears
to be the opinion of the conservative mutationists of the
present day.’’

Aside from his chief emphasis in the selection of ‘‘in-
dividual differences’? Darwin also undoubtedly believed
in the selection of heritable fluctuations of proportion
as illustrated in his classic rebuttal of Lamarck in re-
spect to the long neck of the giraffe:

So under nature with the nascent giraffe, the individuals which were
the highest browsers, and were able during dearths to reach even an
inch or two above the others will often have been preserved; for they
will have roamed over the whole country in search of food.
slight proportional differences will favor survival and will be trans-
mitted to offspring.

If unusual length of neck in the giraffe, as in man, is
a saltatory and heritable character, there is no reason
why this classic case also may not strengthen the opinion
that Darwin was essentially a mutationist. Fluctuations
of proportion, the transmission of which is now in dis-
pute, however, formed a small part of Darwin’s scheme,
nor was fluctuating variability especially connected by
him with the process of evolution.

A very critical reexamination of Darwin’s works
leads us, therefore, to largely dissent from the influen-
tial opinion of De Vries’ that there was always a doubt

* Morgan, T. H., letter, January 11, a ’? Leipzig, 1901 24.
* De Vries, Hugo, ‘‘Die Mutationstheorie,’’ Leipzig, s

190 THE AMERICAN NATURALIST [Vou. XLVI

in Darwin’s mind as to whether ‘‘the selection of muta-
tions” or ‘‘the selection of extreme variants” played
the greater part in the origin of species. As above noted,
the actual cases which Darwin cited and his repeated
emphasis shows that minor saltations of the De Vries
type were chiefly in his mind.

It is obvious that Darwin could not draw such sharp
distinctions either in language or in definition as we may
to-day, profiting by forty years of experiment and of
analysis.

Let us therefore closely examine the kinds of saltation
or discontinuity in mammals which have been recorded
during the last fifty years by Darwin, Bateson and others
and see what they signify.

1. Major and Minor Saltations in Mammals as Supposed
Material for Selection’

The above exposition of Darwin has a very direct bear-
ing on the problem of continuity and discontinuity be-
cause the saltations which he believed to be among the
possible materials of natural selection and of evolution
were chiefly drawn from the very same sources of evi-
dence, namely, hybridization and artificial conditions of
environment, which are now drawn upon by the ad-
herents of discontinuity; the only difference is one of
degree, not of kind. The great saltatory characters of
Darwin cited below (Table I) in mammals are no more
profound than those cited by De Vries as composing the
supposed ‘‘elementary’’ species of (nothera. It is
therefore interesting to compare twenty distinct types or
forms of major and minor saltation in eleven different
types of mammals. Our authorities are Allen, Azara,
Bateson, Brinkerhoff, Castle, Darwin, Davenport,
Haecker, Percival, Poulton, Ridgeway, Root, Seton, Sut-
ton, Twining. The accompanying table presents at once
the very impressive result obtained by this comparison.

*The writer is greatly indebted to Dr. Charles B. Davenport, of the
Carnegie Institution Station for Experimental Evolution, and to Professor

T. H. Morgan, of Columbia University, for criticism and suggestion on
this section.

No. 544] ORIGIN OF UNIT CHARACTERS 191

TABLE I
COMPARATIVE TABLE OF SALTATIONS
1 8| 4/5) 6] 7/8) 9/10\11
a |e
o|2|& n 2a
g/£/3(8|3/8| 2/2 /2\3/2
alal Sajag £ =
©
1. Proopic brachycephaly, abbreviation of face.. x x
2. Sudden development of horns on hornless races} |X x
3. Absence of horns on horned races........... XIX
4. Supernumerary horns on horned races....... x
5. Absence of 1 horn on horned races.........- xX
B Jaw appendages. leo aa eee x x Kii
7. Taillessness, absence of caudals............- XIX x |X x
8. Earlessness, absence of the external ear...... |
9. Single ears, loss of one ear............+--:- | x
10. Short-leggedness, or limb abbreviation ...... XX
11. Consolidation of paired hoofs, syndactylism. . X x
12. Polydactylam -er os eee ee KIXKI XIXIX A x
13. Epidermal thickenings. ............0+---0% x | x
14...Mottled skin markings. ...°006.05 6423152 XIXI X i
15. Excessive hairiness, or length of hair........ XIXI XIX XXIX X
16. Hairlessness, entire absence of hair......... R x x
17. Excessively fine or silky hair...........---- x| |X| x | x
19. Reversed halts 3 re oh ia he eee x x
19. ito hair loGKSs | 4k ee ee ea | |
20. rirletl-RAIr 655s kainic Reed ee XIX TA x
3a. Duplication of horns (tramsverse)........--- ay. eis fe Nom Geel ate

The very uniformity of the result makes us suspicious
as to the significance of saltations, major or minor, in
evolution. In eleven different kinds of mammals, namely,
man, horses, cattle, sheep, deer, pigs, dogs, cats, rabbits,
guinea pigs, mice, we observe that. saltations exactly or
closely similar repeatedly occur. These saltations are of
the same kind, in fact, they partly include those which
were regarded by Darwin as possibly part of the evolu-
tion process through selection, namely, as stable in in-
heritance and as under certain circumstances favoring
the animals which possessed them.

We evidently have to do with abnormal disturbances
of the germinal factors or determiners. Some of these
saltations are very stable in heredity and certain of them
become widespread; some are prepotent and dominant,
others are recessive (e. g., angora, Or ‘long coat”? in
rabbits, Castle); some (e. g., bent tail in certain mice,

192 THE AMERICAN NATURALIST [Vou. XLVI

Plate) follow neither the Mendelian law nor the prin-
ciple of blended inheritance.

On the unit-character doctrine we know that one of
three things is happening in the germ plasm.

First, a ‘‘determiner’’ may drop out and we see a race
of mammals springing up without tails, or color, or hair.
In cattle the determiner for horns is dominant, therefore
something is added.

Second, a ‘‘determiner’’ may be suddenly lost or modi-
fied, and we see excessive hair, curly hair, silky hair,
dwarfed or short limbs, brachydactylism. ,

Third, and even more inexplicable, there occurs the
appearance of a new ‘‘determiner’’ or the removal of an
‘‘inhibitor’’ and we observe horns suddenly arising on
hornless races like horses and rabbits.

That fancy breeds can be established through the ab-
normal behavior and selection of these ‘‘determiners”’
there is no question. That nature works through the
sudden appearance of new and favorable ‘‘determiners’’
is as yet unproved; it is absolutely disproved in the case
of horns, for through paleontology we know that horns
arise in a continuous manner. The only mammal known
to us at present in which it would appear that a duplicate
horn may have sprung into existence through saltation
is T'etraceros, the four-horned antelope of India. Salta-
tion is possibly of significance in the case of the sudden
alteration of hair character because we know of a very
considerable number of curly-haired horses in Mexico
and South America, which are, however, eliminated by
breeders for the reason that correlated with curliness of
the hair are apt to arise certain other characters in the
hoofs and limbs which are unfavorable.

Under wild or natural conditions in mammals we have
as yet secured no direct evidence of such origins or estab-
lishment of saltations either major or minor. There is
reason to believe that peculiar or anomalous mammals if
they do arise are driven away from the herds.

It would appear that the obvious abnormality of the

No. 544] ORIGIN OF UNIT CHARACTERS 193

majority of these characters throws the remainder as
well as saltatory new characters in general under sus-
picion of abnormality.

Paleontology, however, furnishes the most direct evi-
dence of the abnormality of saltations in some of the
hard parts shown in Table I by presenting counter evi-
dence that such profound changes as abbreviation of the
face (proopie brachycephaly), development or loss of
horns, reduction or absence of caudal vertebra, abbrevia-
tion or elongation of the limbs, syndactylism or consoli-
dation of separate metapodials have all been established,
wherever we know their history, through continuity and
not through discontinuity.

2. Bateson’s Evidence (1894) of Discontinuity

Bateson in 1894° was the first to clearly advance the
discontinuity hypothesis as a mode of origin of species
in its modern form. At the time this work appeared it
suffered a searching review from Scott.1° Mutationists,"!
however, still refer to it as laying the foundations for the
discontinuity hypothesis. In order to test the ‘‘ Materials
for the Study of Variation” critically in the light of the
subsequent advance in paleontology, Dr. W. D. Matthew,
who is without bias in the question, was requested to
examine all the cases of discontinuity in mammals cited
by the author with reference to the question whether or
not these cases have any real significance in evolution.
He reports:

Of the 320 cases of discontinuity cited in mammals the greater part
are obviously teratological and have no direct significance in relation
to paleontologie evolution except for a very few instances such as the
supernumerary or fourth molar teeth of Otocyon. While not signifi-
cant [in evolution] these teratological cases are interesting because
they show the prevalence of homeosis, and indicate that many of the

? Bateson, Wm., ‘‘Materials for the Study of Variation Treated with
Especial ogari to Discontinuity in the Origin of Species,’’ Maemillan

& Co., London, 1894.
” Scott, W. B., ‘‘On Variations and Mutations,’’ Amer. Jour. Science

(3), Vol. XLVIII, 1894, pp. 355-374 E
“Darbishire, A. D., ‘‘ Breeding end the Mendelian Discovery, 8vo,

gd

Cassell & Co., London, 1911.

194 THE AMERICAN NATURALIST [ Vou. XLVI

remaining eases which might [otherwise] be considered normal salta-
tions or reversions may actually be teratologic, but disguised by
homeeosis; all of the possibly significant cases (such as the supernu-
merary molars) are thereby placed under suspicion. Setting aside this
suspicion the minority of the “ significant ” cases in teeth and feet may
be said to afford evidence of the meristic variability of vestigial and
rudimentary structures. Bateson’s statement that such variability is
related not to non-functionalism but to terminal position in a series
appears to me directly in conflict with his [Bateson’s] own evidence,
as it certainly is with all my experience. This accords with commonly
observed data in paleontology, for no paleontologist would question
that vestigial teeth or bones are apt to [finally] disappear by “ discon-
tinuous” evolution. As to the appearance by saltatory evolution of
new and primarily functional parts in teeth or feet, I know of no ade-
quate paleontologic evidence in its favor. It is either demonstrably
false or decidedly improbable. In the cases of supernumerary teeth
(Otocyon myrmecobius, Cetacea, etc.) saltatory evolution may be re-
garded as reasonable in default of any paleontologic evidence to the
contrary. Meristic or numerical evolution in fully functional verte-
bre is intrinsically probable as the only method of evolutionary
change.

The fact that so many cases of supernumerary teeth are associated
with asymmetry throws doubt on the significance of all such cases;
asymmetrie variations and those occurring only in upper or only in
lower teeth have no analogy in paleontology; such cases as occur ab-
normally are recognized as of a different and non-significant class than
normal evolutionary changes.

A summary of Matthew’s report is as follows:

Bateson cites 323 cases of discontinuity in vertebre,
teeth and skull. Of these 286 are abnormal, or teratolog-
ical, or reversional, and have absolutely no significance
in evolution; ten cases of supernumerary (or fourth
molar) teeth are possibly significant because among the
mammals there are a few genera with fourth molars
which may possibly have arisen by saltation. There re-
main only thirty-seven cases which may be ranked as
‘probably significant,” and these are the meristic addi-
tions or reductions of vertebre in the spinal column,
significant because of the well-known variations in the
vertebral formule of different mammals, and secondly
because vertebre can be added or subtracted only dis-
continuously.

—
ce
on

No. 544] ' ORIGIN OF UNIT CHARACTERS

SUMMARY OF BATESON’s 323 CASES

E E
2 oa $9
S & A
3 3 55

[r9] = =

Z £ S

E Svoo a a 17 27”
(asymmetry)
L Toh.. A 83 107 67
E OBE Gun ou, Gao yale py cea eee « 110

210 37 67

* Numerical variations of cervical, dorsal and lumbar vertebra.
Additional molars, ef. Octocyon, Myrmecobius, Cetacea. Six cases insuf-
ficiently described.

The fact that the vast majority of germinal anomalies
examined in the above review of Darwin and of Bateson
have no significance in evolution in a state of nature,
throws all germinal anomalies under suspicion as natural
processes, important as they may be in artificial breed-
ing and hybridizing. Yet some of these anomalies in
mammals are less profoundly discontinuous than those
which De Vries has cited in plants under the designation
of ‘‘mutations.’? The most important of these De Vries’
mutations may now be considered.

3. Evidence for De Vries’s Mutation Theory

In 1901 the biological world was aroused as it had not
been since 1859 by the publication of De Vries’s hypoth-
esis.2 Here was a new and apparently sure foundation
for discontinuity in the supposed sudden appearance of
elementary species or ‘‘mutants’’ arising with the acqui-
sition of entirely new characters, new forms of plants or
animals quite free from their ancestors and not linked to
them by intermediates. The influence and vitality of this
great work is shown in a citation from Darbishire (1911,
op. cit., p. 5):

The view that species have originated by mutation is based on Prof.
de Vries’ observations on the Evening Primrose (nothera Lamarck-
iana) (Fig. 1). Working with this form, he was able to witness, for
the first time, the actual process of the origin of new species. —

“De Vries, Hugo, ‘‘Die Mutationstheorie,’’ Leipzig, 1901, p. 24.

196 THE AMERICAN NATURALIST [ Vou. XLVI

- Critical analysis during the past two years by Davis
and by Gates'® of the very species Œnothera Lamarck-
iana on which De Vries chiefly based his monumental
work, tends to show that O. Lamarckiana is possibly a
hybrid of O. biennis and.O. grandiflora and not a natural
species. Thus the ‘‘elementary species’? which are
springing from it in various gardens may prove to be
comparable to the familiar results of hybridization in
mammals and birds.

Davis, on the basis of his prolonged experimental
researches, says:

Indeed, the theory of De Vries may fairly be said to rest chiefly
upon the behavior of this interesting plant, the account of which forms
so large a part of his work “ Die Mutationstheorie ” (2 vols., Leipzig,
(1901). ... In a brief perusal of the work one is struck by the opti-
mism of its author and the brilliancy and breadth of his exposition of
the views set forth. . . . The analysis of the data amassed by Darwin,
in which it is shown that Darwin’s “ single variations ” are the same as
De Vries’s mutations seems to the reviewer particularly effective. .. .
Probably the time will soon come when nearly all biologists will be
ready to admit that mutation or the sudden appearance of new forms
has been an important factor at least in species formation of plants
and animals. Admitting this it remains to be discovered what rela-
tion these sudden appearances bear to the general trends of evolution
which are apparent in so many phylogenies [italics our own] .
for granting the facts of mutation we have only accounted for a miero-
evolution, and it is still to be shown that the larger tendencies can be
sufficiently accounted for by the same means without the intervention
of other factors. . . .

The skepticism of both these botanists is striking.
Their opinions as to the existence of larger evolutionary
trends are exactly in accord with those of paleontologists.

4. Evidence for Discontinuity from Mendelian Heredity
and Experimental Selection

The newest bulwark of the discontinuity hypothesis

is that erected since 1903 by the revival of the great dis-

* Davis, Bradley Moore, ‘‘Genetical Studies on CEnothera. II. Some

Hybrids of Gnothera biennis and O. grandiflora that resemble O. Lamarck-

iana,’’? AMER. NATURALIST, Vol. XLV, April, 1911, pp. 193-233. Gates,

R. R., ‘‘ Mutation in (Enothera,’’ AMER. NaTuRALIST, Vol. XLV, No. 538,
October, 1911, pp. 577-606.

No. 544] ORIGIN OF UNIT CHARACTERS 197

covery of Mendel (1865) and by the negative results of
experiments on fluctuating or quantitative variation.

From the prevalence of discontinuity in heredity, the
separateness of ‘‘unit characters’’ as they appear in the
body and the equally sharp separableness of their com-
plex of ‘‘factors,’’ ‘determiners’? or ‘‘genes’’ in the
germ has arisen the theoretical assumption of the dis-
continuity of origin of all characters in the germ. We
shall now show that this assumption is a non-sequitur.

First, however, the truly marvelous and epoch-making
Mendelian discoveries require our especial examination
in their bearing on the problem of continuity and discon-
tinuity. We have reviewed" the contributions of Allen,
Bateson, Castle, Cannon, Cuénot, Darbishire, Davenport,
Durham, Farrabee, v. Guita, Haacke, Hagedoorn, Har-
mon, Hurst, Laughlin, Morgan, Pearson, Plate, Punnett,
and Rosenoff. This review covers unit characters only
as observed in mammals, to which none the less the prin-
ciples discovered by Mendel in the common garden pea
(Pisum sativum) apply with striking uniformity.

The prevailing field of the researches of these talented
investigators in mammals has been in color characters,
chemical in essence, in various species of rodents, chiefly
mice and guinea pigs, also in Ungulates, such as horses
and cattle, the latter studied less by experiment than from
stud books. Hair form in rodents and in man and skin
pigment have also been exactly investigated. The most
striking general result is the principle of antithesis of
characters which mutually exclude each other, as typified
by the antithesis of Mendel’s ‘‘tallness’’ and ‘‘short-
ness’’ in peas.

The second great feature is that when these antithetic
characters meet in the germ cells, one dominates over the
other; this dominance is a sort of perpetual prepotency.
‘‘Prepotency,’? observes Darbishire, ‘‘is an attribute of
individuals and capricious in its appearance... . What-
ever be the nature of this power . . . it is clear that it

“ With the aid of Miss Mary M. Sturgess, now attached to the Carnegie
Institution Station for Experimental Evolution at Cold Spring Harbor, L. I.

198 THE AMERICAN NATURALIST [Vou. XLVI

has nothing to do with dominance . . . dominance is an
invariable attribute of particular characteristics. ”’!5 Plate
(1910), on the contrary, observes, ‘‘ But a variety of facts
seem to indicate that a reversal of dominance may occur
under certain circumstances and a dominant character
may become recessive, and vice versa.’"® Such reversal
of dominance would appear to be the case in a compari-
son of the mule (cross between ass ¢ and horse 2) and
the hinny (cross between the horse ¢ and the ass 9).

When antithetic characters or functions meet in hered-
ity, there is either ‘‘prepotency,’’ or ‘‘dominance,’’ or
‘‘recession’’ (i. e., latency), or ‘‘inhibition,’’ a something
which indirectly prevents the appearance of characters,
or ‘‘imperfect dominance,’’ or ‘‘blending.’’ In brief,
there are degrees of separableness and antithesis.

Dominance, Conservative or Progressive.—It will be
seen at once that progressive evolution through discon-
tinuity would depend on the dominance of racially new
characters and types. The experimental evidence is con-
flicting, it does not show that new characters are neces-
sarily dominant.

There are many instances of dominance of wild species
(older type) over domesticated species (newer type);
thus De Vries suggested (1902) that the dominant char-
acters are those which are racially older. One case among
the mammals is that the wild gray color in mice domi-
nates over grades below it, black, brown, and white
(Plate, 1910).

Examples of dominance in single characters are that
more intense dominate over less intense colors (Plate,
1910, Davenport, 1907); in the eyes, brown over gray,
gray over blue; in the skin, brunettes over blondes (Dav-
enport, 1909), piebalds over pure albinos (Plate, 1910).
In the hair, wavy or spiral forms dominate over straight
(Davenport, 1908). This would have some bearing onthe
discontinuous fading out of color in desert races like the
quagga, which lost all the stripes of its relative the zebra.

* Darbishire, op. cit., p. 96.
1 Plate

No. 544] ORIGIN OF UNIT CHARACTERS 199

The idea that the positive or present character domi-
nates over the negative, latent or absent character has
become the prevailing one.

It seems highly probable, observes Davenport (1910), that the fu-
ture will show that many more advanced or progressive conditions are
really due to one or more unit-characters not present in the less ad-
vanced condition. In that case it will appear that there is a perfect
accord in the two statements that the progressive and the “ present ”
factor are dominant (pp. 89-90) . . . the specific characteristics are
mostly those that appear late in ontogeny (p. 86) .. . the potency
of a character may be defined as the capacity of its germinal deter-
miner to complete its entire ontogeny. If we think of every character
as being represented in the germ by a determiner, then we must recog-
nize the fact that this determiner may sometimes develop fully, some-
times imperfectly and sometimes not at all [italics our own]...
When such a failure occurs in such a normal strain a sport reaillte.

Potency is variable. Even in a pure strain a determiner does
not always develop fully and this is an important cause of individual
variability (Davenport, 1910, p. ;

Plate similarly favors the hypothesis of dominance of
newer or progressive characters. He observes (1910):

The [Mendelian] laws of inheritance favor progressive evolution in
two ways, for .. . higher, more complicated characters are generally
dominant to the tower, and . . . qualitative characters usually follow
the Mendelian principle in the ease of closely related forms (races,
varieties) while in the crossing of species they follow intermediate [or
blended] inheritance as a rule. In the latter case there is the possibility
that the crossing may have a swamping effect, but this can play no
large rôle on account of the infrequeney of hybrids between species
(Plate, 1910, p. 606).

The same author is of the opinion that phyletic evolu-
tion is discontinuous as regards the transformations of
the determinants [determiners], but in most cases is con-
tinuous in their visible outward workings. He thus main-
tains that while germinal transformations are discon-
tinuous there may be no real antithesis between con-
tinuous and discontinuous somatic variation.

Thus Mendelians appear to agree, first, that there are
grades of continuity and discontinuity, that there are —
antithetic characters which are sharply discontinuous,
others which are ihe continuous, blended or intermedi-

200 THE AMERICAN NATURALIST [ Vou. XLVI

ate. Second, it would appear that complete discontinuity
or entire dominance or recession are qualities in heredity
which may gradually evolve. Many characters show im-
perfect dominance (Castle, 1905) ; gametic purity is not
absolute (Castle, 1906) ; selection is of importance in the
improvement of races (Castle, 1907). . There are a num-
ber of truly blending characters, such as lop-earedness
in rabbits (Castle, 1909), cross blends of long and short
hairs (Castle, 1906), cross blends between short- and lop-
eared rabbits which are permanent (Castle, 1909), blends
in weight inheritance and in skeletal proportion (Castle,

More recent work has tended to show (Hatai, 1911)**
that blended inheritance may be considered to be a lim-
ited case of alternative inheritance where dominance is
imperfect. Thus Mendel’s law of alternative inheritance
may be considered as the standard in all the cases re-
ferred to it (Hatai, 1911, p. 106). Certain characters
which were considered formerly to blend are now re-
garded as showing a certain kind of segregation or unit
inheritance. Thus Davenport (1909) observes:

Skin pigment does not show thorough blending inheritance, but segre-
gation (sometimes imperfect), a more pigmented being imperfectly
dominant over a less... . The reason, the same author observes (1909,
for the blending of hair and skin color in man is the non-development
of distinct color unit-determiners owing to the fact that in man for a
long period there has been no selection for intensity of color, whereas
in the lower mammals definite color determiners have long been main-
tained by selection.

Thus the prevalent recent opinion among Mendelian
observers is that there is a real discontinuity between the
germinal or blastic characters and what the paleontol-
ogist or morphologist generally observes, is only an ap-
parent continuity between somatic characters.

polices Shinkishi, ‘‘The Mendelian Ratio and Blended Inheritance,’’

MER. NATURALIST, Vol. XLV, No. 530, February, ii pp. 99-

* Aaa of the skin seems to depend in man on a series of color

intensity units, possibly one or a few large units, more probably a number

of small units so close together as to be almost continuous (Davenport,
1910).

No. 544] ORIGIN OF UNIT CHARACTERS 201

Since, however, the behavior of somatic characters
forms our only means of knowing whether the deter-
miners are continuous or discontinuous, it is obvious
that this opinion requires further examination in the con-
ceptions of Johannsen.

5. Johannsen’s Pure Line Theory’?

The theoretic contrast between the real discontinuity
of the blastic determiners and the delusive continuity
of visible or somatic form is pushed to its extreme in
the ‘‘pure-line’’ conception which marks the latest de-
velopment in heredity, an advance upon Weismann’s
germ-plasm theory and Mendel’s unit-character law.
Through experiments on successive generations of self-
fertilizing plants (the garden bean), Johannsen has
reached a standpoint which may be briefly stated as
follows:

A “pure line” is composed of the descendants of one pure strain
‘or homozygotie organism exclusively propagated by self fertilization;
such pure lines demonstrate the stability of hereditary constitution in
successive generations where undisturbed by cross breeding or ming-
ling with other strains, showing that the only real changes in organ-
isms are those due to the sudden appearance of new determiners in the
germ.

To replace the word determiner the term gene is proposed. The
genotype represents the sum total of all the genes in the fertilized
germ cell, gamete or zygote: we do not know a genotype but we are
` able in experiment to demonstrate “ genotypical differences.” The
biotype is a group of similar genotypes or pure strain individuals.

Gene, genotype, and biotype are not seen; they are the smaller and

larger units of heredity.
. The phenotype is what we see; it is the developing organism.
Morphology supported by the huge collections of the museums has
‘operated with “phenotypes” in phylogenetic speculation. It is thus
a science of phenotypes and is not of value in genetics because pheno-
type description is inadequate as the starting point for genetic in-
quiries. The adaptation of phenotypes through the direct influence of
environment [Buffon’s factor] or of use and disuse [Lamarck’s fac-
tor] is not of genetic importance. Ontogenesis is a function of the
12 Johannsen, W., ‘‘The Genotype Conception of Heredity,’’ AMER.
Naruratist, Vol. XLV, No. 531, March, 1911, pp. 129-159.

202 THE AMERICAN NATURALIST [ Vou. XLVI

genotype, but the genotype is not a function of ontogenesis. The idea
of evolution by continuous transitions from one type to another has
imposed itself upon zoologists and botanists, who are examining chiefly
shifting phenotypes in very fine gradations. There is such a continu-
ity in phenotypes but not in the genotypes from which they spring.
All degrees of continuity between phenotypes may be found, but real
genetic transitions must be distinguished from the transitions which we
find in museums.

Genotypes, it is true, can only be examined by the qualities and re-
actions of the phenotypes.

Such examination shows that within pure lines—if no new muta-
tions or other disturbances have been at work—there are no geno-
typical differences in the characters under examination. The only real
discontinuity is that between different genotypes. The mutations ob-
served in nature have shown themselves as considerable discontinuous
saltations. There is no evidence for the view that mutations are prac-
tically identical with continuous evolution. In pure lines no influence
of special ancestry can be traced; all series of progeny keep the geno-
type unchanged through long generations. Discontinuity between
genotypes and constant differences between the genes show a beautiful
harmony between Mendelism and pure line work.

Selection will have no hereditary influence in changing genotypes.
Even the selection of fluctuations in pure lines is ineffective to produce
a new genotype

Heredity may thus be defined as the presence of identical genes in
ancestors and descendants, or heredity stands for those properties of
the germ cells that find expression in the developing and developed

phenotype.

Similarly Jennings observes:” What distinguishes the different
genotypes, then, is a different method of responding to the environ-
ment. And this is a type of what heredity is; an organism’s heredity -
is its method of responding to the environmental conditions [p. 84].

. It appeared clear, and still appears clear, that a very large share
of the apparent progressive action of Selection has really consisted
in the sorting over of preexisting types, so that it has by no means the
theoretical significance that had been given to it [p. 88]... . I had
hoped to accomplish this myself, but after strenuous, long-continued and
hopeful efforts, I have not yet succeeded in seeing Selection effective
in producing a new genotype. This failure to discover Selection re-
sulting in progress came to me as a painful surprise, for like Pearson
I find it impossible to construct for myself a “ philosophical scheme of
evolution,” without the results of Selection and I would like to see
what I believe must occur [pp. 88-89]. . . . It would seem that the
diverse genotypes must have arisen from one, in some way, and when
we find out how this happens, then such Selection between m

No. 544] ORIGIN OF UNIT CHARACTERS 203

will be all the Selection that we require for our evolutionary progress
[p. 89]

Thus Johannsen’s general conception of the origin of
progressive or retrogressive new characters is that SIS
is sufficient to state that the essential point in evolution
is the alteration, loss or gain of the genes or constit-
uents of the genotype . . . all evidences as to ‘muta-
tions’ point to the discontinuity of the changes in ques-
tion.’

6. Negative Results of Experiments on Quantitative
Variation

We agree with Johannsen that a delusive appearance
of continuity might arise through selection of degrees
of hereditary fluctuation in structure or function, for ex-
ample, of tallness or shortness of stature, of intensity or
faintness of color. Some Mendelians discard fluctua-
tions altogether as non-hereditary; thus Punnett (1911,
p. 138) observes: ‘‘At the present time we have no
valid reason for supposing that they [fluctuations] are
ever inherited.’’

The question, however, is not as to quantitative onto-
genic variations caused by favorable or unfavorable en-

‘vironment or by changes of habit, but as to heritable
fluctuations springing from the germ plasm. Experi-
ments have been directed to the point whether variations
in size, in proportion, etc., of unit characters as distin-
guished from the unit characters themselves are trans-
mitted.

Davenport has reached negative results; he observes
(1910) :

In the last few decades the view has been widespread that char-
acters can be built up from perhaps nothing at all by selecting in each
generation the merely quantitative variation that goes farthest in the
desired direction. The conclusion upon which De Vries laid the great-
est stress, that quantitative and qualitative characters differ funda-
mentally in their heritability, is supported by our experiments (p. 96).
I have made two tests of this view using the plumage color of poultry

» Punnett, R. C., ‘‘Mendelism,’’ Macmillan Co., 1911 (3d edition).

204 THE AMERICAN NATURALIST [ Vou. XLVI

(p. 94). ... After three years of selection of the reddest offspring no
appreciable increase of the red was observed except in one case, which
looks like a sport (p. 96). These fluctuating quantitative conditions
depend on variations in the point at which the ontogeny of the char-
acter is stopped; and the stopping point is in turn often if not usually
determined by external conditions which favor or restrict the ontogeny.
Thus the selection of redness of comb, of polydactylism, of syndactyl-
ism, have not proved the inheritance of quantitative variations. Ap-
parently, within limits, these quantitative variations have so exclusively
an ontogenie signification that they are not reproduced so long, at
least, as environmental conditions are not allowed to vary widely.

Similarly Love?! from experiments on the yielding
power of plants remarks:

Unless further studies produce different results we can say from the
facts at hand that there is no evidence to show that a basis exists for
cumulative selection.

Similar conclusions have been reached by Pearl
(1909)?? in the breeding of fowls for laying purposes.

All the above results are negative.

Even the positive or affirmative results obtained by
Cuénot and later by Castle, wherein quantitative char-
acters may be shifted in one direction or the other by
selection are now given a new interpretation by certain
Mendelians. For example, Cuénot showed by continued
selection of lighter colored mice that the coat became
paler; and Castle has shown that in rats the coat through
selection may be made darker. Castle remarks (1911) :**
_I prefer to think with Darwin that selection . . . ean heap up
quantitative variations until they reach a sum total otherwise unattain-
able, and that it thus becomes creative.

He cites cumulative results in the development of a
fourth toe in the hind foot of guinea-pigs and in the
modification of the dorsal striping of hooded rats.

2 Love, Harry H., ‘‘Are Fluctuations Inherited?’’ Contr. VI, Lab.
Experim.. Plant-Breeding, Cornell Univ., AMER. NATURALIST, You XLI IV,
No. 523, Tuy, 1910, pp. 412—423..

™” Pea ond, ‘‘Is there a ee Effect of Selection?’ é
foes ad ERARE 2, 1909, H, 4.

*% Castle, W. E., ‘‘The Nature of Unit aui The Harvey Lec-
tures, delivered andes the Auspices of the Harvey Society of New York,
8vo, J. B. Lippincott Co., pp. 90-101.

No. 544] ORIGIN OF UNIT CHARACTERS 205

Morgan’s remarks (1912) on these positive experi-
ments are as follows:

Castle has been very guarded in regard to the interpretation of: the
results of selection in this case. It is probable that extreme selection
is necessary to maintain the higher stage reached. It does not breed
true and slips back easily. If this is correct it suggests: first, that
nothing permanent has been effected in the germ-cells; and second,
that the result is due to the discovery of more extreme eases of fluctu-
ating variations than ordinarily occur.

The general import of these experiments and opin-
ions is that fluctuations in the determiners, or genes, can
not be utilized to establish a new quantitative mean. It
is obvious that what have been measured by biometri-
cians as hereditary ‘‘fluctuations’’ might be regarded
as ‘‘saltations’’ of all degrees, but such saltations do
not represent new determiners in the Mendelian or
Johannsen sense; they are mere fluctuations in existing
determiners. Pure Mendelians would allege that tall-
ness in man or other mammals can only be accumulated
through the saltatory origin of ‘‘tall’’ determiners which
are not connected continuously through intermediate
forms with the antithetic ‘‘short’’ determiners. As to
Stature Brownlee observes (1911, p. 255) :**

I think that I have shown that there is nothing necessarily antagon-
istie between the evidence advanced by the biometricians and the Men-
delian theory. . . . (1) If the inheritance of stature depends upon
a Mendelian mechanism, then the distribution of the population as
regards height will be that which is actually found, namely, a distri-
bution closely represented by the normal curve.

6. Summary as to Discontinuity and Mendelism

Genetics is the most positive, permanent and trium-
phant branch of modern biology. Its contributions to
heredity are epoch-making. But heredity is the conserv-
ative aspect of biology, and experimental genetics thus
far reveals the laws of conservation.

* Brownlee, J., ‘<The Inheritance of Complex Growth Forms, such as
Stature, on Mendel’s Theory,’’ Proc. Roy. Soc. Edinburgh, Vol. XXXI,
PE D. 1911, pp. 251-256.

206 THE AMERICAN NATURALIST [ Vou. XLVI

Genetics has not yet brought us one single step nearer
the solution of the problem of the progressive origin of
new characters in mammals.. The very independence,
multiplicity and discontinuity of the units leave us
farther afield. In place of what used to be regarded as
the instability of the organisms, as a whole, we now have
to conceive of the instability of thousands, nay hundreds
of thousands of units.

As shown in our analysis of the saltations cited by
Darwin and Bateson, Mendelism has revealed the fact
that the majority of saltations simply reflect failures in
the germinal mechanism. The inference is natural that
the remaining minority also represent anomalies, or law-
less conditions. Over a half century of anatomical re-
search among mammals has failed to demonstrate in a
state of nature the sudden origin of a single new pro-
gressive character which has become fixed in the race.

Nor have Mendelism and experimentalism released us
from the hard confines of examination of the germ
through the soma; behavior of unit characters in the
soma is the sole means of knowing the behavior of the
‘‘determiners’’ in the germ. If the unit characters in
the soma behave discontinuously we are forced to the
conclusion that their determiners behave discontinu-
ously; if, on the contrary, these unit characters behave
continuously, are we not forced to the conclusion that
there is a continuity in the behavior of the correspond-
ing determiners?

Let us therefore proceed to consider the value of some
of the evidence for continuous behavior in the origin of
certain new characters, again repeating our opinion that
certain other characters are antithetic, without interme-
diates, and consequently discontinuous both in heredity
and in origin.

(To be concluded*

THE BOTANICAL SOCIETY OF AMERICA

At the Washington meeting of the Botanical Society of
America, the following papers constituting a symposium
on Modern Aspects of Paleobotany were read by invita-
tion of the council, on December 28th.

I. “The Relations of Paleobotany to Geology.”
F. H. Know. Ton.

II. ‘‘The Relations of Paleobotany to Botany.”
1. ‘Phylogeny and Taxonomy.’’ Joun M. CoULTER.
2. ‘‘Morphology.’’ Epwarp (©, JEFFREY.
3. OOO. oe. ArtHur HOLLICK.

In accordance with the instructions of the Society these
papers are here printed in full.
‘Grorce T. Moore,
Secretary

MODERN ASPECTS OF PALEOBOTANY
I. Tue Revations oF PALEOBOTANY TO GEOLOGY `

DR. F. H. KNOWLTON
UNITED STATES GEOLOGICAL SURVEY

AurHovucH there is vague mention of fossil plants in
literature as early as the thirteenth century, and unscien-
tific adumbrations in the faintly growing twilight of the
succeeding centuries, the real science of paleobotany did
not have its beginning until well on in the nineteenth cen-
tury. With the publication, in 1828, of Brongniart’s
“Histoire des végétaux fossiles”? and the ‘‘Prodrome,’’
there was given to paleobotany ‘‘that powerful impetus
which found its immediate recognition and called into its
service a large corps of colaborers with Brongniart, rap-
idly multiplying its literature and increasing the amount

207 .

208 THE AMERICAN NATURALIST [ Vou. XLVI

of material for its further study.’’ Ward. In the suc-
ceeding decades, even to the close of the century, the
students of paleobotany were mainly occupied in accumu-
lating data as regards distribution, both areal and
vertical, and the opening decades of the present century
find the subject a recognized, respected, coequal part of
the general field of paleontology.

Paleobotany, together with all the other branches of
paleontology, admits of subdivision into two lines or fields
of study—the biological and the geological—depending
upon the prominence given to the one or the other of
these phases of the subject. The biological study is, of
course, concerned especially with the evolution of the
vegetable kingdom, that is, with the tracing of the lines
of descent through which the living flora has been devel-
oped. As this side of the question will be taken up by
other contributors to this discussion, it may be dismissed
from further consideration, as the geological aspect is
almost exclusively the phase of the subject to which the
present paper is devoted.

In the first place it will be necessary to call attention
to the fact that the successful use of fossils of any kind
as stratigraphic marks is—or at least may be—entirely
independent of their correct biological interpretation.
To most botanists, and indeed to some paleobotanists,
this statement will doubtless come as a surprise, since
they have come to imagine that the impressions of plants,
the form in which they are most made use of in this con-
nection, are so indefinite, indistinct and unreliable that
they can not be allocated biologically with even reason-
able certainty, and hence are of little or no value. Asa
matter of fact hardly anything could be further from the
truth, and it can be confidently stated that it makes not
the slightest difference to the stratigraphic geologist
whether the fossils upon which he most relies are named
at all, so long as the horizon whence they come is known
and they are clearly defined and capable of recognition
under any and all conditions. They might almost as well

No. 544] BOTANICAL SOCIETY OF AMERICA 209°

be referred to by number as by name, so long as they fill
the requirements above demanded, though of course every
stratigraphic paleontologist seeks to interpret to the very
best of his knowledge the fossils he studies. He may—
doubtless often does—make mistakes in his attempts to
understand them, but his errors are undoubtedly fewer
than he is not infrequently charged with! His faculty
of observation is rendered acute from the close study of
the restricted and often fragmentary material available,
and he has learned to see and make use of characters
which are often overlooked or wholly neglected by the
botanist. The latter, even when he has before him the
complete living plant, including root, stem and foliar and
reproductive organs, sometimes experiences difficulty in
correctly placing his subject, and, to judge from some
recent work, there are paleobotanists who study only the
internal structure of fossil plants and yet are beset with
extreme difficulty in interpreting their biological sig-
nificance.

It may then be taken as settled that the needs of the
stratigraphic geologist will be met if he is supplied with
a series of marks or tokens by which he may unfailingly
identify the various geological horizons with which he
deals, while to the historical geologist who makes use of
fossils in unraveling the succession of geological events,
the correct biological identification is of the greatest im-
portance, for upon this rests his interpretation of the
succession of faunas and floras that have inhabited the
globe. As the late Dr. C. A. White has said: ‘‘If fossils
were to be treated only as mere tokens of the respective
formations in which they are found, their biological clas-
sification would be a matter of little consequence, but
their broad signification in historical geology, as well as
in systematic biology, renders it necessary that they be
classified as nearly as possible in the manner that living
animals and plants are classified.”’ :

While it is in no way desired to overlook or underesti--
mate the biologie value of such fossil plants as have for-

ô

210 THE AMERICAN NATURALIST [ Vou. XLVI

tunately retained their internal structure in condition for
successful study, it is probably safe to say that their
value to geology as compared with the impressions of
plants is as 1,000 to one, and had we only the former, there
never could have been developed the science of strati-
graphic paleobotany. For example, the collections of the
U. S. National Museum embrace over 100,000 specimens
of the impressions of Paleozoic plants, whereas of those
showing internal structure there is hardly a half dozen
unit trays full. In the Mesozoic and Cenozoic collections
belonging to the same institution there are thousands
upon thousands of specimens from hundreds of localities
and horizons, while of those retaining their internal
structure there are so few that they can almost be num-
bered in tens.

There is another and an excellent practical reason why
the impressions of plants are, and will always remain, of
more value to geology than those exhibiting internal
structure, no matter how well this structure may be pre-
served. As soon as a plant impression is exhumed it is
instantly ready for study and may be interrogated at
once as to the stratigraphic story it has to tell, whereas
the plant with the structure preserved usually shows
little or nothing on a superficial examination, and re-
quires laborious, expensive preparation before it can be
identified. For example—to make a personal applica-
tion—for the past five years I have annually studied and
reported on from 500 to 700 collections, each of which em-
braced from one to hundreds of individuals, and with
them have helped the geologists to fix perhaps fifty hori-
zons in a dozen states. If it had been necessary to cut
sections of these specimens before the geologist could
have had his answer, it is safe to say that very little
would have been accomplished.

All fossil plants must be interpreted by and through
the living flora. In the more recent geological horizons
the plants are naturally found to be most closely related
to those now living, but as we proceed backward r time

No. 544] BOTANICAL SOCIETY OF AMERICA 211

the resemblances grow less and less and finally we find our-
selves in the presence of floras a large percentage of which
are without known or clearly recognized living represen-
tatives. In describing these and making them available
for stratigraphic use it has been necessary to give them
generic and specific names, after the analogy of the living
floras, so that we may have convenient handles by which
to use them. Many of these are confessedly what may
be called genera of convenience, such, for example, being
many of the genera of the so-called ‘‘ferns’’ of the Paleo-
zoic. Some—but especially botanists—unfamiliar with
the geological use of fossil plants, have argued that it is
unsafe, or even actually unwise to venture to give names,
not only to those without living representatives, but even
to those obviously belonging to living groups. A reply
to this objection seems unnecessary in view of what has
been said. _

The practical application of fossil plants as an aid to
geology may be briefly mentioned. There have been de-
seribed from—let us say—North America, upwards of
5,000 species, of which number some 1,200 are confined to
the Paleozoic, perhaps 2,000 to the Mesozoic, and 1,500 to
the Cenozoic. During the sixty or seventy years that
this information has been accumulating it has developed
that certain species or other groups enjoy a considerable
time range, and therefore are of little value in answering
close questions of age, while others are of such limited
vertical distribution that their presence may indicate in-
stantly a definite horizon. Thus, if hej d in association

impressions that we have named Sequoia N ordenskiöldi,
Thuya interrupta, Populus cuneata, etc., it is known in-
stantly that we are dealing with the lower Eocene Fort
Union formation, since not one of these species, together
with several hundred others, has ever been found outside
thishorizon. Innumerable other concrete examples could
of course be given, though hardly necessary, yet it may
be instructive to note that within a single geographic
province—the Rocky Mountain region—the several plant-

212 THE AMERICAN NATURALIST [Vor. XLVI

bearing formations present are characterized as follows:
The Kootenai by 120 species, the Colorado by perhaps 50
species, the Dakota by 460 species, the Montana by 150
species, the Laramie by 140 species, the Arapahoe by 30
species, the Denver by more than 150 species, the Fort
Union by from 500 to 700 species, etc., ete. This shows
that, as Professor J. W. Judd once said: ‘‘ We still regard
fossils as the ‘Medals of Creation,’ and certain types of
life we take to be as truly characteristic of definite periods
as the coins which bear the image and superscription of _
a Roman emperor or of a Saxon king.’’

Just a word may be said on the economic application
of stratigraphic paleontology. It is perhaps safe to say
that never in the history of American geology has there
been so close an interrelation and dependence of geology
on paleontology as at present, and of this confidence
paleobotany may justly claim its full share. Thus, of the
even dozen of paleontologists in the employ of the U. S.
Geological Survey and covering all branches of the sub-
ject, four are paleobotanists.

Among the many subsidiary problems connected with
the application of paleobotany to geology, the use of
fossil plants as indices of past climate occupies a most
important place. As the majority of plants are attached
to the substratum and hence are unable to migrate like
most animals when the temperature of their habitat be-
comes unfavorable, they must either give way or adapt
themselves gradually to the changed conditions of their
environment. Tlé@efore, fossil plants have always been
accorded first place as indices of past climates. ‘‘They
are,” as Dr. Asa Gray has said, ‘‘the thermometers of
the ages, by which climatic extremes and climate in gen-
eral through long periods are best measured.”’

To those who have not given especial consideration to
the subject, the idea appears to obtain that climatic varia-
tions, such as now exist, are normal or essential and that
they were present without marked differences during all
geological ages.’ It is now established, however, that this

No. 544] BOTANICAL SOCIETY OF AMERICA 213

conclusion is entirely without geological or paleobotan-
ical warrant, and that the most pronounced climatic dif-
ferentiation the world has known extends only from the
Pliocene to the present. As a matter of fact we of to-day
are living in the glacial epoch in what possibly is only an
interglacial period, and we know that the time which has
elapsed since the close of the last ice-invasion has been of
less duration than was one, and possibly two, of the Pleis-
tocene inter-glacial periods. We also know that the cli-
mate was milder during these inter-glacial intervals than
has obtained since the final retreat of the ice, as shown by
the fact that in eastern North America certain species of
plants then reached a point some 150 miles further north
in the Don Valley than they have since been able to
attain. The development of strongly marked climatic |
zones, at least between the polar circles, is, then, ‘‘excep-
tional and abnormal, and we have no evidence that in any
other post-Silurian period, with the possible exception of
the Permo-Carboniferous period, has the climatic distri-
bution and segregation of life been so highly differen-
tiated and complicated as in post-Tertiary times.’”

The regular and normal conditions which have existed
for vastly the greater part of geologic time, have been
marked by relative uniformity, mildness and comparative
equability of climate. This is abundantly shown by the
almost world-wide distribution and remarkable uniform-
ity of the older floras: When, for instance, we find the
middle Jurassic flora extending in practical uniformity
from King Karl’s Land, 82° N., to Louis Philippe Land,
63° S., we have conditions which not only bespeak a prac-
tically continuous land-bridge, but exceptionally uniform
climatice conditions. To have made this possible there
could have been neither frigid polar regions nor a torrid
equatorial belt, such as now exist. Theabsence of growth-
rings in the stems of these plants, as well as the presence
of such warmth-loving forms as cycads and tree-ferns,
point to the absence of seasons and the presence of mild
and equable climatic conditions.

1See White and Knowlton, Science, N. S., Vol. 31, 1910, p. 760.

214 THE AMERICAN NATURALIST [ Von. XLVI

Another example of similar import is afforded by the
early Pennsylvanian flora, that is, the flora of the lower
part of the Upper Carboniferous. Wherever terriger-
ous beds of this age have been discovered, representa-
tives of this peculiar flora, which includes such common
genera as Lepidodendron, Sigillaria, Sphenophyllum,
etc., have been found, this distribution ranging from
South Africa to Brazil and Argentina, and thence over
the northern hemisphere.

Similarly, the Mississippian flora (Lower Carbonifer-
ous) has been found in Spitzbergen, Greenland and arctic
Alaska, and thence south over Europe and America, and
although somewhat older than the last, is distinctly re-
lated to that in Argentina.

On passing up in the geologic time scale we find that
during late Mesozoic and early Cenozoic time the present
dominant types of vegetation were firmly established.
With what probability of success may these floras be in-
terrogated as to the climatic conditions under which they
existed? We find from a study of the present flora that
certain types of vegetation, as well as certain plant asso-
ciations, have definite climatic requirements. Thus, Ar-
tocarpus, or the bread fruit trees, are now confined to
within 20° of the tropics, showing that they require the
moist heat of the torrid regions. If, now, we find that
Artocarpus once throve in Greenland 70° or more north,
during Cretaceous time, we feel justified in assuming that
its climatic requirements were not very different from
those of its living representatives. And when we find
that it was then in association, as it is to-day, with cycads,
tree-ferns, cinnamons, palms and other distinctly tropical
forms we are confirmed in the opinion that at that time
Greenland must have enjoyed a tropical or at least a
sub-tropical climate.

Another example is afforded by the Fort Union forma-
tion. In the rocks of this horizon, which now occur on
the wind-swept, almost treeless plains of the Dakotas,
Wyoming and Montana and thence northward to the val-

No. 544] BOTANICAL SOCIETY OF AMERICA 215

ley of the Mackenzie, are found remains of Sequoia, Tax-
odium, Thuya, Ulmus, Populus, Vitis, Platanus, Sapin-
dus, Viburnum, Corylus, Juglans, Hicoria, etc., ete. From
this array we feel justified in assuming a cool to mild
temperate climate for this early Eocene flora, and further,
from the presence of numerous, often thick beds of lig-
nite, that there was a much higher precipitation than at
present.

A layer of fan-palm leaves a foot in thickness in a for-
mation in northern Washington indicates climatic re-
quirements in which the minimum temperature did not
fall below 42° F. The presence of numerous West In-
dian types in the Miocene lake beds of Florissant, Colo-
rado, would alone point to almost tropical conditions, but
as these are associated with others of more northern
affinities, it seems safe to predicate at least a warm tem-
perate, or possibly sub-tropical climate.

II. Tue RELATIONS or PALEOBOTANY To Botany
1. Phylogeny and Taxonomy

PROFESSOR JOHN M. COULTER
UNIVERSITY OF CHICAGO

Ir is impossible to disentangle morphology and phy-
logeny, for the largest motive in modern morphology is
to construct phylogenies. An excessive amount of over-
lapping will be avoided in this paper by laying the em-
phasis upon the inferences to be drawn from morpho-
logical investigations as to probable lines of descent,
rather than upon the morphological results themselves.

It should be kept clearly in mind that the material of
paleobotany, as indicated by the program, is not always
used to contribute to the science of botany. The determi-
nations of plants inthe interest of geological horizons are
of immense service to geology, but of comparatively little
value to botany. This means that some paleobotanists
are geologists, and some are botanists, and it is the work
of the latter that concerns us at this time.

216 THE AMERICAN NATURALIST [ Vou. XLVI

The recent rapid development of our knowledge of the
structures of fossil plants is familiar to botanists, con-
stituting as it does one of the most remarkable chapters
in the history of our science. This has been due not only
to the elaboration of a technique for sectioning petrifac-
tions, but also to the inclusion of the vascular system
among the morphological material that is recognized to
be significant in conclusions concerning phylogeny. From
the standpoint of paleobotany, the vascular system is its
most valuable asset for vascular plants, for its chances
of preservation exceed those of any other structures ex-
cept the seed, and its significance in phylogeny is far
more apparent and extended than that of the seed. Asa
result of this paleobotanical connection, the phylogeny of
the vascular groups can be made now a resultant of com-
parative structures and actual history. Many an old
phylogeny, based upon the comparative structures of ex-
isting plants alone, has been contradicted by history,
which, in the nature of things, must furnish the final
check upon any proposed phylogeny.

The topic of this paper is really an invitation to indi-
cate some of the recent reactions of modern paleobotany
upon the phylogenies of vascular plants. The title in-
cludes taxonomy, but in so far as this deals with great
groups, defined or discovered, it is covered by the state-
ments concerning phylogeny. So far as it deals with the
recognition of individual forms, it is clear that paleo-
botany must learn to recognize the relationships of fossil
plants, or there will be no reliable taxonomy or phy-
logeny. So long as paleobotany depended upon the form
resemblances of detached organs, there could be no tax-
onomy in the real sense. It was merely a cataloguing of
plant material. But when it learned to uncover struc-
ture, it began to establish a real taxonomy. The contri-
butions of paleobotany to taxonomy, therefore, may be
summed up in the statement that it has begun to extend
our schemes of classification into the ancient floras; that
this has resulted in a far truer view of the great groups

No. 544] BOTANICAL SOCIETY OF AMERICA 217

than their expression in the present flora can possibly
give; and that this makes a rational phylogeny possible.

I will address myself in the main, therefore, to phy-
logeny, as involving all the taxonomy that is of large im-
portance. When paleobotany to-day assembles the great
series of paleozoic pteridophytes, the impression is very
different from that of a few years ago. It is true that we
have always heard of the giant forms of the Paleozoic
and their dwarf representatives of to-day. We con-
trasted Lepidodendron with Lycopodium, and Calamites
with Equisetum, and the total impression was the strik-
ing difference in size.

Now we have learned that these paleozoic Lycopodiales
and Equisetales were not merely comparatively large, but
that they were also comparatively complex. For exam-
ple, we have learned that their huge bodies developed
secondary wood and had attained heterospory. We be-
gin to understand that vascular plants, with the exception
of Angiosperms, were as completely differentiated, so
far as the great groups are concerned, at the beginning
of our records as now; and that the phylogeny in sight is
not that of one great group following another, but of all
the great groups spraying out into more and more modern
expressions.

Not long ago, our morphology taught that the homo-
sporous Lycopodium is the modern representative of the
arborescent, paleozoic club mosses, and that the hetero-
sporous Selaginella is a modern offshoot. Now we find
that the paleozoic forms, with their heterospory, their
ligules, and their other structures, link up with Selagi-
nella; and we are asking, where are the ancestors of Ly-
copodium? Not long ago Isoetes, once suspected of being
responsible for the monocotyledons, was fluttering be-
tween Filicales and Lycopodiales, and at last had settled
down as an anomalous member of the latter group; and
now modern paleobotany assures us that its whole struc-
ture suggests that it is a much reduced and compacted
Lepidodendron. Thus the anomalous Isoetes has a ten-

218 THE AMERICAN NATURALIST [ Vou. XLVI

tative connection, and the very normal Lycopodium has
none.

Once we talked of the evolution of the strobilus as
shown by Lycopodium, Selaginella, and Equisetum, and
fancied that we saw in these modern forms the highest
expression of the pteridophyte strobilus. Now we know
that these strobili of to-day are excessively simple as
compared with those of the Lycopodiales and Equisetales
of the Paleozoic. The evolution of the pteridophyte
strobilus that is in sight, therefore, is an evolution from
a complex strobilus to a simple one.

According to the old morphology of external form, the
club mosses were entirely capable of having given rise to
the conifers. With some knowledge of structure this
idea faded away, except in certain quarters, and the bril-
liant discovery of the so-called ‘‘ seed-ferns ’’ seemed to
dispose of it entirely. Now we find among the paleozoic
lycopods, notably some of the herbaceous ones, that a
seed-like structure has been attained, and we have ‘‘seed-
club mosses’’ as well as ‘‘seed-ferns.’’ This does not
mean that the club mosses gave rise to conifers or to any
other existing group of seed plants, for more important
structures forbid it; but it does mean that the paleozoic
groups had advanced very far; that the seed-condition
may have been attained by several groups of paleozoic
pteridophytes; and that it takes more than a seed to dis-
tinguish a ‘‘seed-plant.”’

These are a few illustrations of the upsetting facts that
modern paleobotany has been introducing into the old-
time phylogenies of Lycopodiales and Equisetales.

It is among the phylogenies of Filicales, however, that
modern paleobotany has wrought the greatest change, so
far as pteridophytes are concerned. The old picture in-
cluded a luxuriant fern vegetation during the Paleozoic,
which culminated in the Coal-measures, whose so-called
‘‘fronds’’ made up at least one half the vascular flora.
The occasional attached synangium or sorus was plainly
like those of the Marattiacee, and since the Marattias

No. 544] BOTANICAL SOCIETY OF AMERICA 219

are eusporangiate, these paleozoic ferns were eusporan-
giate. And so we settled back comfortably to the convic-
tion that the paleozoic ferns were Marattiacex, repre-
sented to-day by a small tropical family; that the euspo-
rangiate habit and synangia are historically older condi-
tions than the converse; and that the leptosporangiate
habit and sori of free sporangia are comparatively
modern.

The change of view that modern paleobotany has intro-
duced must be familiar to most of you. The abundant
fern vegetation of the Paleozoic has vanished, having
been replaced by a great group of fern-like gymno-
sperms; many of the marattiaceous synangia have proved
to be the microsporangiate structures of these same gym-
nosperms; there is no indication that the eusporangiate
habit is older than the the leptosporangiate; and it is en-
tirely clear that our earliest known ferns had free spo-
rangia and not synangia. In fact, after the first revul-
sion of feeling, following the discovery of the fern-like
gymnosperms, the question was seriously raised, is there
any evidence of paleozoic ferns? Of course no one
doubted their existence, but where is the evidence?

Paleobotany has now begun to answer this important
question. Evidence of the existence of a group of arbo-
rescent, Marattia-like ferns during the Upper Carbonifer-
ous is accumulating. Much of this evidence is negative,
for it consists simply of the fact that many species of cer-
tain of the large, so-called ‘‘frond genera” have not been
found to be fern-like gymnosperms. On the doctrine of
chance this may be worthy evidence. It simply means
that a vast display of fern-like leaves must contain some
ferns. The positive evidence, however, is supplied by
vascular anatomy, and is found in the stems called Psaro-
nius. The structure of these stems has been compared
recently with the very characteristic structure of the
stems of the living marattias, and the close relationship
is obvious. Moreover, the leaves of Psaronius belong to
one of the largest frond genera, and some of them bear

220 THE AMERICAN NATURALIST [ Vou. XLVI

Marattia-like synangia. So clear is the evidence that
Scott has called these arborescent Marattias, of the later_
Paleozoic, Palzeo-Marattiaceae, and they may well stand
for the precursors of the much humbler Marattias of
to-day. |

But these ancient Marattia-like forms were not the
oldest ferns, for paleobotany has revealed a much older
assemblage of undoubted ferns, so old, in fact, that the
assemblage is called Primofilices. They are represented
not only by numerous detached sporangia, many of which
have the annulus characteristic of modern leptosporan-
giates, but all the essential structures of one well-char-
acterized family, the Botryopteridex, are known. These
still somewhat vague Primofilices are extremely suggest-
ive and perplexing. The free sporangia and the annulus
of the Botryopteridee suggest leptosporangiate connec-
tions; but the sporangia are not borne as among the lep-
tosporangiates, nor is the annulus of the same character.
The sporangia are in clusters terminating the bare
branches of a rachis; and the annulus is a vertical, multi-
seriate band on one side or both sides of the sporangium.
In fact, the so-called ‘‘rudimentary’’ annulus of Osmunda
suggests a reduced multiseriate annulus of one of the
Botryopteridee. The vascular system, however, is very
characteristic and very primitive, and is so varied as to
suggest a synthetic type that might have given rise to all
the diversities found in modern ferns.

The possibilities of paleobotany are well shown in con-
nection with these Primofilices. One of its form-genera,
Stauropteris, known only as sporangia, has been discov-
ered by Scott with germinating spores. The germination
of a fern spore is so different from that of a microspore
of gymnosperms, or of any other heterosporous plant
with which we are acquainted, that it is clear that Staur-
opteris is a fern sporangium and not the microsporan-
gium of some gymnosperm. When germinating spores
are preserved, and also the swimming sperms of paleo-
zoic gymnosperms, as recently described, we may expect

No. 544] BOTANICAL SOCIETY OF AMERICA 221

that paleobotany will presently be able to uncover all of
the essential morphology of the great fossil groups.

This picture of paleozoic ferns is somewhat dim yet,
but were it not for the recent work in paleobotany we
should have no picture at all, or, what is worse, an
entirely false one.

The most conspicuous contribution of modern paleo-
botany, however, is its remarkably complete reconstruc-
tion of the phylogeny of gymnosperms. Our present
records of this group extend through a longer period and
are more continuous than for any other vascular group.
It was not only associated with the paleozoic pterido-
phytes of the Coal-measures, but it was contemporaneous
with pteridophytes throughout all their recorded history.
Seed-plants, therefore, are just as old as any vascular
plants, so far as our records go. It is clear that seed-
plants have descended from pteridophytes, but when our
records begin, they had already descended.

Our conception of gymnosperms before the paleobo-
tanical work of the last decade, will emphasize the change
that has taken place. We thought of them as cycads,
conifers, and gnetums; and the morphology of that time
undertook to develop a phylogenetic sequence with these
three groups. Cycads were clearly the most primitive
gymnosperms, and when Ginkgo was found to share with
them in the retention of swimming sperms, it was just as
clearly a ‘‘transition’’ form, on its way from cycads to
conifers. In the dim paleozoic background Cordaitales
lurked, but they were quite detached from living gymno-
sperms, a group that belonged to the gelogist rather than
to the botanist. I fancy that many a student of gymno-
sperms in those distant days never even heard of Cor-
daitales. The contrast between such a perspective of the
group as I have just indicated and the perspective we
possess to-day is due almost entirely to paleobotany.

In the first place, it resurrected the most primitive
group of gymnosperms, the Cycadofilicales, called by our
English brethren the Pteridosperms. It is a group that

222 THE AMERICAN NATURALIST [ Vou. XLVI

had always been with us, disguised as paleozoic ferns, and
the story of its recognition and rehabilitation is about the
most sensational one in the annals of paleobotany. This
meant that Cordaitales had a companion group, and that
there were two great gymnosperm assemblages during
the Paleozoic. Then the question arose as to the rela-
tionship of these two groups. There was no historical
sequence to answer the question, for the two groups were
observed side-by-side, and very distinct, as far back as
the records go. This threw the answer back upon com-
parative structures, and this testimony is clear. If the
two groups have been derived from ancient ferns, and
the structure of Cycadofilicales hardly admits of any
other conclusion, it is evident that the Cordaitales have
departed much further from that origin. Therefore, if
the two groups are related to one another genetically, and
the discovery at various paleozoic horizons of persisting
synthetic forms that combine features of both groups
seems to make this reasonably assured, the Cordaitales
were derived from Cyecadofilicales older than any we
know. :

The beginning of our perspective, therefore, is that
very ancient ferns, earlier than any of our records of
vascular plants, gave rise to a Cycadofilicales stock, which
persisted throughout the Paleozoic; and that this stock,
also earlier than any of our records, gave rise to the Cor-
daitales branch. The records begin with these two stocks
working along towards their modern expression, and all
Mesozoic and modern gy perms can be referred to
them. In other words, the gymnosperm genealogical tree
comes into sight as two strong branches of a dichotomized
trunk whose existence and character are hypothetical.

The Mesozoic branches from these two paleozoic stocks
furnish us with another triumph of paleobotany, and this
achievement is of peculiar interest to Americans.

The Cycadofilicales of the Paleozoic contributed to the
Mesozoic the gy perms once known as ‘‘fossil cy-
cads,’’ a group so dominant and so characteristic that the

No. 544] BOTANICAL SOCIETY OF AMERICA 223

Mesozoic was long called ‘‘the age of cycads,” so far as
its vegetation is concerned. These characteristic Meso-
zoic forms are known now as the Bennettitales, and their
resurrection from rich Mesozoic deposits of America is
due to the skill and patience of our American colleague,
Dr. Wieland. They retained many of the primitive fea-
tures of the Cycadofilicales, but departed from them
chiefly in the development of a strobilus. Not only so,
but the strobilus is peculiar among gymnosperms, all of
which have strobili except the fern-like Cycadofilicales.
The strobilus of Bennettitales is bisporangiate, and the
two sets of sporophylls are related to one another as they
are in the flowers of Angiosperms. With the investment
of sterile bracts, the strobilus as a whole bears a remark-
able structural resemblance to such a flower as that of
Magnolia. This resemblance has proved to be very se-
ductive, for it has led to the claim that Bennettitales rep-
resent the ancestral forms of angiosperms. It is not the
province of the present paper to discuss this claim. It
has in it as a basis just enough of structural resemblance
and of historical timeliness to make a plausible argu-
ment, but hardly enough to carry conviction to those who
must take other facts into consideration. In any event,
the Bennettitales are a notable group, and paleobotany
has revealed them to us.

Along with them, the true cycads, or Cycadales, began
to appear, apparently never a dominant group, and they
have persisted in the tropics to the present time. The
cycads as we know them, therefore, are the modern rep-
resentatives of an old phylum, which included Bennetti-
tales in the Mesozoic and Cycadofilicales in the Paleozoic,
a phylum aptly called the Cycadophytes. The cycads
present us with that paradox with which students of phy-
logeny are familiar, namely, they are structurally very
primitive, but historically modern. So far from being
the oldest of living gymnosperms, they are younger than
the conifers and the ginkgos.

The Cordaitales of the Paleozoic had already developed

224 THE AMERICAN NATURALIST [ Vou. XLVI

strobili, and had made notable changes in the vegetative
body, changes which characterize the second great gym-
nosperm phylum, fittingly called the Coniferophytes.

The connections of the paleozoic Cordaitales with the
Mesozoic ginkgos and conifers are very obvious, and
these two groups, associated with Bennettitales and some
Cycadales, made up the mesozoic gymnosperm flora.
The Ginkgoales are really a mesozoic type, for their rep-
resentation to-day by a single species is probably due to
preservation by culture. It is an interesting fact that
the ginkgos, directly connected with the paleozoie Cor-
daitales, have retained the primitive reproductive struc-
tures of that group and of the Cycadophytes, but ad-
vanced in vegetative structures as did the conifers. The
group is very old in its reproductive methods, and modern
in its vegetative body.

The great gymnosperm group of the Mi aoabis, as of
the present time, is the Coniferales, and their deployment
during the Mesozoic is a subject of fascinating interest.
The so-called tribes of conifers are found distinctly dif-
ferentiated during the Mesozoic, and the determination
of their relative ages is one of the triumphs of vascular
anatomy. My colleague in this symposium, Professor
Jeffrey, has lived in the storm center of this work, and
it would not be fitting for me to invade his own special
field. It may be said, however, that to discover that the
Abietineæ (the pine tribe) are the oldest conifers is up-
setting the older phylogeny fully as much as paleobotany
has done for other gymnosperm groups.

The conifers are distinguished among the other gym-
nosperms that have been discussed, in being modern both
in reproductive methods and vegetative structure.. In
contrast with cycads, therefore, they present the same
paradox conversely, that is, they are structurally younger
than cycads, but historically older.

This hasty outline of gymnosperm history, which pa-
leobotany has interpreted for us, proves convincingly
that no plant phylogeny is adequate until it has included

No. 544] BOTANICAL SOCIETY OF AMERICA 225

the historical record, which of course is the province of
paleobotany; and furthermore, that general conclusions
= based upon the study of the living flora alone are more
apt to be false than true.

The great problem of paleobotany to-day is the history
of angiosperms. Having perfected a weapon in the at-
tack upon gymnosperms, it remains for the paleobotanist
who is a vascular anatomist to uncover the origin of our
greatest group, with its comparatively brief history.
The origin is probably recorded in the Mesozoic, and we
wish to see the significant structures, and not guess at
external form, and much less guess at purely hypothetical
connections. To this great task paleobotany is turning.
We have had the guesses; and I am confident that pres-
ently we shall have the facts.

II. Tar RELATIONS or PALEOBOTANY TO Botany

2. Morphology

PROFESSOR EDWARD C. JEFFREY

HARVARD UNIVERSITY

Tue morphology of Goethe’s metamorphoses, a hun-
dred years ago, was entirely external morphology, illum-
inating at the time and for many decades later; but now
for over a quarter of a century extinct, except so far as
it lives on for descriptive purposes in manuals of sys-
tematic botany. It has been replaced by a conception
of morphology, based not on external form but on inter-
nal structure. The replacement has been slow in this
country, where, unfortunately, morphology is still very
largely a thing of external threads and patches. The
father of plant morphology in its modern form, was, as
you all know, Wilhelm Hofmeister, who, over fifty years
ago, began to put the subject on a truly evolutionary
basis. Like many other men of genius, he was before
his time and received little hab reer from his less
oo BO det

226 THE AMERICAN NATURALIST [Vou. XLVI

But it is Professor Coulter’s task to deal with morphol-
ogy, particularly in its paleobotanical aspects, in rela-
tion to systematic botany, as is specially fitting, since the
laboratories over which he presides at Chicago have up
to the present time been the greatest single influence in
putting American morphology on a modern and progres-
sive basis. His important relation to systematic devel-
opment in this country is also well known, representing
as it does an earlier but not less striking phase of his
botanical activity. My excuse for even referring to sys-
tematic matters in these preliminary remarks is the
close relation in which they necessarily stand or should
stand to morphology and paleobotany. Darwin in his
immortal ‘‘Origin of Species,’’ although but little given
to figurative language, has described morphology as the
soul of natural history. It is to be feared that there is
just ground for complaint, that the botanical natural
historian has in the past too often worn his soul upon his
sleeve or has even appeared to lack that necessary ad-
junct of higher existence. Internal morphology now
holds the field and all other lines of botanical activity
from systematic botany to plant physiology must take
account of its doings, if they are to make solid and en-
during progress. |

If a change has come in recent years in the point of
view of morphology, an equally important shifting of
position has taken place, during the past few decades in
the attitude of paleobotany. Until the late seventies of
the nineteenth century, paleobotany had to do practically
with the external form of plants alone, as they appear as
impressions in the rocks. It is true that Brongniart in
the earlier years of the last century realized the impor-
tance of internal structure in the case of fossil plants and,
as Dr. Scott has recently pointed out, his views on this
subject are so clear and fit actual conditions so well, that
they read as if they were written only yesterday. But
Brongniart had few followers even among his own coun-
trymen. To the Englishman, Williamson, belongs a large

No. 544] BOTANICAL SOCIETY OF AMERICA 227

- part of the credit of putting paleobotany on a really
satisfactory footing. He insisted on the absolute neces-
sity of taking into consideration internal structure as
well as external form, and went so far in some of his
writings as to state that imprints or impressions alone,
of extinct plants, had little scientific value. What may
be called the evolutionary bias of Englishmen, a fortu-
nate inheritance from the greatest of all biologists,
Charles Darwin, has led them far in the pursuit of paleo-
botany on anatomical lines. We need only call to mind
the demonstration on anatomical grounds that the
greater part of the trees of the Coal Period were, in spite
of their arboreal habit, in reality vascular cryptogams,
a demonstration absolutely confirmed later when it was
possible to study, in detail, their reproductive organs.
An even more striking illustration of the same kind is
supplied by some of the fern-like forms of the Paleozoic,
which have long appeared in the catalogue as vascular
cryptogams. Here too it was first shown on anatomical
grounds, and afterwards from the examination of the or-
ganization of the reproductive structures, that outward
appearances were entirely deceptive and that the organ-
isms in question were in reality seed plants of a prim-
itive type. These two illustrations, which might be in-
definitely multiplied, serve to indicate the very impor-
tant services which the study of internal organization
has rendered both to the natural system and to the doc-
trine of evolution. They further make it clear that the
aphorism of mundane existence, which admonishes us
not to trust to appearances, holds equally well in the
domain of plants. Paleobotany, like morphology, has
accordingly at the present time entered into the anatom-
ical phase of development.

So long as morphology concerned itself mainly with
the external form of the reproductive structures of exist-
ing plants and paleobotany had perforce to content itself
for the most part with the impressions upon the rocks
of the foliar organs of extinct ones, there was little to

228 THE AMERICAN NATURALIST [ Vou. XLVI

bring these two branches of botanical science together. -
The systematic botanist was accordingly quite safe in
ridiculing the insufficiency of the evidence upon which
the conclusions of the older paleobotany was founded.
With the advent of the anatomical phase of development
in both morphology and paleobotany, the two sciences
have become united on the basis of common interests and
are both enormously strengthened by the union. Morphol-
ogy, from being the exponent of a priori philosophical
ideas, applied to the question of the evolution of plants,
and derived for the most part from the inner conscious-
ness rather than from any truly scientific and inductive
study of facts, has become the logical fancy-free hand-
maiden of evolution. Paleobotanical science, on the other
hand, having realized, especially in the case of the older
plants, which are naturally of the greatest importance
from the evolutionary standpoint, that the external form,
even the external form of the reproductive structures,
is often very deceptive as to real affinities, has come to
regard as most important the much less variable internal
organization of extinct plants. Paleobotany is in the
position to supply us now, for the first time, with reliable
facts regarding the organization and true systematic
affinities of the ancient vegetation of the earth, and
morphology has reached a condition of maturity where
facts are infinitely more important than philosophical
fancies, however charmingly expressed.

It is unfortunately the case that morphology, until
comparatively recently, has been quite as unscientific in
its methods as the systematic botany with which in the
earlier years of its existence as a branch of botanical
science it was intimately allied. The recent important
change of attitude in plant morphology is practically en-
tirely due to extensions in our knowledge of the organiza-
tion of extinct plants. Sachs in the second part of his
classic ‘‘History of Botany,’’ which deals with plant
anatomy and similar matters, deplores this unsatisfac-
tory condition in the following words: ‘‘Owing to the ex-

No. 544] BOTANICAL SOCIETY OF AMERICA 229

traordinary diversity of opinion that exists among botan-
ists even on the most general questions in the science, it
is extremely difficult to ascertain what can be considered
a common possession,—an unfortunate condition of things
from which no science suffers perhaps so much as
Botany.’’ Until comparatively recently these words
were as true of plant morphology and plant physiology
as when they were written, nearly forty years ago. If
there is any mitigation of the situation it is because of
the application of chemical and physical laws to the un-
derstanding of the functioning of plant structures and
the facts of modern paleobotany to the elucidation of the
structures themselves. There can be no doubt whatever,
that, without the background supplied by our increasing
knowledge of fossil plants, the picture painted by the
morphologist and embryologist of the evolution of plants
is without depth and entirely without perspective. We
literally can not see our wood for the countless trees
which have crowded into the foreground representing its
most modern stage of development. It is certainly im-
possible to formulate even the rudiments of the evolu-
tionary perspective of plants in the absence of paleo-
botanical facts as the enduring and fundamental back-
ground.

It is accordingly impossible to deny that in the past
morphology has been largely fanciful, where it has de-
parted in any way from the bare description of the facts
of structure in modern plants, or has -attempted to
arrange them in accordance with any general or scien-
tific principles. True the situation has been relieved not
a little, as a result of the study of development and com-
parative anatomy, but the difficulty has always been
here to decide the direction in which a comparative or
developmental series should be read. The indispensable
sense of direction can alone be supplied by sighting down
` the fingerposts of the past, the records of fossil plants,
which show us the real path by which evolutionary ad-
vancement has been made in any given case. — -

230 THE AMERICAN NATURALIST — [Vou. XLVI

One of the commonest fallacies of the older morphol-
ogy was to regard simpler structures as more ancient
and complicated ones as more specialized and modern.
This error is very deeply implanted in the existing highly
artificial systematic arrangement of the higher plants.
For example in the case of the conifers, the Taxaceæ
are put lower than the Pinacex, on account of the simpler
organization of their reproductive and vegetative struc-
tures. The paleobotanical record, however, shows us
clearly that the Pinacex, particularly the abietineous
Pinacee, are among the most ancient repr tatives of
the coniferous stock, while the Taxacex, particularly the
genus Taxus, stand for the class in its most modern con-
dition of development. Let us take another illustration
from the Angiosperms. Systematic text-books invari- |
ably place the Monocotyledones below the Dicotyledones,
on the basis of their simpler organization. This view
of the matter does not accord, however, with the results
of anatomical and paleobotanical research, which clearly
show that the Monocotyledones are neither ancient in
their occurrence nor primitive in their organization, but
represent a condition of reduction from ancestors which
were essentially dicotyledonous in their more important
characteristics. These two illustrations, which might be
multiplied indefinitely from systematic works, show that
the older morphology was essentially fanciful and phil-
osophical in its methods and by no means worthy of the
name of an inductive science.

The new morphology, purely inductive in its pro-
cedure, is solidly founded on the testimony of the rocks
and considers no sequence valid, unless it is clearly
supported by the evidence derived from the study of
fossil forms. It shows itself moreover scientific in
its firm adhesion to valid general principles and
its disregard of uncoordinated facts. These general
principles too are few in number and are as easily
grasped in their simplest form as are the great general-
izations of the sister sciences, chemistry, physics and

No. 544] BOTANICAL SOCIETY OF AMERICA 931

astronomy. The three most important laws or general
principles of morphology are those which have to do with
recapitulation, reversion and retention. The first is
common to both plants and animals, while the other two
are infinitely better illustrated by botanical than zoolog-
ical facts. The validity of these laws has long been ad-
mitted in a somewhat hazy and unpractical fashion. It
has remained for morphology based on paleobotany to
bring them into prominence as the fundamental working
principles of the investigator of plant evolution. They |
are in fact the rudiments, the three R’s of biological sci-
ence, with which even the tyro should become well
acquainted.

We can not do better than take a particular illustra-
tion of these laws as applied to botanical facts. The
conifers show themselves particularly favorable for the
elucidation of general evolutionary principles, because
they not only constitute a large, varied and widely dis-
tributed element of our existing flora, but can be traced,
with trifling intérruptions, continuously far into the past.
Of the conifers there is one tribe at the present time
entirely confined to the southern hemisphere. I refer
to the Araucariinex, of which the New Zealand Kauri and
the Norfolk Island pine may appropriately stand as
examples. In the Mesozoic period, the middle ages of
our earth, they flourished throughout the entire world.
It is often considered that the Araucariinee represent
the most ancient group of conifers. This belief appears
to be based on a too common fallacy, that groups nearly
extinct in the existing flora necessarily represent ancient
forms. A further basis for the hypothesis of the extreme
antiquity of the Araucarian conifers is derived from
their habit and the general features of organization of
their wood, in both of which respects they present points
of resemblance, by no means complete, however, with
those Paleozoic gymnosperms to which the origin of the
general coniferous class is by common consent of morphol-
ogists and paleobotanists referred, namely, the Cordai-

232 THE AMERICAN NATURALIST [ Vou. XLVI

tales. If we follow the Araucariinee backward step by
step, as recent additions to our knowledge now permit us
to do, into the past, we find that although in the Tertiary
they retained very largely the characteristics which they
present to-day in the Cretaceous, particularly in the
Lower Cretaceous and in the Jurassic, they become less
and less like their existing representatives and more and
more like the Abietinex, the actually dominant conifers
of the northern hemisphere, both in general organiza-
tion and wood-structure. . To save time let us consider
only three points of anatomical organization, for com-
parison. The existing Araucariinee are remarkable
among conifers and gymnosperms generally, in possess-
ing leaf-traces which continue to'be formed in the wood
by the cambium, long after the leaves which they origi-
nally supplied have ceased to exist. The old trunk of a
Kauri or an Araucaria, for example, shows on the outside
of its wood marks representing the traces of leaves,
which may have disappeared hundreds of years pre-
viously. This feature has been seized upon by Professor
Seward as one of undoubted primitiveness. It is cer-
tainly bizarre, and if one accepts the unusual as the cri-
terion of antiquity, the Araucariinee certainly could
qualify for an ancient lineage on this character. This
view does not however accord with paleobotanical facts. .
In the Lower Cretaceous we find very many undoubted
Araucarian trunks, in which the leaf-traces were not
persistent as in the existing representatives of the tribe
and in the Jurassic and Triassic trunks of this type
become practically universal. This at once makes it
clear that the persistent leaf-trace, so characteristic of
the living Kauri and Araucaria, is not an ancestral fea-
ture of the Araucarian stock. Recently we have sent
out from Harvard a botanical expedition to Australasia.
At my suggestion Messrs. Eames and Sinnott, who were
its personnel, have brought back old seedlings of the two
existing Araucarian genera. On investigation of the
lower region of these in proximity to the cotyledons, it

No. 544] BOTANICAL SOCIETY OF AMERICA a

was found that here the leaf-traces were ephemeral in
their persistence, exactly as in the older Mesozoic repre-
sentatives of the Araucarian stock from the Lower Cre-
taceous, Jurassic and Triassic deposits. Here appears
a very striking example of the validity of the law of re-
capitulation as exemplified by the young individual, the
seedling stem for a short part of its vertical length re-
peating ancestral conditions which have long. disap-
peared in the adult. Let us consider two remaining
characters together, in order -to economize time. The
water pores of the tracheids in existing Araucarian con-
ifers occur in a crowded and alternating condition and
are deformed or flattened by mutual contact, a feature
of resemblance to the oldest known gymnosperms. This
feature is in marked contrast to the pit arrangements of
the other existing tribes of conifers, where the water
pores are loosely grouped and when numerous and mul-
tiseriate are opposite. Another unusual feature of the
wood of existing Araucariiner is the absence of wood
parenchyma, a feature likewise illustrated by the woods
of the Paleozoic gymnosperms. If we follow the Arau-
carian conifers below the horizon of the present, we ob-
serve in the older representatives a condition of pitting
of the tracheids, which the lower we go geologically,
becomes more and more like that found in the other tribes
of conifers, particularly the Abietinex, so much so that
Gothan, who has recently described the Jurassic woods
of Spitzbergen and other arctic islands, which in the
Mesozoic supported a luxuriant flora, has identified them
as of Abietineous affinities. It is further found in the
case of those Mesozoic fossil woods which most nearly
resemble the living genera Agathis and Araucaria, that
wood parenchyma was exceedingly abundant. Now let
us examine our Araucarian seedlings in regard to these
features. The wood of the cotyledonary region here ex-
emplifies both the characteristic pitting and presence of
wood parenchyma of the Mesozoic Araucarioxylon type.
Thus the Araucariinee, which we are able to follow very

234 THE AMERICAN NATURALIST [ Vou. XLVI

far back into the past, present a very striking illustra-
tion of the biological law of recapitulation. Other fea-
tures of the seedling might have been considered with the
same result.

But ancestral conditions are not in plants confined to
the young individual. We find them also illustrated in
connection with the principles of retention and reversion.
The law of retention is well exemplified by the root, cone
and first vegetative annual ring of the Araucariinee,
where the Mesozoic features of structure appear only
less completely than they do in the seedling. The law
of reversion is likewise illustrated readily in this tribe
of conifers in connection with injuries, for the wood
formed subsequently to wounds shows Mesozoic charac-
ters, which are not a feature of normal structure.

It appears clear from the illustration which I have
chosen, only one among many possible ones from the re-
sults of recent coordinated anatomical, developmental,
experimental and paleobotanical investigations, that we
have definite laws of plant evolution. It is further clear
that like Ulysses we must ‘‘follow knowledge like a sink-
ing star” into the night of the past, if we are to reach
durable general biological principles. The laws of re-
capitulation, retention and reversion, founded on cases
where it is possible to trace an unbroken sequence of
forms, are likewise applicable to the elucidation of con-
ditions in groups the past of which is as yet insufficiently
known. A very notable example of this kind is pre-
sented by the Angiosperms. Here more than anywhere
else among the higher plants philosophical views in re-
gard to evolution prevail. You are doubtless all famil-
iar with the generally accepted dictum, forming a feature
of all elementary botanical instruction, that the woody
stem of perennial Dicotyledones is derived from one of
herbaceous texture. This conclusion is based on the old
fallacy, that simpler conditions are necessarily antece-
dent to more complex ones and antedate them in time.
If we investigate the primitive type of stem organization

No. 544] BOTANICAL SOCIETY OF AMERICA 235

in the Dicotyledones, in accordance with the principles
of recapitulation, retention and reversion, we arrive at
very different conclusions indeed and of much greater
significance from the standpoint of evolution.

It will save time to take a concrete illustration. In
the cross section of a small branch of the oak we find
the woody part of the stem, composed of ten alternat-
ingly outstanding and depressed segments, separated
from one another by ten large so-called primary medul-
lary rays. In accordance with the herbaceous hypothesis
of the origin of the woody stem, the projecting segments
are supposed to correspond to five originally entirely
separate bundles. Further the intervening depressed
segments of the cylinder, set off from the others by the
ten large rays, are supposed to represent parts of the
wood which have been secondarily interpolated through
the activity of a so-called interfascicular cambium. If
we examine the facts in the light of the general principles
of plant evolution, a very different and much sounder
conclusion is reached. Let us take only the seedling
evidence, for the other accords absolutely with it. In
the cotyledonary region of the young stem we find for a
number of years a completely continuous woody cylinder,
without either large rays or depressed segments. At
this level in fact the rays are entirely narrow ones of the
gymnospermous type. Both the broad rays and the seg-
ments which they delimit appear only later. The so-
called large primary rays are in fact formed as a result
of the fusion of smaller rays into aggregates around the
incoming leaf-traces. The broad rays of the oak are-
consequently fusion products and not at all primitive
structures. They are clearly related as a storage device
for the strands which enter the stem from the leaves.
There are two broad rays to each leaf, corresponding to
its two strongly developed lateral traces, and since the
phyllotaxy of the oak is of the 2/5 type, there are nor-
mally and originally five pairs of large rays in the stem.
The depression of five of the segments delimited by the

236 THE AMERICAN NATURALIST [ Vou. XLVI

large leaf or foliar rays below the level of their fellows
presents another interesting problem in evolution, into
which there is not time to enter at present. It is enough
to say that the so-called primary or large rays of dico-
tyledonous stems are not primary or primitive struc-
tures at all, but are of secondary origin and formed from
what was originally wood, in order that the assimilates
_ from the leaves may more readily and conveniently be
stored for future use. We have thus an explanation of
the relation of the woody to the herbaceous stem, which
is at once in harmony with the general principles of plant
evolution and at the same time with physiological neces-
sities. The fallacy of considering the herbaceous as the
primitive type is further made clear. There is in fact
every reason to believe that the early Dicotyledones were
entirely woody perennials, just as the forbears of the
existing herbaceous Pteridophyta have been shown by
paleobotanical and anatomical investigations to have
been derived from arboreal ancestors of the Paleozoic.
The clear realization of the universal validity of the
primary laws of plant evolution depends almost entirely
on our increased knowledge of the organization of fossil
plants. There has come to morphology as a consequence
of the study of ancient forms a new birth, quite compar-
able to the remarkable intellectual awakening in Europe,
at the beginning of the modern period, resulting from
the rediscovery of the ancient classics of the Greeks and
Romans. This awakening, known as the renaissance, has
its exact counterpart in plant morphology and our en-
thusiasm over the discovery of new fossil remains which
throw a light on the origin of existing plants is not dif-
ferent from that experienced by the enlightened citizens
of the Italian cities at the end of the middle ages, over .
the unearthing of a new manuscript of Virgil or of
Horace. As a result of this impetus the morphologist
has already in the past decade revolutionized the system-
atic arrangement of the Gymnosperms and the work in
this special field has scarcely begun. In the case of the

No. 544] BOTANICAL SOCIETY OF AMERICA 237

Angiosperms, change is also in the air and I venture to
predict that within two or three decades at most we shall
have made substantial progress towards a scientific, that
is, a natural, classification, of what is at present a huge
and hopelessly confused labyrinth, penetrated at best in
a halting way by the tenuous and insecure thread of our
present highly artificial system.

The future presents many interesting possibilities for
the morphologist and the paleobotanist and let us hope
for the systematist as well. Obviously we are now on the
threshold of the discovery of a system of plants which
shall depict their evolutionary sequence. Would that we
might count on the sympathetic cooperation of all sys-
tematic botanists in this stupendous and intellectually
attractive task. Unfortunately in the thirties of the last
century systematic botany parted company with plant
morphology and has appeared since somewhat to resem-
ble the man with the muckrake in the vision of the im-
mortal tinker. Although there hangs above it the shin-
ing diadem of a natural system, with which to crown its
arduous labors of many years, the raking together of
straws, sticks and even antique dust seems to present an
invincible attraction. The assiduous strokes of the rake
may glean additional straws and sticks; but although
we may all agree that he is a benefactor of mankind who
makes two living blades of grass sprout where there was
one before, we shall scarcely consent with unanimity
that the making of more new species out of old ones, is
a highly commendable scientific occupation. A great
danger on the systematic side appears at the present
time to inhere in the doings of so-called world congresses.
Recently the paleobotanists of those countries which are
mainly active in the investigation of fossil plants have
agreed unanimously and publicly to repudiate the vote
of the recent congress at Brussels, imposing upon them
Latin diagnoses of extinct plants. The students of fossil
plants, although they have to do with organisms no

238 THE AMERICAN NATURALIST [ Vou. XLVI

longer living, regard their science as a live one and are
consequently strongly unwilling to shroud its doings in
the pall of a dead language, perhaps not the less because
the mandate to do so comes as the result of a Russian
ukase. The present world congresses of botanists seem
to present certain ominous resemblances to the world
councils of the church about the time of the Reformation
in Europe. The so-called ecumenical councils of Chris-
tianity were unfortunately characterized at that time by
reactionary tendencies, including among others a prefer-
ence for the Holy Writ in the Latin rendering of the
Vulgate. Unless there is some relaxation of the medie-
val attitude upon the part of the majority of systematic
botanists, there is reason to fear a reformation in botany,
as uncontrollable as that led by Martin Luther and John
Knox in religion. It is sometimes said in favor of Latin
diagnoses of plants, that the older literature of system-
atic botany is in the Latin tongue. The same is true of
the equally venerable sciences of chemistry and astron-
omy, yet for that reason the chemists and astronomers
of to-day do not think it necessary to perpetuate the writ-
ten language of the alchemists and astrologers. Botany
has likewise, let up hope, passed through what corre-
sponds to the alchemistical and astrological phase of
development and need not conceal its doings in the
dog Latin of a Paracelsus or an Albertus Magnus. Con-
ditions in this country are hopeful in this respect, for
although we are not wholly free from the pedantry that
maketh and loveth a Latin diagnosis, the large majority
of American systematists are entirely progressive in
their point of view. May they prevail and in the course
of time, with all modesty, impress their attitude on the
European cultivators of this important field of botanical
activity.

No. 544] BOTANICAL SOCIETY OF AMERICA 239

II. Tue RELATIONS oF PALEoBoTANY TO BOTANY

3. Ecology

DR. ARTHUR HOLLICK

New York BOTANICAL GARDEN

As I understand the object of a symposium it is not to
provide opportunity for the reading of exhaustive or
highly technical dissertations, or for the presentation of
new material, but rather to present recognized facts as
clearly as may be, with recent interpretations of their
meaning or significance, in order to enlist interest in and
to stimulate discussion of the subject under considera-
tion; and this seems to have been the view which was
taken by those who have preceded me. In such connec-
tion it is my privilege to present the claims of ecology to
recognition, in indicating the relations between botany
and paleobotany.

Plant ecology, as the term is commonly defined and
understood, is that branch of botany which deals with the
study of the interrelations of plants and their relations
to environment. As a distinct science it is practically a
product of the present generation. I do not know exactly
when the term was first employed in scientific literature,
but it certainly was not in general use in connection with
botany at the time of my earliest contributions to the
subject, and I did not then know that I was dealing with
ecology when discussing certain floras and their accom-
panying geologic and physiographic features of enyl-
ronment.

The ecologie relations between botany and paleobotany
are mostly concerned with the problems of phytogeog-
raphy. Paleobotany has supplied the explanations of
numerous puzzling facts in regard to the geographic
isolation of certain genera ; the occurrence of some genus
or species only in certain widely separated regions of the
earth; and the problems in connection with many endemic
floras. Indeed the phenomena of plant distribution 1n

210 THE AMERICAN NATURALIST [Vor. XLVI

general at the present time would be lacking in logical or
adequate explanation but for the facts which have been
brought to light by the study of fossil plants in regard to
distribution in the past. Many such instances might be
cited, but for the purposes of this symposium a few of
the most striking only need be recalled to serve as con-
crete examples of the general abstract propositions.

In the earlier part of the last century, when the science
of paleobotany was in its infancy, and much that we now
know about living plants had not been learned, numerous
remains of coniferous trees were found in Europe and
elsewhere in the Old World, in deposits of relatively
recent geologic age. For the most part these remains
were either identified as living genera or were given
generic names designed to indicate their nearest appar-
ent relationships with such (Pinites, Tawites, Arauca-
rites,etc.). Other similar remains, however, which could
not be satisfactorily compared with any living ones, were
given new generic names. Among these latter may be
mentioned certain small cones associated with leafy
twigs, which were assumed at the time to represent an
extinct coniferous genus. Large areas of the New World,
however, were yet unexplored and many hitherto un-
known living genera were awaiting discovery and de-
scription. One of these was Sequoia, a genus of two
species only, confined in their distribution to scattered
groves on the western coast of the United States. Ecol-
ogy was an unknown science when these groves were dis-
covered, but the relatively limited number of the indi-
vidual trees, and their geographic isolation, at once
attracted attention and aroused interest and discussion
in regard to their ancestry and the phenomenon of their
peculiar distribution. Paleobotany supplied the desired
information. When the generic characters of Sequoia
were made known they were seen to be identical with
those of the supposed extinct fossil coniferous genus of
the Old World. Further than this, however, similar re-
mains, comprising numerous different species, were sub-

No. 544] BOTANICAL SOCIETY OF AMERICA 241

sequently found extending through Siberia to the eastern
coast of Asia and through Europe, Iceland, Greenland
and the Arctic regions to Alaska, and thence southward
to the home of the two remaining living species on the
western coast of North America. The phytogeographic
problem of the genus Sequoia, as we know it to-day, was
thus resolved into the geologic problem of the causes
which produced the climatic changes resulting in the ex-
tinction of the genus over vast areas where it formerly
existed, and the total extinction of all except two of the
numerous species by which it was formerly represented.
Modern areal limitation of the genus was thus shown by
paleobotany to be a result of former areal elimination.

(Incidentally it may be remarked that this example
also involves a question of nomenclature which, however,
I trust our chairman may declare to be not germane to
the subject and hence ineligible for discussion. It is one
of the few instances in which a genus was known and
named as a fossil before it was discovered and named in
its living form.)

The genus Taxodium, comprising two, or possibly three
living species, is confined to the southern United States
and Mexico, so far as its present distribution is con-
cerned. Up to the close of the Tertiary period, however,
it flourished throughout what are now the temperate and
arctic zones of North America and Eurasia,—not only
the genus, but apparently the identical species yet living
and others now extinct. Paleobotany has adduced ample
proof of these facts, so that, as in the case of Sequoia, the
present distribution of Taxodium is explained as merely
the result of its elimination from other regions where it
formerly existed.

The monotypic genus Ginkgo, which by many is also
regarded as representing a monotypic family and order,
is confined, so far as its natural distribution is concerned,
to eastern China and Japan. No known facts could ade-
quately account for its taxonomic and geographic isola-
tion until paleobotany revealed the multiplicity of its ex-

242 THE AMERICAN NATURALIST [ Vou. XLVI

tinct specific and allied generic forms, and its former
wide distribution throughout the Eurasian and North
American continents.

Another monotypic living genus, Sassafras, limited in
its present distribution to eastern North America, repre-
sents an ancient type of angiosperm vegetation whose
fossil remains have been found not only throughout
North America, but also in many parts of the Old World.

These are merely a few of the many examples of
generic isolation—geographic and taxonomic—the expla-
nations of which have been supplied by the study of
paleobotany.

Other apparent peculiarities of distribution, such as are
represented in our living flora by Liriodendron and Ne-
lumbo, genera which are restricted to eastern Asia, east-
ern North America and the central American regions, are
exceedingly difficult to explain satisfactorily on any
theory of migration in recent times; and the theory of
origin de novo in each of the widely separated regions is
too thoroughly discredited to merit discussion. None of
the known facts of recent plant migration, or evolution,
or mutation, are adequate to explain the conditions as we
now find them. Paleobotany, however, has demonstrated
that such apparent peculiarities of generic distribution
are readily explained when the facts of former distribu-
tion are ascertained. Each of these genera was formerly
world-wide in its distribution and prolific in species; but
changes in environment caused their extinction in all ex-
cept the widely separated regions which they now inhabit,
reducing the species of Nelumbo to two, and of Lirioden-
dron to one. What have been regarded as problems of
distribution, explainable by improbable theories of mi-
gration or evolution, have thus been shown by the facts
of paleobotany to be merely some of the many examples
of isolation due to elimination in intermediate regions. `

‘The peculiar endemic flora of Australia did not origi-
nate de novo by reason of its isolation. Paleobotany has
shown that the living endemic flora of Australia in many

No. 544] BOTANICAL SOCIETY OF AMERICA 243

of its characteristic elements, such as the genus Huca-
lyptus, represents what the general vegetation of the en-
tire earth was like at the close of Mesozoic time, when
the continent of Australia was isolated from the rest of
the world. Elsewhere than in Australia climatic and
physiographic changes subsequently eliminated the Me-
sozoic types of vegetation and evolved new ones; but in
Australia the conditions remained almost stationary, and
that continent to-day, so far as its native flora and fauna
are concerned, is still in a late Mesozoic or early Neozoic
stage of development. It is an endemic flora not because
it has evolved new types by reason of its isolation, but
because it has remained stationary by virtue of this rea-
son, while the great bulk of the world’s vegetation has
changed.

And so the paleobotanist extends the right hand of fel-
lowship to the botanist and says, ‘‘when you are puzzled,
or in doubt, don’t despair, come to us,’’ for individually
or collectively we can, probably, suggest reasonable ex-
planations, not only as to why living plants have come to
be where they are, but also how they have come to be
what they are. |

SHORTER ARTICLES AND DISCUSSION

THE INFLUENCE OF CAVE CONDITIONS UPON
PIGMENT DEVELOPMENT IN LARVA OF
AMBLYSTOMA TIGRINUM

One of the possible methods of approach to the problem of
the origin of the modifications of cave animals is by experiments
in which outside forms are kept under conditions normally
encountered by animals living in caves. In following up this
method of approach, it seemed best to select forms as plastic as
possible (if perchance there are forms which are plastic with
reference to modifications by cave conditions). Typical cave
animals, 7. e., species highly adapted for cave life, perhaps
without exception belong to families and genera having many
members showing an inclination toward cave habitation either
by actually frequenting or living in caves or by inhabiting
similar dark and retired situations elsewhere; so that it might
seem that these groups possess a certain plasticity toward modi-
fications by cave conditions. Especially might this plasticity
be expected in the particular species which already show a tend-
ency to live under conditions resembling those of caves.

Young animals may be expected to be more responsive to
changed environment than adults, and since many of the uro-
deles live in caves and similar situations, and their eggs and
young can be obtained in numbers and reared with comparative
ease, amphibian larve were selected for some of the experiments.

Newly laid eggs of Amblystoma tigrinum were collected (in
some cases were laid in jars in the laboratory where some of the
adults were confined) from March 30 to April 4, 1911. The
eggs, in small lots, were placed in a number of 6-inch battery
jars, and the jars divided into two lots, Series A and Series B.
Series A was placed in the artificial cave while Series B was
kept on a laboratory table adjacent to and directly in front of
a west window. Otherwise the two series were, as far as pos-
sible, subjected to like conditions. Unfortunately it was not
possible to keep Series B at as low or as uniform a temperature
as Series A. Development was somewhat slower on the average
with Series A than with Series B, and both of these developed
slightly less rapidly than a third series, Series C, observed in the
pool where the eggs were laid and the larve allowed to develop

244

No.544] SHORTER ARTICLES AND DISCUSSION * 245

under natural conditions, but as I shall attempt to show later
neither temperature differences nor different rates of develop-
ment materially influenced the amount of pigment developed.

The animals were fed on alternate days with an abundance
of daphnids and cyclops, and later with small bits of tender
beef. The amount of pigment developed was judged by making
accurate determinations of the color of two definite regions of
the skin of each individual. For the body color, determinations
were made for a region to one side of the median line and about
midway between the pectoral and pelvic limbs. This is the
most heavily pigmented region of the larva. For the head
color the region immediately anterior to a line from one eye
to the other was used. This is usually the least pigmented por-
tion of the animal. The colors were obtained by means of the
Milton Bradley color tops and the records made in percentages
of black, white, orange and yellow, which when blended most
nearly matched the color of the skin of the animal. The orange
used in these tops most nearly resembles the No. 101 of
Klincksieck et Valette’s ‘‘Code des couleurs’’ and the yellow
is intermediate between Nos. 206 and 211. All the color records
under consideration for Series A and B were made within two
weeks and under uniform conditions so that the results ought
to be strictly comparable. The records for the somewhat more
rapidly developing Series © were made about three weeks
earlier. The results tabulated include color records for all the
surviving individuals of Series A and B and sixteen individuals
selected at random to constitute Series C.

The average body color of the Series A, numbering 22 indi-
viduals, reared in the cave, contained 49.7 per cent. of black,
16.9 per cent. white, 9.3 per cent. orange and 24.1 per cent.
of yellow; or 49.7 per cent. black and 50.3 per cent. non-black.
The extreme range in degree of pigmentation was from 32 black
and 68 non-black to, in the case of one individual decidedly
darker than its fellows, 70.5 black to 29.5 non-black. The
average head color was 38 per cent. black and 62 per cent. non-
black (26.7 white, 7.0 orange and 28.3 yellow) with 27 per
cent. and 51 per cent. as the extremes in amounts of black.

On the average, Series B (7 individuals), reared in labora-
tory light, had a body color of 86.1 per cent, black and 13.9 per
cent. non-black (4.7 white, 2.4 orange and 6.8 yellow), and a
head color of 76.1 per cent. black and 23.9 per cent. non-black
(8.4 white, 4.0 orange and 11.5 yellow). The range for the

body color was from 82 to 90.5 black and for the head color : : : : -

740° THE AMERICAN NATURALIST [ Vou. XLVI

from 62.5 to 90 black, ranges distinct from and not at all over-
lapping those for Series A reared in darkness.

Series C (16 individuals), living in the outdoor pool under
natural conditions, had an average body color of 86.6 black and
13.4 per cent. non-black (4.5 white, 3.1 orange and 5.8 yellow),
the extremes ranging from 77.5 to 93 per cent. black. The
head color averaged 78.3 black and 21.7 non-black (7.7 white,
4.2 orange and 9.8 yellow) with 61 per cent. and 90 per cent.
the extremes in amount of black. The data for all three series
are shown together in the adjoining table.

| Body | Head
4 xs
x| s| 8] 2 \|gs|/4|2| | & \32
5 a 8 D on Z a si ee O'i
et o a ee ee ee:
| ae
Series A (cave): |
et ea »: -| 70.5 29.5 51.0 | 49.0
PAPRIGRG a a 32.0 68.0 |27.0 |
AVOPAQOR Wr uae ra | 49.7|16.9| 9.3 124.1 50.3 38.0 26.7 | 7.0 28.3 62. 0
|
Series B (laboratory) | |
es:
Darkat. n 90.5 9.5 90.0 | 10.0
Piehtest 23 82.0 18.0 (62.5 37.5 5
AVARO, oe oo ae 86.1| 4.7 | 2.4 | 6.8 |13.9 | 76.1; 8.4| 4. A ‘11.5 5123. 23.9
Series C (pool):
Extremes:
Parkent. oA 93.0 7.0 90.0 10.0
ETES PERAE E E E Tho 5161.0 39.
Avera a | 86.6| 4.5 | 3.1 | 5.8 |13.4 |78.3 | 7.7 | 4.21 9.8 (21.7

It will be noted that the series reared in laboratory light
(Series B) and the one developing under natural conditions in
the outdoor pool (Series C) agree very closely in the amount of
pigment development, the average percentages of black in the
body color determinations being 86.1 and 86.6, and for the head
color 76.1 and 78.3 respectively, while the extremes for the two
series differ but slightly. However, the most significant fact
is that these series differ so sharply from Series A reared in the
dark and having an average body color of only 49.7 per cent.
black and an average head color of 38 black. On the average
those reared in darkness have about four times as much non-
black in the body color as those reared in light, while the darkest
individual of the cave series, really aberrant for that series
because it is so dark with 70.5 per cent. black in the body color

No. 544] SHORTER ARTICLES AND DISCUSSION (247

and 51 per cent. in the head color, has decidedly less pigment
than the lightest individual of either Series B or Series C, with
77.5 per cent. black in the body color and 61 black in the head
color.

The percentage differences in color between the series reared
in darkness and the two series reared in light do not adequately
represent the color differences as perceived by the eye. The
average person probably would not hesitate to call the lighter
individuals of Series A ‘‘white,’? and on the other hand would
probably class the various individuals of Series B and C as
‘dark gray’’ or ‘‘coal black.’”’

There appear to be no reasons for thinking that the marked
differences in pigment development were due to differences in
nutrition or differences in temperature. Series B so closely
resembled Series C in amount of pigment development that it
is evident the differences in temperature and food did not
greatly influence their pigmentation, yet Series B was reared in
a west room subject to great fluctuations in temperature because
of the extreme exposure to the summer sun during the afternoon,
and the temperature of this series on the average was probably
higher than in the outdoor pool in which Series C developed,
while the food was exclusively daphnids, copepods and bits of |
beef. On the other hand Series A was maintained at a quite
uniform temperature consistently lower than either of the other
series. Fischel (’96) and Flemming (’97) found that with
certain salamander larve (the former at least used larve of
Salamandra maculata) temperature differences influenced the
amount of pigment development, the individuals reared at from
4° to 7° C. being much darker than individuals reared at from
15° to 20° C., indicating that for Salamandra maculata the lower
temperature favors greater pigment development. Inasmuch
as my lightest series, Series A, was reared at the lowest tempera-
ture of any and yet was markedly lighter than the others, and
that Series B and C did not differ widely in amount of pigment
development although reared under somewhat different tem-
perature conditions renders it improbable that temperature in-
fluences produced the differences in pigmentation observed.

Series A and B were fed in every way alike and so far as
could be judged took about the same quantities of food. While
as a whole Series A developed less rapidly than Series B, part of
Series A developed more quickly than some of Series B yet each
individual possessed an amount of pigment similar to the others

of its series, Hence the evidence strongly points to ‘the con- . oe

248 THE AMERICAN NATURALIST [ Vou. XLVI

clusion that the differences in amount of pigmentation were not
due to differences in temperature, food or rate of development
but solely or in much the greater part to the absence or presence
of light.

The difference in pigmentation between Series A on the one
hand and Series B and C on the other is fully as great as the
difference in pigmentation between certain species that live
habitually in caves and others which do not. To this point the
case looks most significant. With the approach of transforma-
tion, however, the amount of pigment rapidly increases, par-
ticularly on the lighter regions of the larve, and this increase
is more pronounced with those reared in darkness than with the
others. Nevertheless there still remains a marked difference in
pigmentation between the transformed individuals of Series A
reared under cave conditions and Series B reared in the light.

Color records of a few recently transformed salamanders
were taken before the spots so characteristic of the adult were
definitely formed. An attempt was made to get careful records
of the general blended color effect independent of the mottling
which is quite pronounced at this stage, during which the pig-
ment is being segregated to form the definite color pattern of
the adult. The average body color of the transformed sala-
manders of Series A was 70.2 per cent. black and 29.8 per cent.
non-black, and the average head color 68.2 black and 31.8 non-
black. These averages for Series B are: Body color, 89.7 black
and 10.3 non-black; head color, 87.7 black and 12.3 non-black.
No transformed individuals of Series C were obtained. Unfor-
tunately records of transformed individuals of Series A and B
were made of only 2 and 5 individuals, so these are scarcely more
than sample records, but they are probably fairly representative
of their respective series, and they clearly indicate that after,
as well as previous to, transformation a distinet difference in
pigmentation persists between the salamanders reared and al-
lowed to transform in the cave and those developing and trans-
forming in the light. A. M. Banta

CoLD SPRING HARBOR, N. Y.

REFERENCES

Fischel, A. ’96. Ueber Beeinflussung und Entwickelung des Pigments.
Archiv f. mikr. Anat. u. Entw., Bd. 47, pp. 719-734.

Flemming, W. ’97. Ueber den Einfluss des Lichts auf die Pigmentierung
der Salamanderlarve. Archiv f. mikr. Anat. u. Entw., Bd. 48, pp. 369-
374, also pp. 690-692, ;

Klincksieck, P., et Valette, T. 08. Code des Couleurs. Paris. 46 pp.

VOL. XLVI, NO. 545~

The American Naturalist

MSS intended for publication and books, etc., intended for ra should be
sent to the Editor of THE AMERICAN NATURALIST, Garrison-on-Hudson, New York.

Articles Sees research work bearing on the problems of ‘organic evolu-
tion are especially welcome, and will be given preference in publica

One hundrea reprints of ope e are supplied to authors ae of charge.
Further reprints will be supplied at cos

Su iptions and ate aan << be sent to the A gg toga The

subscription price is four dollars a yea Foreign postage cents and

Canadian geri weny Fick cents additional. The we for single copies is
forty cents. g rates are Four Dollars

THE SCIENCE PRESS
Lancaster, Pa. Garrison, N. Y.
| : _ NEW YORK: Sub-Station 84

natter, April 2, 1908, at the Post Office at Lancaster, Pa., under the Act of
Congress of March 3, 1879.

THE BULLETIN —Por bargains ia Extsotograpt : TENTH EDITION.

— ial and Pre-historio Specimens. Hooke oa Metotal | THE MICROSCOPE,
| | ethods and to His ners
History, Science, Travel, Voyages, etc. See THE : by Sron Hiway Gaee, of Co 7 pa aaae
T a octavo |

and ii ion i .00,
4 Duke St., Adelphitondon—England | COMSTOCE - Prios $2.00, postpaid.

University of Colorado
- Mountain Laboratory

Fourth summer session June 24 to ie T
gust 2, 1912. Courses” in Field Biology x

i Botany, Biology of Ponds and Streams. — 4

x (animal and plant), Ecology, Systematic ‘2 :

THE
AMERICAN NATURALIST

Vout. XLVI May, 1912 No. 545

THE CONTINUOUS ORIGIN OF CERTAIN UNIT
CHARACTERS AS OBSERVED BY A
PALEONTOLOGIST!

DR. HENRY FAIRFIELD OSBORN

RESEARCH PROFESSOR OF ZOOLOGY, COLUMBIA UNIVERSITY, CURATOR EMERITUS
OF VERTEBRATE PALEONTOLOGY IN THE AMERICAN MUSEUM OF
NATURAL HISTORY, VERTEBRATE PALEONTOLOGIST
UNITED STATES GEOLOGICAL SURVEY

Il. EVIDENCES ror CONTINUITY

Abandoning the historical background, we come to our
own subject, the origin and establishment in continuity
of characters which when established exhibit many of the
distinctive features of unit characters, namely, segrega-
tion, stability, pure heredity, and possibly, although this
has not yet been demonstrated, dominance and recession
im successive generations.

In fifteen previous papers of the writer beginning in
18893 the observation is repeatedly made that all abso-
lutely new characters which we have traced to their very
beginnings in fossil mammals arise gradually and con-
tinuously. One by one these characters, which are inde-
pendently changing in many parts of the organism, at
the same time accumulate until they build up a degree of
change which paleontologists designate as a ‘‘mutation’’

* Osborn, H. F., ‘‘The Paleontological Evidence for the Transmission
of Acquired Characters,’? AMER. NaTuRALIST, Vol. XXIII, No. 271, July,
2399, pp. 561-566.

249

250 THE AMERICAN NATURALIST [Vov. XLVI

in the sense of Waagen, who proposed this inter-specific
term in 1869. Finally they reach a sufficiently impor-
tant phase to designate the stage as a species.”°

These new characters were first (1891) termed ‘‘defi-
nite variations’’; subsequently (1907)? the term ‘‘recti-
gradations’? was applied to them. Rectigradation is
merely a designation for the earliest discernible stages
of certain absolutely new characters. It involves no
opinion nor hypothesis as to genesis; it is a simple mat-
ter of observation. Referring to the figure (p. 274) of
the upper grinding teeth of the horse, the majority of
the fourteen characters have been observed to arise as
rectigradations.

Quite different is the allometron. This is a new desig-
nation for the continuous change of proportion in an ex-
isting character which may be expressed in differences
of measurement. Since 1902 and especially during the
past year the behavior of allometrons has been very
carefully investigated by myself and by my colleague,
Dr. W. K. Gregory.

REcTIGRADATION—a qualitative change, the genesis
of a new character.

ALLOMETRON =a quantitative change, the genesis
of new proportions in an ex-
isting character.

The distinction between a rectigradation and an allo-
metron is readily grasped: when the shadowy rudiment
of a cusp or of a horn first appears it is a rectigradation ;
when it takes on a rounded, oval or flattened form this

2 This sentence may be contrasted with that of Punnett (op. cit., p. 15):
‘í Speaking generally, species do not grade gradually from one to the other,
but the differences between them are sharp and specific. Whence comes
this prevalence of discontinuity if the process by which they have arisen is
one of accumulation of minute and almost imperceptible differences? Why
are not intermediates of all sorts more abundantly produced in nature than
is actually known to be the case???
H. F., ‘‘Evolution of Mammalian Molar Teeth to and from
the Triangular Type,’’ 8vo, Macmillan Company, September, 1907.

No. 545] ORIGIN OF UNIT CHARACTERS 251

change is an allometron. In mammals rectigradations
are comparatively few; allometrons comprise the vast
number of changes in the hard parts. In the origin of
cusp and horn rudiments rectigradations are parallel
(see Fig. 3), in the changing proportions of a skull
allometrons are divergent (Figs. 1, 3).

Granting, without at present considering the evidence,”®
that these rectigradations and allometrons arise con-
tinuously through entirely unknown laws, that they are
blastic or germinal characters, the question’ arises, do
they become separable as unit or alternating characters
in heredity.

In general, paleontology furnishes quite as strong
proof as Mendelism or experimental zoology as to the
individuality, separableness, and integrity of single char-
acters in evolution. But, whether both rectigradations
and allometrons are separable in heredity can only be
demonstrated through experiments on cross breeding or
hybridizing. |

The special object of this Harvey lecture is to show
that certain at least of the rectigradations and allome-
trons observed in mammals are separable in heredity,
that they split up into larger and smaller groups or
units, some into partially blending units, others into ab-
solutely distinct or non-blending unit; finally that at
least in the first cross they exhibit dominance. .

The very important remaining question whether, like
the quality of ‘‘tallness’’ or ‘‘shortness’’ in Mendel’s
classic experiments on the pea, these allometrons con-
tinue to split into dominants and recessives in later
crosses, has not been investigated but is probably ca-
pable of investigation in mammals which do not become
sterile in the first hybrid generation. _

Five examples of the continuous evolution of recti-
gradations and allometrons may be cited, namely: —

3 This evidence is for the first time fully presented in the writer’s mono-
graph on the ‘‘Titanotheres,’’ in preparation for the U. 8. Geological

ey. P

252 THE AMERICAN NATURALIST [Vor. XLVI

1. Skull and horns of titanotheres (Figs. 1, 3, 4).

2. The horns of cattle (Fig. 2).

3. The cranium of man (Fig. 1).

4. The skull of horses (Figs. 4, 5, 6, 7).

5. Teeth (Fig. 8).

One of the most salient examples of the genesis of
unit characters through continuity is that of the evolu-
tion of horns, i. e., of the osseous prominences on the
skull. Horns are now known definitely to be ‘‘unit char-
acters,’’ first through their sudden and complete disap-
pearance in the niata and polled breeds of cattle; second,
because they conform to the laws of sex-limited inherit-
ance. The question is, do horns originate continuously
or discontinuously ?

ance

Brach.

Mes, ol.
Fig. 1. CONTINUOUS ORIGIN OF ALLOMETRIC ‘“ UNIT CHARACTERS ” IN THE
CRANIUM (A) AND SKULL (B) oF MAN AND TITANOTHERES.
A, Man Brachycephaly Mesaticephaly Dolichocephaly
B, Titanotheres Brachycephaly Mesaticephaly Dolichocephaly
(Paleosyops) (Manteoceras) (Dolichorhinus)

1. Horns of Titanotheres

The titanotheres are an extinct family of quadrupeds
distantly related to the horses, tapirs and rhinoceroses,

No. 545] ORIGIN OF UNIT CHARACTERS 253

to the evolution of which the author has devoted twelve
years of investigation, assisted by Dr. W. K. Gregory.
As set forth in an earlier contribution?’ the genesis of
horns as rectigradations has been observed in four or
five distinct phyla of titanotheres. These phyla descend
independently from a single ancestor of remote geologic
age. Both in respect to new cusps on the teeth and new
horn rudiments on the skull there is observed what in
our ignorance may be called an ancestral predisposition
to the genesis of similar rectigradations. This predis-
position betrays the existence of law in the origin of cer-
tain new characters; it recalls a sagacious remark of
Darwin:

. . . The principle formerly alluded to under the term of analogical
variation has probably in these cases often come into play; that is, the
members of the same class, although only distantly allied, have in-
herited so much in common in their constitution, that they are apt to
vary under similar exciting causes in a similar manner; and this would
obviously 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.” :

Briefly, the origin of the titanothere horns is as fol-
lows: (a) from excessively rudimentary beginnings, 1. e.,
rectigradations, which can hardly be detected on the sur-
face of the skull; (b) there is some predetermined law
or similarity of potential which governs their first exist-
ence, because (c) the rudiments arise independently on
the same part of the skull in different phyla at different
periods of geologic time; (d) the horn rudiments evolve
continuously, and they gradually change in form (4. €.,
allometrons) ; (e) they finally become the dominant char-
acters of the skull, showing marked variations of form
in the two sexes; (f) they first arise in late or adult
stages of growth, but are pushed forward gradually into

™<<The Four Inseparable Factors of Evolution. Theory of their Dis-
tinct and Combined Action in the Transformation of the Titanotheres, an
Extinct Family of Hoofed Animals in the Order Perissodactyla,” Science,
N. S., Vol. XXVII, No. 682, January 24, 1908, pp. 148-150. —

”<<Origin of Species,’’ Vol. II, p. 221.

Fic. 2, CONTINUITY E ONTOGENESIS OF THR HORN AND Horn SHEATH IN CATTLE IN Seven STAGES, 1-7. After preparations by Mr, 8S.
EL Chubb in the aaa A i American Museum of Natural History

1. Adult, 9 years, completed osseous horn and horny sheath. 2. Yearling, 18 months, sea Wwe shifting of osseous horn to occiput. 38. Calf, 2 months,
continuous shifting of osseous horn to occiput. 4. Calf, 2 weeks, Roope shifting of osseous horn to occiput. 5. Fatal stage, 9th month, bony swell-
ing, and epidermal swelling pointed, 6. Fetal stage ?6-7th month, epidermal swelling, paren with pointed hair tuft. 7. Foetal stage, ?5th month,
epidermal swelling, covered with 40 scattered hairs

No. 545] ORIGIN OF UNIT CHARACTERS 255

earlier and earlier ontogenic stages until they appear to
arise prenatally.

In the titanotheres (Fig. 3) the bony swelling is seen
at the junction of the nasals and frontals (black sha-
ding), in dolichocephalic skulls it appears chiefly on the
nasals, in brachycephalic skulls chiefly on the frontals.
Its original low, rounded shape is like that seen in the
ontogeny of the horns in eattle.

2. Horns of Cattle

The phylogenesis of the horns in titanotheres (Fig.
3) is exactly similar to the ontogenesis of the horns in
Bovidae (Fig. 2), in which the dermal rudiments first ap-
pear soon after the complete formation of the bones of
the skull in the unborn young, and the osseous rudiments
appear as rounded protuberances in the 8th month.

In the ontogenesis of horns in cattle three distinct ele-
ments are involved: (a) a psychic predisposition to use
the horn, (b) a dermal thickening over the bony horn
swelling which in ontogeny precedes the swelling, (¢) ap-
pearance of the bony swelling itself.

The ontogenesis is observed to be accompanied by a
marked allometric change in the skull which shifts the
horn backward from the side of the cranium to the side
of the occiput by the obliteration of the parietal bones.

3. Cranium of Man.

A third instance of continuous development is that of
the form of the cranium in man (Fig. 1), an allometric
evolution, or change of proportion, which is of especial
Significance because, according to the unanimous testi-
mony of anthropologists,*! head form is the result of
very gradual change either in the elongate (dolicho-
cephalic) or broadened (brachycephalic) direction.

* Ripley, Wm. Z., ‘‘The Races of Europe, a Sociological seas ae
D. Appleton & Co., 1899, 624 pp:

256 THE AMERICAN NATURALIST [Vor: XLVI

The matter is directly pertinent to the present discus-
sion because ‘‘long heads’’ and ‘‘broad heads’’ are con-
tinuously crossing and we know what the direct and ulti-
mate effects of such crosses are. The evidence has im-
portant bearing also on the influence of selection, environ-
ment, and inheritance or the effects of use and disuse.

Determination of the proportions of the cranium or
the cephalic index is one of the standard tests or race;
it is an expression of the greatest breadth of the head
above the ears and the percentage of its greatest length
from the forehead (glabella) to back, the latter measure-
ment being taken as 100. Three types adopted by anthro-
pologists are:

Extreme Range
Brachycophali¢, 80.1 and Shove .sise-sessoeridiser ire 100-80
Mosocephabe, T9180 (oo oes Seas cases es tse eee 80-75
Detiehocepialis, Td aud below 6. a. 5s i ek wees 75-62

Among the present races of Europe the widest limits
of variation between brachycephaly and dolichocephaly
are in the averages between 73 and 87; individual ex-
tremes of 62 and 100 have, however, been observed.
These extremes in European head form do not coincide
either with geographic or political boundaries, but are
attributed to the entrance into Europe of brachycephalic
and dolichocephalice types which evolved in Asia. Simi-
larly among the aborigines in America the indices range
from a low dolichocephaly as among the Delaware, Pima
Indians, etc., to a decided brachycephaly as among the
Athabascan tribes in Panama, Peru, and other localities.
A significant fact in Europe is that dolichocephaly and
brachycephaly are extremely stable characteristics in
heredity. The significant fact in America is that through
a very long period of time the various races of Indians,
who are believed to have had originally a similar origin,
have acquired under conditions of geographic isolation
considerable diversity in the proportions of the head.
Similarly A. Keith?? from the present distribution of the

2 Keith, A., Journ. Royal Anthropological Institute, 1911. See Nature,
Vol. 88, No. 2195, November 23, 1911, p. 119.

No. 545] ORIGIN OF UNIT CHARACTERS 257

Negro tribes in equatorial Africa has reached the follow-
ing conclusions:

There has been free intermigration; in the course of their evolution,
the tendency of one tribe has been towards the accentuation of one set
of characters, of another towards another set. Thus the Dinka ac-
quire high stature and narrow heads; the typical Nigerians low stature
and narrow heads; the Basoko wide, short heads and low stature; the
Buruns wide heads and high stature. Interbreeding may have played
its part; but if it had played a great part we should have found
greater physical uniformity than there is. The influence of Arab blood
on these tribes has probably been exaggerated.

It appears that environment has not any direct in-
fluence on head form, but that geographical isolation
affords the several varieties of man as well as other
mammals a chance to develop their peculiar head charac-
ters. Elliot Smith states (letter, August 12, 1911):

In my opinion the conditions of dolichocephaly and brachycephaly
must have developed very slowly through exceedingly long periods of
time and in widely separated areas amidst widely different environ-
ments. Brachycephaly is especially distinctive of the Central Asian
high plateau populations, dolichocephaly of the littoral and plain-
dwelling peoples; but these “unit characters” are now so fixe that
environment is powerless to modify them in a thousand years or so.
...1 do not believe for a moment in Boas [that is, in Boas’s observa-
tions (1911) on the rapid influence of environment in modifying head
form].

Elliot Smith takes very strong ground as to the lack of
evidence that environment directly produces any modi-
fication of head form; he implies that such modification,
if natural, would only show itself after thousands of
years of residence; environment no doubt has indirect in-
fluence. Hrdlička, on the other hand, has obtained defi-
nite results in the influence of environment on the vault
and face form of the Eskimo;** it remains to be shown
how far these changes are ontogenic. The recent con-
clusions of Boas (1911)* that dolichocephaly and brachy-

2 Hrdlička, Ales, ‘‘Contribution to the Anthopology of Central and
Smith Sound Eskimo,’’ Anthr. Paper Am. M. N. H., V, Pt. TI, 1910, p. 214.

“ Boas, Franz, ‘The Mind of Primitive Man,’’ 8vo, Macmillan Com-
pany, New York, 1911, 924 pp.

258 THE AMERICAN NATURALIST [Vou XLVI

o>.

P H M

Fig. 3. ue AND ALLOMETRONS IN TITANOTHERES, Continuity
in pa Se oe a seous horns in titanotheres. P =ł2d lower gapen
H= of na and — (shaded) showing osseous horn; S = S;
M= ae i a i bon

nope lay a oo (dolichocephalic) titanothere.
, a

Il. Paleosyops, a broad-headed (brachycephalic) titanothere. |
I. Eotitanops, an ancestral (mesaticephalic) titanothere.

II-V belong A four eperen phyla which Reon in their allometric
evolution of head (8) d foot si een on (M) but give rise to independe:

pans ae e pong the o
us horn

usps on Pt premolar teeth (P) and of
rudiments (H) on the seat

No. 545] ORIGIN OF UNIT CHARACTERS 259

cephaly are congenitally altered by environment in the
first generation are modified by his statement that this
action in bringing diverse head forms together would not
go so far as to establish a uniform general type.

No anthropologist has offered any satisfactory expla-
nation as to the adaptive significance of dolichocephaly
or brachycephaly. It is well known that these differences
of head form are not associated with intellectual ability
or mental aptitude. Boas writes (April 8, 1911):

So far the matter is very perplexing to me. I feel, however, very
strongly with you that changes in type are very liable to be progres-
sive in definite directions. . . . To my mind it seems no more diffieult
to assume that this predetermined direction should continue from
generation to generation than to make the much more difficult assump-
tion that notwithstanding. all internal changes the egg-cell of one
generation should be absolutely identical with that of the preceding

generation.

Apart from the disputed question of the direct influence
of environment and of human selection there is absolute
unanimity of evidence and of opinion on the one point
essential to the present discussion, namely, as to the con-
tinuity of allometric variation which establishes different
extremes of head form under conditions of geographic
isolation. ;

Granted that these extremes evolve continuously, do
they become discontinuous in heredity? |

One of the general results of crossing long-headed and
narrow-faced types with broad-headed and broad-faced
types is what is known as disharmonic heredity, namely,
that condition in which the face and cranium do not hold
together, but broad faces may couple with long skulls, or
vice versa (Boas, 1903).*° Boas concludes that there can
be no question that the mixture of a long-headed and of
a short-headed race may lead to disharmonism, one race
contributing head form, the other facial expression.

As to stability or segregation in heredity the latest

3 Boas, Franz, ‘‘ Heredity in Head Form,’’ Amer. Anthropologist, Vol.
5, No. 3, July-September, 1903, pp. 530-538.

260 THE AMERICAN NATURALIST [Vou XLVI

opinions of Boas, Elliot Smith and Hrdlicka have been
sought. Boas is one of the most positive as to the hered-
itary stability of head form. He observes (1911, pp.
7-9):

Among European peoples head proportions are considered among
the most stable and permanent of all characteristics. In intermarriage
of “ dolichocephalice ” and “ brachyeephalic ” individuals the children
do not form a blend between their parents but inherit either the

dolichocephalie or brachyeephalie head form. Head form thus con-
` stitutes a case of almost typical alternating heredity (p. 55). No evi-
dence has been obtained, however, to show that either brachyeephaly
or dolichoecephaly is dominant. Children exhibit one head form or the
other, and the cephalie index or ratio of breadth to length undergoes
only slight alteration during growth, or ontogeny.

Elliot Smith (letter of August 12, 1911) is ‘‘firmly
convinced that the form of cranium, orbits, nose, jaws,
limb bones, etc., in the ‘Armenoid’ and ‘Proto-Egyptian’
series are very stable or even fixed ‘unit characters’
which do not really blend, but that certain elements of
mosaic assemblage of characters pe be grafted on to
others belonging to the other race.’

Opinions as to Blending.—It will be noted that Boas
(1895) admits a certain blending of head form in crosses.
Hrdlička (letter, November 1, 1911) speaks even more
guardedly as to the hereditary stability of head form.
He says:

As to the head form constituting a “ unit character ” which does not
blend in intermixture, I am not able to give a conelusive opinion, but
my experience and other considerations lead me to be very skeptical
that such is the case to any great extent. The subject is a very com-
plex one and requires considerable direct investigation in different lo-
calities and with different peoples before the exact truth can be known.

. As to the statement that long or broad head form is a stable or
unit character not blending in intermixture, I think that only the first
part of the proposition may be held as fairly settled. But even then I
should change the word “stable” to “ persistent,’ and qualify the
phrase by adding “under no greatly differing and lasting environ-
mental conditions.”

That prolonged interbreeding or intermixture tends
to break down the stability of hereditary head form is

No. 545] ORIGIN OF UNIT CHARACTERS 261

indicated by Boas, Elliot Smith, and Ripley, as well as
by Hrdlicka, as quoted above. Thus Ripley (1899), p.
55) observes: —

The plotting of cephalic indices on a map of Europe shows that
there is a uniform gradation of head form from several specifie centers
of distribution outward.

In Italy over 300,000 individuals taken from every
little hamlet have been measured. In the extreme south
we find the dolichocephalic head form of the typical —
Mediterranean race; the type changes gradually as we
go north until in Piedmont we find an extreme of brachy-

Brontother.™

a Eohippus

Fig. 4. CONTINUOUS ORIGIN or ALLOMETRIC “ UNIT CHARACTERS ” IN THE
SKULL OF VARIOUS UNGULATES.
© Cytocephaly, Bubalis Rangifer.
D Dolichocephaly, Opisthopic : Pr
(Titanotheres) (Equines).
In the ancestral Eotitanops and Eohippus the facio-cranial index is very
similar. In the descendants of these Is, as i ed by the dotted lines,
the facio-cranial indices are widely divergent; in the Titanotheres (Bronto- —
therium) the cranium is elongated; in the horses (Equus) the face is elongated.

262 THE AMERICAN NATURALIST [Vou. XLVI

‘cephaly of the Alpine type, recalling the broad-headed
Asiatic type of skull. Thus (Ripley, p. 56) ‘‘pure phys-
ical types come in contact and this means ultimately the
extinction of extremes.” Applying these principles to
the present case, it implies the ultimate blending of the
long and the narrow heads and the substitution of one
of medium breadth.

Elliot Smith also (letter, August 12, 1911) implies a
gradual modification or blending of head form through
prolonged intermixture. He observes:

Egypt does not give a clear answer to your queries because her ex-
ceedingly dolichocephalie brown race [related to the Mediterranean
race of southern Europe] underwent a double admixture (cirea
3,000 3.c.) with moderately brachyeephalie “ Armenoids” from Asia
and dolichocephalie Negroes from Africa. The Mediterranean Egyp-
tians are on the whole a little broader-headed than they were 6,000
years ago, and this may be due in part to a slow development toward
mesaticephaly; but it is mainly the result of an admixture with alien
bracephalies and mesaticephalics. There is an unquestionable tendency
toward the elimination of the extremes of narrowheadedness and
broadheadedness.

Hrdlička (letter, December 5, 1911) observes:

As to the effect of the mixture of brachycephalic and dolichocephalic
individuals or peoples, I am led to believe that there is in the results
of such mixtures a large percentage of more or less intimate “blend ”
of the two forms, for such a condition is indicated by the curves of
distribution of the cephalic index among such national conglomerates
as the French, Germans, different tribes of the American Indians, ete.
These curves, if sufficiently large numbers of individuals have been ex-
amined, all approach more or less the ideal camel-back curve. If no
“blend” existed, we should be bound to get the double or dromedary-
back eurve. Of course the effects of mixture and the effects of environ-
ment are with our present means often impossible of separation, they
often obseure each other. Yet the indications are that there is gener-
ally a considerable amount of more or less mixture of the many ele-
mentary constituents of the hereditary characters [known collectively
as] dolichocephaly and brachycephaly.. With this there coexists doubt-
less some tendency toward a differentiation into the two opposite forms
of the head.

Thus in human head form we have proofs of continuous

No. 545] ORIGIN OF UNIT CHARACTERS 263

allometric change strictly comparable to that which oc-
curs in the crania of lower mammals, especially as ob-
served in the horses and titanotheres; the extremes are
produced in so-called pure human races under geographic
isolation; when these pure races are brought together
there arises disharmonism or alternating heredity or
both. Neither the dolichocephalic nor brachycephalic
type is as yet known to be dominant; opinion is divided
as to whether in the first cross the heredity is pure or
whether there may be a tendency to produce an inter-
mediate form; opinion is nearly unanimous that pro-
longed interbreeding produces blends.’

4. Skull of Titanotheres.

The continuity of allometric evolution in the skull of
the titanotheres (Fig. 4) has been the subject of pro-
longed investigation by the writer, assisted by Dr. W.E
Gregory, involving thousands of measurements, many of
which belong in strictly successive phyletic series.
Allometry (i. e., the measurement of allometrons) here
applies to the skull as a whole. We secure the cephalic
index by dividing the breadth across the cheek arches by
the total basilar length of the skull. There are also other
indices, such as the facio-cranial, in which we measure
continuous trends of allometric change; brachycephaly
and dolichocephaly arise independently in four different
phyla or lines of descent. The adaptive significance is
sometimes apparent, sometimes obscure. As shown in
Fig. 1 the titanotheres, like man, exhibit facial abbrevia-
tion and cranial elongation (postopic dolichocephaly) in
contrast with the facial elongation (proopice dolicho-
cephaly) of the horses. These phenomena are similar
to those of cytocephaly, or the bending down of the face
upon the base of the cranium as observed in the reindeer

T. H. Morgan observes that a blend may occur in the first generation,
F,, even where perfect segregation occurs in F.. The results of crossing
the equine skull as described below indicate a tendency to blend in the
first cross.

264 THE AMERICAN NATURALIST [Vowu. XLVI

(Rangifer) and the hartebeest (Bubalis). Cytocephaly
is an ontogenetic and phylogenetic new character, aris-
ing or developing continuously.

As in the case of the human skull; the causes of these
profound changes in head form are entirely unknown;
the mechanically adaptive significance is sometimes ap-
parent, sometimes obscure. The evidence is strengthened
by the examination of the titanotheres that human selec-
tion has little or no influence on human cranial form.
The great point to emphasize is that this allometric evo-
lution in the skull and all parts of the skeleton is the pre-
vailing phenomenon of change in the skeleton of mammals.
It is constantly in progress and is universally, so far as
we can observe, a continuous process. As displayed in the
four phyla of titanotheres (Fig. 3), the elongation or
broadening of the foot bones proceed independently and
are divergent, while in the same mammals the rectigra-
dations exhibited in the rise of similar cusplets on the
teeth and similar horn rudiments on the face are parallel;
in the former ease no ancestral predisposition seems to
be operating, in the latter case ancestral predisposition
certainly seems to operate; this is why the internal laws
controlling the origin of new allometrons and of new
rectigradations and allometrons are regarded as essen-
tially dissimilar.

Paleontological analysis of these rectigradations and
allometrons even unaided by experimental heredity re-
veals the essential feature of the ‘‘unit character’’ prin-
ciple, namely, that what we are observing is. an incredibly
large number of unit elements each of which enjoys a
certain independence of evolution at the same time that
each unit is adaptively correlated with all the others. For
example, in the upper and lower grinding teeth of horses
alone there are 504 cusp units, each of which has an inde-
pendent origin and development; at the same time each
cusp is more or less distinctly correlated in form with
the all-pervading dolichocephaly or brachycephaly of the
skull; in fact, from certain single cusps of the teeth we

No. 545] ORIGIN OF UNIT CHARACTERS 265

can often determine whether the animal is brachycephalic
or dolichocephalic

N:P,

Monophyletic

-= _

opera

MULE

arta wes trne NAg
Polyphyletic RSE
IMPERFECT BLENDING OF

CROSS-BREEDING AND
HE FacraAL BONES IN ASS (MALE),

5. ALLOMETRIC “ UNIT
CHARACTERS ” OF TH Horse (FEMALE) AND
MULE.

Bones of the side of the face, Ass.

Bones of the side of the face, Mule.
Bones of the-side of the face, Horse.
The horse is oo polyphyletic,
arrow points to C, a distinct bum
ass, “ihe

the ass is probably sonm e
ho

k point at ‘wich feast De of

F.P

oO. The
rse and mule, not observed in the
the nasals is i L =la

266 THE AMERICAN NATURALIST [Vou. XLVI

The question arises as a result of the somewhat con-
flicting evidence as to the crossing of brachycephals and
dolichocephals in man, what happens when we cross two
phyla of lower mammals which have been diverging along
separate allometric lines and in the meantime have ac-
quired a greater or less number of new characters which
when sufficiently developed attain specific rank.

The answer is given very distinctly in the cross between
the dolichocephalic horse (E. caballus).and the meso-
cephalic ass (E. asinus). Here we learn again that pro-
found differences have been established through con-
tinuity and that we are enabled to split up these differ-
ences into distinct or partially blending units through
cross breeding.

5. Blended or Alternating Heredity in Horses.**

So high an authority as J. Cossar Ewart (1903) has
sustained the prevailing view that in the mule there is
generally an imperfect blending of the characters of the
immediate parents; the same author, however, notes that
mules occasionally serve as examples of unit or exclusive
inheritance. He cites two cases: (1) a mule out of a
well-bred, flea-bitten New Forest pony closely resembles
her sire, the ass; (2) a ‘‘calico’’ mule, on the other hand,
is surprisingly like his dam, an Indian ‘‘painted’’ pony.
This painted mule demonstrates that the ass is not always
more prepotent than the horse. From this author’s very
extensive breeding experiments the following conclusions
are reached: the less fixed or racially valuable characters

* The writer is indebted to Mr. S. H. Chubb, Mrs. Johanna Kroeber
Mosenthal and to ee W. F Gregory for many of the observations and all
of the measurem ich this comparison is based. The materials
studied are three pe of fe ass (¢ E. asinus), ten of the horse (9 E.
caballus), and four of the mule, all adult with teeth in approximately the
same stage of we

3 The most resent (1912) opinion of Ewart is much more gues as to-
the operation of Mendel’s law in pure breeding strains of

horses. See
‘t Eugenics and the Breeding of Light Horses,’’ The Field, kia 10.
1912, pp. 288, 289.

No. 545] ORIGIN OF UNIT CHARACTERS 267

of zebras either blend with or are dominated by the cor-
responding characters in their horse and ass mates,
Thus, as influencing dominance or prepotency, the value
which a character has attained in the past struggle for
existence seems to count for something. In zebras and
in horses certain physical and mental traits are more
highly heritable than others. Among the characteristics
which are often handed down unblended in zebra-horse
hybrids and to a less extent in zebra-ass hybrids are the
size of the ears, the form of the hoofs, the massiveness of
the jaws; while among psychic characters are transmitted
the extreme caution, the wonderful alertness and
quickness.

The new results brought forward in this Harvey lecture
from the comparison of the skull and teeth of the horse,
ass and mule on the whole strengthen the theory of unit in-
heritance both in rectigradations and inallometrons. The
measure of unit character inheritance as contrasted with
blended inheritance is very precisely brought out in the
detailed study of the twenty-two characters which are
examined below. Before discussing these characters in
detail it is interesting to point out that the ancestors of
the horse and the ass have probably been separated for
at least 500,000 years. In the meantime the horse has
become extremely dolichocephalic, the ass has remained
comparatively mesocephalic; the horse has a relatively
long, the ass a relatively short face; the horse has highly
complex, the ass has somewhat simpler grinding teeth;
the horse exhibits advanced adaptation to grazing habits
and has become habituated to a forest and plains life in
comparatively fertile countries, while the wild ass is by
preference a browsing animal, finding its food in exces-
sively arid countries where there is a marked dearth of
water and water courses. The physical and psychical
divergences in these two animals have developed over an
enormously long period of time. Every single tooth and
bone of the horse and ass show differences both in recti-
gradation and in allometric evolution.

268 THE AMERICAN NATURALIST [Vor. XLVI
One feature which tends to make the results of the
cross less clear and distinctive than they are is that

while the ass is monophyletic (being descended with

S
ASS
S
MULE
S
HORSE |

Fic, 6. CROSS-BREEDING AND IMPERFECT BLENDING OF SUB-ALLOMETRIC “ UNIT
CHARACTERS ” OF THE NASAL BONES IN ASS (MALE) AND HORSE (FEMALE).
view of nasals and naso-frontal suture, Ass.
Top view of nasals and naso-frontal suture, Mule.
Top view of nasals and naso-frontal suture, Horse.
8 = point of section shown in Figs. 3 and 5

No. 545] ORIGIN OF UNIT CHARACTERS 269

modification from the wild E. asinus of northern Africa),
the domestic horse is not a pure strain and is certainly
polyphyletic, having in its blood that of several races,
such as the Arab and the Forest or Norse horse, animals
which have specific distinctness although they still inter-
breed.®® To this mixed strain or polyphyletic heredity of
the horse, are probably attributable many of the allometric
variations in the bones of the skull and in the enamel
pattern of the teeth of the mule in some of which we
observe a nearer approach to the ass type than in others.
If we could cross the ass with a pure horse race like the
Steppe or Prjevalsky horse we should probably obtain
more precise results. Another disturbing feature in the
comparisons and indices given below is that we do not
know the exact structure of the skull of either of the
parents from which the mule skulls examined were
derived.

Despite these sources of fluctuation and of error, the
general results obtained are fairly positive and definite.

The first point of interest in the segregation of unit
characters in the mule is that connected with the three
germinal layers, namely, the epiblast, mesoblast and hypo-
blast. All the characters of epiblastic origin appear to be
derived from the sire, namely, the epidermal derivatives,
the distribution of the hair, especially in the mane and
tail, the hoofs, ete., are those of the ass, although the color
pattern, as in the ‘‘calico’’? mules described by Ewart,
may be derived from the mare. The nervous system and
psychie tendencies, all of epiblastic origin are also
derived from the ass, including minor psychic character-
isties, such as aversion to water. Still more striking,
perhaps, is the fact that the enamel pattern of the grind-
ing teeth, again of epiblastic origin, is mainly that of the
ass, although, as shown below, there are some inter-
mediate and some distinctive horse-like characters in the

* There are many absolute charactérs which separate the Arab from the _

Norse horse, among them the invariable presence of one less vertebra in the ee ae

lumbar region of the back. ee

270 THE AMERICAN NATURALIST [Vow XLVI

teeth of the mule; this may be partly connected with the
mesoblastic derivation of the dentine of the teeth.

Monophyletic

MULE

Polyphyletic
CROSS-BREEDING AND raglan SEPARATION OF ALLOMETRIC “ SUB-

Fig. 7.
— Cnanscrres ” OF THE NASAL BONES IN ASS (MALE), HorSE (FEMALE)

Mid-section of nasal bones, Ass.
Mid-section of nasal bones, Mule.
Mid-section of nasal bones; Horse.

Y1, Y? = variations in the depth of pa — in the mule. V1, V? = varia-
tions ta the depth of the nasals in the ho

Mesoblastic derivatives, on the other hand, are divided
between the sire and dam, the skeleton and limbs of the
mule being mainly proportioned as in the ass, while the
skull of the mule, as we shall see, is almost purely that
of the horse.
Blended and Pure Inheritance in the Bones of the Face

Blending—A comparison of the bones of the side of

No. 545] ORIGIN OF UNIT CHARACTERS Zil

the facial or preorbital region shows intermediate or
partly blended form and proportions both of the nasals,
premazxillaries, frontals, and lachrymals, in which, how-
ever, the mule approaches E. caballus rather than E.
asinus. Attention may be called to some of the details
of the comparison: (1) Suture between the nasals and
premaxillaries: in E. asinus short and elevated, in the
mule intermediate but more like the horse; in the horse
elongated and depressed (see Fig. 5). (2) Naso-frontal
suture on the top of the skull: in the ass straight or trans-
verse; in the mule incurved, more like the horse than the
ass; in the horse arched or incurved (see Fig. 6). (3)
Depth and convexity of the nasals: in the ass shallow and
flattened; in the mule deeper, more like the horse; in the
horse highly arched. (4) Bump or convexity on posterior
third nasals: in the ass very slight; in the mule moderate,
more like the horse than the ass; in the horse strong (see
Fi. y).

The same tendency in the mule to exhibit a slight de-
parture from the horse toward the ass type is shown in
the outlines of the bones of the face (Figs. 3, 4, 5). Com-
paring step by step the premaxillaries, maxillaries,
nasals, and lachrymals, while the proportions and the
sutural outlines are mainly those of the horse, there is a
more or less distinct blending, or intermediate character
in the direction of the ass; see especially the naso-pre-
maxillary suture, the degree to which the nasals extend
downward on the sides of the face to join the maxillaries,
and the degree to which the nasals extend on the sides of
the face to join the maxillaries. In this naso-maxillary
junction certain horses approach the ass type. The char-
acteristic bump on top of the nasals of the horse is trans-
mitted to the mule, and the highly characteristic trans-
verse suture between the frontals and the nasals, as seen
from the top (Fig. 4), is rather that of the horse than of
the mule.

Non-blending.—More definite results are shown in the
heredity of the indices or ratios between the various por-

1. Cephalic Index:

272 THE AMERICAN NATURALIST [Vowu. XLVI

tions of the skull and of the teeth; these indices are ex-
tremely constant allometric specific characters, they are
independent of size. For example, the indices of a
diminutive pony and of a giant percheron would be the
same. Similarly the indices of a diminutive donkey and
of a very large ass would be the same.

The index is the best and most exact form of express-
ing mathematically the profound differences between the
skull of the horse and that of the ass. Indices have the
value of specific characters; they are of especial signifi-
cance in the present discussion in comparison with those
in the face, cranium and palate of man and of the titano-
theres above considered.

Chief among the allometric differences are the follow-
ing: (1) Inits proportions the ass has a relatively shorter
space between its grinding and its cutting teeth, the
bit-opening; this is correlated with the fact (2) that the
ass has a relatively broader and shorter skull than the
horse; also with (3) the fact that the ass has a relatively
longer cranium (postorbital space) and shorter face
(preorbital space) than the horse; (5) the ass also has
relatively broader grinding teeth correlated with the
broader skull; (6) correlated also with its less elongate
skull the ass has a relatively rounder orbit than the
horse, 7. e., the vertical and horizontal diameters are more
nearly equal. (7) A very distinctive feature is the angle
which the occiput makes with the skull; this is one of the
marked specific features of the ass.

NoN-BLENDING OR PURE INHERITANCE INDICES IN THE SKULL

Width of skull X 100 Ass 46,9-49.9

Basilar len gth

Diast
2. Diastema Index: echoed i at ee
Basilar length of skull

Length i
3. Cranio-facial Index: mgri of crantem X 190

Length of face

Mule 40.8—43.6
Horse 40.4—44.1

Ass 15.6-17.6
Mule 18.6-21.9
Horse 18.2—23.0

Ass 56.3—61.0
Mule 48.9-51.8
Horse 45.3—49.9

4,

5.

6.

f

No. 545] ORIGIN OF UNIT CHARACTERS 273

Vertical diameter of orbit X 100 Ass 96.0-104.2
> Mule 78.7— 99.1

Horse 84.2— 93.5

Orbital Index :

: Ass 15.2-16.0
Transverse diameter of M? X 100
Molar Index: iia ee ~ sage Mule 14.2-14.9

otal length of entire molar series Horse 13.9-15.7

Angle between vertex of skull and line Ass 52.5-60.0
connecting most posterior points of Mule 61.0-66.5
Occiput-vertex occipital crest with condyles, Horse 64.0-76.5
angle Index: i. e., nearly all horse skulls will stand
when set up on end, some mule skulls
(one out of four), no ass skulls

=,

Distance from palate to posterior end of Ass 93.8-111.7
Vomer Index: vomer X 100 Mule 95.5-110.3

Distance from vomer to foramen magnum Horse 72.8- 86.5

The above indices prove that the mule has not a primi-
tive skull like that of the ass on a larger scale, but has
essentially the skull of the horse, namely :

1. A long, narrow skull, as a whole.

2. A long diastema, or space for the bit.

3. A short cranium and a long face.

4. A long, oval orbit.

5. A relatively elongate and narrow set of grinding
teeth.

6. A vertically placed occiput.

The one character in which the mule resembles the ass
is the elongation of the vomer behind the bony palate. It
should, however, be distinctly stated that while the indices
given above are those which probably prevail in mules,
there are overlaps in the (4) orbital index and (6) occi-
put-vertex angle. Thus in one mule the orbital index
agrees with that of one of the asses.

Enamel Pattern of Grinding Teeth.—In the marvel-
ously complex pattern of the grinding teeth the ‘‘unit
character” transmission is quite sharply defined in the
majority of characters, while intermediate or slightly
blended in the minority. In general in the grinding teeth
of the ass the main enamel folds are less complicated

274 THE AMERICAN NATURALIST [Vou. XLVI

than in the horse and there are fewer secondary or sub-
sidiary folds; the ass especially lacks the ‘‘pli caballin’’
(fold 5) which is usually a very pronounced specific
character of the horse. The mule shows a very slight
indication of this fold and thus resembles the ass. The

folds: (1,3,4) ASS

folds: (1,2,3,4) MULE

folds: (1,2,3; 4,556) HORSE
ROSS- PRORDING AND SEPARATION OF espadana aes DISTINCT
N THE ENAMEL FOLDINGS AND PATTERN OF THE GRINDING
A o

OF Section through the crown a the third
eau ainia (p* or 4th premolar) ass (male), horse (female) and mule.

No. 545] ORIGIN OF UNIT CHARACTERS 275

subsidiary folds in the grinders of the mule are simpler
than those in either the horse or the ass. The grinder
of the mule would be pronounced by any systematist not
knowing its mixed parentage to belong to the ass rather
than to the horse, especially in the absence of the ‘‘pli
eaballin’’ (fold 5), in the form of the hypostyle (hs, fold
6), in the smaller size of the protocone (pr), the large
size of which is very distinctive of the horse. A very
detailed study and comparison of the grinding teeth in
the horse, ass and mule made by an independent ob-
server, Dr. W. K. Gregory, gives the following result:

Secondary folds: protocone, pr.
paracone, pa
metacone, me
hypocone hy.
protoconule pl.
metaconule ml.
Secondary elements : parastyle, ps.
ostyle, ms
hypostyle, hs
Primary elements:
fold (ie es Horeca. <2 Male . 3.55% Ass
Told 2 Ue ees Se cea Mule.
fold: 3 i428; Horse =: Males. Ass
Tod é eo oe Horie 3005), jo hd eres ASS
Tod Oo es Hor

UNIT CHARACTERS IN GRINDING TOOTH OF THE MULE

Distinctly ass-like: 5 character

say ‘ tal peculiar to ass.
Less distinctly ass-like: 6 characters

Common to horse and ass: 5 characters 5 common to horse and ass.
Distinctly horse-like: 2 characters a samdir to horse.
Less distinctly horse-like: 4 characters } r

It would be especially desirable to compare the same
enamel characters in the hinny, which is a cross be-
tween the male horse and the female ass, in which it is
well known that the E. caballus and E. asinus char-
acters are differently distributed.

Summary.—Out of the 28 characters examined in the
skull and teeth of the mule, 18 are distinctly derived
either from one parent or the other with very slight, if
any, tendency to blend, 10 characters show a distinct
tendency to blend.

276 THE AMERICAN NATURALIST [Vou. XLVI

This evidence, in the opinion of T. H. Morgan, is in
entire accord with the modern views of hybridizing;
parallels for each instance can be given; without the
evidence of the F, generation no conclusions adverse
to Mendelism are possible. Even the differences in re-
ciprocal crosses, i. e., horse ĝ, ass 9, can be understood
if sex-limited inheritance prevails in some characters.

To confirm the results suggested by this F, genera-
tion of the horse and ass, it would be necessary to inter-
breed races of mammals to F, or F, to ascertain whether
these characters of the skull and teeth really mendelize.
It is doubtful whether such specific types of mammals can
be found, and whether sufficient stability of character
exists in artificially produced races.

Sufficient evidence has been adduced, however, to show
that a very large number of characters which are to the
best of our knowledge of continuous origin, present all
the appearance of ‘‘unit characters?’ in the first genera-
tion of hybrids.

II. Conciuston

Is it not demonstrated by this comparison of results
obtained in such widely different families as the Bovide,
Hominide, Titanotheriide and Equide that discontinu-
ity in heredity affords no evidence whatever of pages
tinuity of origin?

As to origin, it is demonstrated in paleontology that
certain new characters arise by excessively fine grada-
tions which appear to be continuous. If discontinuities
or steps exist they are so minute in these characters as to
be indistinguishable from those fluctuations around a
mean which seem to accompany vey stage in the evolu-
tion and ontogeny of unit ch

IV. THEORETICAL CONSIDERATIONS

After having attempted to confine our discourse to
facts it is a pleasure to relax into the more genial at-
mosphere of opinion and hypothesis.

No. 545] ORIGIN OF UNIT CHARACTERS Zit

The principle of pre-determination, which results in |
the appearance of rectigradations, involves us in radical
opposition to the opinions of the Bateson-DeVries-
Johannsen school. There is an unknown law operating
in the genesis of many new characters and entirely dis-
tinct from any form of indirect law which would spring
out of the selection of the lawful from the lawless. This
great wedge between the ‘‘law’’ and the ‘‘chance’’ con-
ception, which since the time of Aristotle has divided
biologists into two schools of opinion, is driven home by
modern paleontology.

Paleontology, in the origin of certain new characters
at least, compels us to support the truly marvelous phil-
osophie opinion of Aristotle, namely:

Nature produces those things which, being continu-
ously moved by a certain principle contained in them-
selves arrive at a certain end.

While recent biology has tended to sharply distin-
guish bodily from germinal processes and to place chief
emphasis upon evolution appearing to originate in the
germ cells, we must not forget that for a hundred million
years or more,,or from the beginning of life, the germ
plasm has had both its immediate somatic and its more
remote environmental influences. Because the grosser
form of Lamarckian interpretation of transmission of
acquired characters has apparently been disproved, we |
must not exclude the possibility of the discovery of finer,
more subtle relations between the germ plasm and the
soma, as well as the external environment. There are
several phenomena, which have been observed only in
paleontology, that afford evidence for the existence of
such a nexus; because it appears that certain germinal
predispositions to the formation of new characters, con-
nected, as Darwin conjectured, in some way with com-
munity of descent, are only evoked under certain so-
matic and environmental conditions, without which they
appear to lie in a latent, potential or unexpressed form.

All that we may be able to observe are the modes of

278 THE AMERICAN NATURALIST [Vou. XLVI

operation in the genesis of new characters and in the
adaptive trends of allometric evolution without gaining
any intimate knowledge of what the causes are. This
thought may be made clear through the following anal-
ogy. Naturalists observed and measured the rise and
fall of the tides long before Newton discovered the law
of gravitation; we biologists are simply observing and
measuring the rise and fall of the greater currents of
life. It is possible that a second Darwin may discover
a law underlying these phenomena bearing the same re-
lation to biology that the law of gravity has to physics,
or it is possible that such law may remain forever un-
discovered. Another analogy may make our meaning
still clearer. Ontogenesis is inconceivable, for example,
the transformation of an infinitesimal speck of fertilized
matter into a gigantic whale or dinosaur; we may watch
every step in the process of embryogeny and ontogeny
without becoming any wiser; in a similar sense phylo-
genesis may be inconceivable or beyond the power of
human discovery. Not that we accept Driesch’s idea of
an entelechy or Bergson’s metaphysical projection of
the organic world as an individual, because we must be-
lieve that the entire secret of evolution and adaptation
-is wrapped up in the interactions of the four relations
that we know of, namely, the germinal, the bodily, the
environmental, with selection operating incessantly as
the arbiter of fitness in the results produced. In the
meantime*” we paleontologists have made what appears
to be a substantial advance in finding ever more con-
vineing evidence of the operation of law rather than of
chance in the origin and development of new characters,
something which Darwin had clearly in mind.*!

* Osborn, Hick. “Ths peta Mechanism and the Search for the
Unknown Factors of Evolution,’’ Biol. Lect. Marine Biol. Lab., 1894,
AMER. N ON Vol. XXXIX, No. 341, May, 1895, pp. 418—439.

“1 Darw. has.: ‘‘I have spoken of vyarlations sometimes as if they
were due i dha This is a wholly incorrect expression; it merely serves

to acknowledge plainly our ignorance of the cause of each particular
variation. ’’

THE BIOLOGY OF THE CRAYFISH

F. E. CHIDESTER
RUTGERS COLLEGE

INTRODUCTION

Tue first reference to the crayfish in scientific litera-
ture is in Aristotle’s ‘‘History of Animals,’’? where he
speaks of the ‘‘small Astaci which breed in the rivers.’’

Aristotle and the older naturalists used the term Astaci
to include both the crayfish and the lobster.

Faxon divides the crayfish into two great groups (24):
One, restricted to the northern hemisphere, is found
in Europe, Asia and North America. The other is found
in the southern hemisphere, in Australia, Tasmania, New
Zealand, Fiji Islands, Madagascar and South America.

The islands now inhabited by crayfish, such as Eng-
land, Japan and Cuba, were probably once connected
with the mainland.

In speaking of the distribution of the crayfishes, Faxon
Says:

The northern family of crayfishes contains two genera, Astacus and
Cambarus. These groups occupy distinct geographical areas. The
genus Astacus is found in the old world in Europe and western Asia
as far as the Aral and Caspian Seas, and in America in the region west
of the Rocky Mountains, draining into the Great Salt Lake and the
Pacifie Ocean. It is thus seen to occupy the western sides of the two
northern continents. Cambarus is found in North America east of the
Rocky Mountains, in the region which is bounded on the north by Lake
Winnipeg and New Brunswick, and on the south by Guatemala and
Cuba. Crayfish thus are discontinuous genera, that is, genera which
now occupy widely separated areas, such as Astacus in Europe and
Pacifie North America, but which once ranged over the intervening
ranges as well.

It is comparatively easy to distinguish the common
Cambarus from the Astacus of Europe and western
America. Members of the genus Astacus have eighteen

279

280 THE AMERICAN NATURALIST [Von. XLVI

gills, while those of the genus Cambarus have but seven-
teen. The female of the genus Cambarus has a false
pouch, the annulus ventralis, which serves as a sperm re-
ceptable, while in Astacus the sperm is deposited on the
posterior part of the thorax in spermatophores.

Dr. A. E. Ortmann has made most careful studies of
the distribution of the crayfish. References to his papers
will be found in my bibliography. Dr. Ortmann writes
me that there are in the United States and Central Amer-
ica, 74 species of Cambarus and 5 of Potambius (Asta-
cus). Inthe United States excluding Mexico, Guatemala
and Cuba, there are 64 species of Cambarus and 5 of
Potambius.

The European word ‘‘crayfish’’ is used by teachers of
zoology, probably because of Huxley’s classic, ‘‘The
Crayfish. ”’

Ortmann found (38) that not only was ‘‘crawfish”’
used by Say, 1817, earlier than ‘‘crayfish’’ by Huxley,
1880, but that in this country ‘‘crawfish’”’ is the popular
name.

‘‘Crayfish,’’ ‘‘crawfish,’’ or, as it is sometimes incor-
rectly called, ‘‘erab,’’ come from the same root, Old Ger-
man, ‘‘ Krebis,’’ from which are derived, on the one hand,
the modern German ‘‘ Krebs ”’ and the English ‘‘ crab ’’;
on the other hand the French ‘‘ ecrevisse,’’ the English
and the American ‘‘ crayfish.’’

The crayfish on which my own observations have been
centered belong to the species Cambarus bartonius bar-
toni, the only species which has migrated into New Eng-
land.

My work was carried on in the field and in the labora-
tory continuously for nine months. In the field I have
watched the activities of the crayfish in the small ponds
with which Worcester, Mass., is so well supplied. At
night I used a powerful acetylene gas lamp. In the lab-
oratory I made use of two large aquaria, one of them an
ordinary running water aquarium with a pile of sand at

No. 545] THE BIOLOGY OF THE CRAYFISH 281

one end, and the other a still water aquarium arranged
to furnish a more nearly natural habitat.

In this paper I have not touched upon the anatomy or
the work on regeneration, but have confined myself to
what is generally known as ecology or biology. In the
attempt to make the paper fairly complete I have referred
in the text to the numbers in the bibliography.

It is a pleasure to acknowledge my indebtedness to Dr.
C. F. Hodge, Dr. Newton Miller and Dr. J. P. Porter, of
Clark University, Dr. A. E. Ortmann, of the Carnegie
Museum of Pittsburgh, and Dr. E. A. Andrews, of Johns
Hopkins University.

SENSES

Touch.—Touch is probably the sense of greatest value
to the crayfish. It is sensitive to touch over the whole
surface of the body (16), especially on the chelae and
chelipeds, mouth parts, the ventral surface of the abdomen
and the edge of the telson.

Vision.—The crayfish, in common with the insects, has
a compound eye. It is believed by many that the com-
pound eye is a visual apparatus which is almost worthless
for detecting the forms of objects, especially if these
objects are stationary; but that it may furnish a very defi-
nite response to stimuli of moving objects.

Bell’s experiments with the crayfish (16) showed that
there was no response to stationary objects. The case was
entirely different with large, moving objects. The
response was not due to any change in the intensity of
light such as that caused by a shadow falling on the ani-
mals, for they would react to a movement made on the
opposite side of them from the window. Reaction to
smaller moving objects was not so marked.

Crayfish are sensitive to strong light and hide during
the day under stones, among roots of plants near the bank,
and in burrows in the bank. It is a noteworthy fact that,
in France, the people catch crayfish by building huge fires
on the bank at night to attract them.

282 THE AMERICAN NATURALIST [Vou XLVI

My own experiments indicate that in nature the cray-
fish will retreat from a strong light, but will approach a
dim one. In the spring I found that it was extremely diffi-
cult to frighten a crayfish from its food by means of my
acetylene light. In collecting at night it is very easy to
attract crayfish from some distance by setting a light on
the bank so that it dimly illuminates some little space of
water.

Smell and Taste.—Very little experimental work to.
determine the senses of smell and taste in any of the crus-
tacea was done until Bell, in 1906, tested the reactions of
the crayfish (15) to chemical stimuli, applying meat juice
by means of a fine pointed pipette to various parts of the
body. He found that the antenne, antennules, mouth
_ parts and chelipeds were especially sensitive.

Recently (1910) Holmes and Homuth published the
results of an extended series of experiments on crayfish
in which the outer or inner rami of the antennules were
removed; the antennules were removed entirely; the
antenne were removed; the chelipeds removed; and in
some specimens the brains were destroyed. (33).

They found that the outer rami of the antennules bear-
ing the olfactory sete were especially sensitive to
olfactory stimuli, that the inner rami of the antennules,
the antennz, the mouth parts and the tips of the chelipeds
were all sensitive to some extent to olfactory stimuli.

It is probable that in the crayfish we have a very highly
developed topochemical sense, or contact-odor sense.
Forel uses this term (25) in speaking of the fact that in
ants, odors are apparently detected by the contact of the
antenne.

Bell found that the crayfish was sensitive to food when
not in contact with it. I experimented with freshly cut
meat and with meat which had been exposed to the air for
some time so that the cut surfaces had dried, and found
that the crayfish would go toward and seize the fresh meat
first. Evidently the diffusion of the meat juices was
readily detected.

No. 545] THE BIOLOGY OF THE CRAYFISH 283

Hearing.—It has been pretty clearly demonstrated by
Bell that the crayfish has no sound reactions. He tried
experiments (16) such as rapping on a board floating in
the water, snapping a metal snapper in and out of the
water, and setting tuning forks in vibration in the water,
but got no response.

It is possible that the crayfish is sensitive to the sound
made by the movement of the mouth parts of another
crayfish. This has not been proved.

Equilibrium.—Bunting found that young crayfish with
the statocysts removed would swim upside down as
readily as right side up (18). It is also pretty certain
that the older crayfish have a sense of equilibrium,
although the response to rotation in their case is not
definite, but purely individual.

MATING, SPAWNING AND DEVELOPMENT.

The process of mating in Astacus differs from the
process in Cambarus. In the case of Astacus, the males
approach the females in October, November and January.

The male seizes the female with his pincers, throws her on her back
and deposits the spermatic matter, firstly on the external plates of the
caudal fin, secondly on the thoracic sterna around the external open-
ings of the oviducts. During this operation the appendages of the
first two abdominal somites are carried backwards, the extremities of
the posterior pair are enclosed in the groove of the anterior pair; and
the end of the vas deferens becoming everted and prominent, the semi-
nal matter is poured out and runs slowly along the groove of the an-
terior appendage to its destination, where it hardens and assumes a
vermicular aspect (20

After an interval of from ten to forty-five days, oviposition takes
place. The female rests on her back and bends the abdomen forward,
forming a chamber into which the oviduets open. The eggs are passed
into the chamber by one operation, usually during the night, and are
plunged into a viscid, gray mucus with which it is filled. The sperma-
tozoa pass out of the spermatophores and mix with this fluid, fertiliz-
ing the ova, but just how, and what becomes of them, are unknown (20).

The female of Cambarus differs from the female of
Astacus in having a false pouch, the annulus ventralis.
Andrews found that this pouch does not appear in Cam-

284 THE AMERICAN NATURALIST [Vou. XLVI

barus affinis until the individual has reached a third stage
after leaving the egg.

The method of sperm transfer in C. affinis, and that
of C. b. b. as well, is as follows:

The male everts the bent, nozzle-like papille at the mouth of the
vasa deferentia and through them discharges sperm into an actual
tube that passes down each of the first two abdominal appendages or
stylets. Both first and second pairs of stylets are locked together by a
peg and groove contrivance. The sperm thus passes through a closed
tube from the vasa deferentia into the annulus ventralis without com-
ing into contact with the water. Copulation lasts from two to ten
hours and may be repeated by either animal with some other (3, 4).

In a previous paper, I pointed out (21) that the males
do not distinguish the females and that males ‘‘ repeat-
edly grasp other males, and sometimes, in spite of their
frantic struggles turn them over and attempt to copulate
with them.’’ The crayfish is at such a state of nervous
tension during the period of sexual activity, that the
female will curl her abdomen at the slightest touch and the
male will at first grasp any rounded object presented to
him and attempt to overturn it. A stimulus so slight as
the slow lowering of the water when I siphoned it from
the closed tank, was sufficient to cause violent activity
among the males, with the result that all the females were
soon held by males.

Pearse, in a study of crayfish made in the laboratory
with no attempt to reproduce natural conditions (40), has
made many interesting observations and experiments,
verifying my statements (21) and adding the discovery
that a male will copulate with a dead female. He dis-
covered that the male of one species had succeeded in ad-
justing his stylets to the annulus of a dead female of
another species.

Andrews has just published a paper (13) in which he
mentions seeing males attempt to copulate with dead and
bound or paralyzed males and to actually go through all
the activities of mating with dead females except the in-
jection of the spermatophores and plugging of the annu-
lus. He agrees with my previous statement that the males

No. 545] THE BIOLOGY OF THE CRAYFISH 285

do not recognize the females, and suggests that the differ-
ence from the standpoint of the crayfish between the sexes
is a difference of behavior, which difference is perceived
by muscle and touch sense.

The passivity of the female when seized is marked in the
crayfish, but as I shall show in another paper, in the
marine crabs the female is not passive but aids in the
movements preliminary to conjugation.

It is possible, though I have not at present enough
observations to support the theory, that in the crayfish
and the lobster, deposition of sperm is most effective when
the female has just moulted and the annulus ventralis or
the ventral surface, as the case may be, is clean. In crabs
where fertilization is internal, it is necessary that the shell
be soft; softness is of course of no use where the fertiliza-
tion is external, in fact it might be injurious; but the
cleanness of a new coat may facilitate the deposition of
the spermatophores, and the retention of the plug.

Anatomically there should be no difficulty in crossing
the different species of Cambarus. It would be interesting
to see if spermatophores deposited by a male Astacus on
the shell of a female Cambarus would fertilize the eggs.
It is quite probable, however, that the female would not
leave the spermatophores on her thorax and abdomen
until the time of egg extrusion.

Andrews (13) transplanted sperm receptacles of several
females to females of another species and the mutilated
females lived to lay eggs but the eggs did not develop.
Males would not fill the transplanted receptacles.
Andrews found that conjugation between species may take
place to some extent, but did not succeed in any case in
securing sperm transfer and actual crossing of species.

There seems to be no well-marked mating season in the
cold-water species, including the species on which my
observations were made. In the ponds, mating crayfish
were not found later than November 1, but in the labora-
tory copulation occurred at intervals during the fall,

286 THE AMERICAN NATURALIST [Vowu. XLVI

winter and spring. It is probable that in its native haunts
the crayfish behaves differently.

In the spring the males die off in great numbers. This
is a phenomenon which is noted in many arthropods, and
seems to be a wise provision of nature to prevent the now
useless males from using up the food required for the
spawning females and the young crayfish. Some of the
males, however, live to a good old age. I have found
several that were over 90 mm. long.

In the case of C. bartonius bartoni there are two more
or less well-marked spawning seasons, fall and spring.

The fall laying, as indicated by females brought into
the laboratory, is during the latter part of September and
all through October and November. The spring laying
extends from about March 15 to about May 15.

Andrews observed the process of laying in Cambarus
affinis. For four or five days previous to laying, the
female cleans her abdomen diligently and is exceedingly
sensitive to disturbances during that time. The actual
laying is done in deep water at night. It takes from ten
to thirty minutes to extrude the two hundred to four
hundred eggs. Each egg is attached by a tiny filament to
the abdominal hairs (9).

The time of fertilization is supposed to be when the
eggs are laid, as they pass over the annulus ventralis.
Andrews found that on the removal of the annulus before
the eggs were extruded, the eggs were unfertilized and did
not develop.

When first extruded the eggs are almost black, but as
development goes on they become reddish in color and at
the end of about four weeks, when the young crayfish are
hatched, they are nearly transparent. The time of devel-
opment, from the extrusion of the eggs till the crayfish are
detached from the parent, is about E weeks in the
species which I studied.

Even after the young are detached Front the swimmerets
of the mother, for several days they do not venture far
from her, and taking warning at any apparent danger,

No. 545] THE BIOLOGY OF THE CRAYFISH 287

scuttle under her abdomen. It is probable that here the
visual sensitivity to moving objects is more highly
developed than in the adult in comparison with other
senses.

The young crayfish moults very frequently during the
first year.

I found that two or three days before moulting the adult
crayfish come up into the shallows exposing their cara-
paces and drying them out thoroughly. The first time that
I saw this prolonged drying-out process I did not think it
significant, for I have seen crayfish in ponds where the
water was pure and fresh, elevating their carapaces for a
few minutes at a time. It is a habit which is not neces-
sarily caused by impure water, for the same thing was
noted by me in the laboratory with animals in the run-
ning-water tank.

When I noted by the number (in oil paint) on its back
that the same individual was continually remaining only
partly submerged, I made a note and watched develop-
ments. Later, in three other crayfish I noted this pre-
liminary drying out, and predicted the approximate time
of the moults. This was convenient knowledge, for a
crayfish in difficulty with his half removed os coat falls
easy prey to his brethren.

It is possible that the aeration of the attached young by
the mother is for the purpose of enabling the young cray-
fish to moult more readily. Observations like these have
not been reported for other crayfish or for marine crusta-
ceans, but it seems possible that in the crayfish we have
such a drying out of the old exo-skeleton as we find taking
Place i in insects, like the dragon fly, which live for a time
in the water.

Andrews made a thorough study of the young of both
Astacus and Cambarus and found that in Cambarus the
young four months old averaged about 41 mm. in length.
During the winter of the first year there is no increase in
size, but the second summer of life marks an increase of
thirty per cent. in length (12).

288 THE AMERICAN NATURALIST [Vou. XLVI

I have found females but one year old with eggs, and the
development went on in the laboratory just as in the more
mature females. The largest female that I captured was
102 mm. in length. The largest male was 90.5 mm. long.

Foon.

Crayfish are omnivorous. I have previously shown
that C. bartonius bartoni prefers fresh animal food to
stale animal food or either fresh or stale vegetable
food (21).

Some crayfish eat a great deal of vegetable matter,
one species, the chimney builder, Cambarus diogenes,
seeming to prefer it. The vegetable matter eaten con-
sists of dead leaves, potato, onion, young corn and
buckwheat,

The animal food consumed by the crayfish consists of
worms, insects, insect larve, a few fish, frog, toad and
salamander eggs, and occasionally a dead fish or frog.
I have seen crayfish devour a hapless relative who was
endeavoring to rid himself of his old shell. Sometimes
females eat eggs from their own abdomens and even
devour their own freed offspring.

Enemies.—The crayfish suffers from internal and
external enemies. Among the plants which live symbiot-
ically with the crayfish are diatoms, bacteria and sapro-
legnia. Internally, Distoma cerrigerum and Branchiob-
della have been noted. But these are not all the enemies
of the crayfish. Besides man, who uses thousands of
dollars worth of crayfish for food and as a garnish, many
small animals find them palatable.

Many fish, including the black bass, Micropterus,
which fishermen find very partial to crayfish, eat them.

Professor Surface reported (38) that the salamanders
Cryptobranchus allegheniensis and Necturus maculosus,
are among the chief enemies of the crayfish.

Ortmann mentions seeing the water snakes, Natris
sipedon and N. lebens, when captured, disgorge crayfish

No. 545] THE BIOLOGY OF THE CRAYFISH 289

and has also found garter snakes, Eutenia sirtalis, in
the holes of Cambarus monongalensis.

In the laboratory and in the field I have found that the
common box turtle catches many crayfish.

Many birds, including the eagle, king-fisher, wild ibis
and turkey, have been observed with crayfish in their
claws; or the remains have been seen at the nests.

CRAYFISH as [INJURIOUS CREATURES.

The river species do not especially injure human in-
terests except in occasionally capturing a few toads, fish
and frogs, but the burrowing species are cited by Ort-
mann (38) as being very injurious, especially in the low-
lands of Pennsylvania, Maryland and West Virginia.

They make mud piles which clog harvesting machines,
and are considered by the farmers in Maryland as such
pests that it is common to throw unslacked lime over the
fields in order to kill the unwelcome tenants.

West Virginia farmers claim that the crayfish de-
stroy crops of buckwheat, corn and beans by eating the
young sprouts.

Great damage is done by the burrowing species Cam-
barus diogenes, in burrowing into dams on ponds and
reservoirs, one notable instance being the levees of the
Mississippi (38).

To destroy crayfish it is customary to throw Uae,
slacked lime over the fields, or to pour carbon bisulphide
into the holes, or to drain the infested area.

None of these measures is efficacious, the first two
methods being impracticable on account of the difficulty
in reaching the bottom of the burrow and the second,
simply lowering the water level, only delays matters a
little.

VALUE OF THE CRAYFISH.

è?
At the present time, with the lobster fishery in a state
of decline, it seems as if the crayfish could be profitably
substituted for its larger cousin.

290 THE AMERICAN NATURALIST [Vou XLVI

In a carefully written paper (11) Andrews sets forth
the possibilities of crayfish propagation. |

He states that, from the small region on the Potomac
between Washington and Fort Washington, it was esti-
mated that there were half a million crayfish sent
annually to New York.

New York, New Orleans, Chicago, Milwaukee and
San Francisco, and many other large cities consume
large quantities as food.

In 1902, the U. S. Fish Commission reports state the
crayfish catch of Monroe Co., Florida, was 55,664 pounds,
worth $3,382.

In Oregon, 116,400 pounds, worth $7,760, were caught
in one year.

With crayfish maturing in one season and growing to a
length of from four to five inches in three years; and con-
sidering the large number of eggs (100-600) laid by one
female, there should be but little difficulty in supplying
a large demand for these animals.

When we consider that the large Astacus readily
adapts itself to the slight difference in environment in
the east, we see that the crayfish is a very practicable
substitute for the lobster.

There should be no difficulty in disposing of the
smaller Cambarus, either as fresh food or canned, as we
get the abdomens of shrimps.

In the school and college laboratories, the anatomy of
the crayfish has been studied ever since Huxley wrote
‘‘The Crayfish.’’ The habits and activities of the young
and adult crayfish are of great interest and profit for
study. The animal is suited for many kinds of experi-
ments, and the large ganglia and nerve cells are readily
removed and are excellent for neurological work. The
psychologists should find a profitable subject for study
in the relations of mother and offspring for the few days
just after the young are detached from the mother’s
swimmerets.

Daily Life—From a lengthy series of observations,

No. 545] THE BIOLOGY OF THE CRAYFISH 291

including the continuous study of several specimens for
twenty-four hours, I have concluded (21) that the cray-
fish shows his greatest activity at nightfall and at day-
break. In nature the crayfish is less active during the
day than he is in captivity, since, as a rule, he has more
hiding places in his natural habitat.

Pearse has stated (40) that the number of matings
occurring in two boxes, one being painted black and
closed entirely, and the other being exposed to light, did
not vary to any extent. It is obvious from the work of
Andrews and myself that the fact that crayfish in a state
of sexual tension, stimulated by transference to different
receptacles, copulate as well in the light as in the dark,
does not bear on the question of normal activity. The
difference between night and day must not be assumed
to be entirely that of light, in experiments on higher
invertebrates. _

It is possible that the tendency of the crayfish to re-
main in hiding during the day is to some extent lessened
when sexual feeling is strong, but this seems rather im-
probable under natural conditions.

In my specimens hibernation was well marked. I was
careful to change the water daily in my still-water
aquarium, thus keeping it fairly cool. Several of my
crayfish hibernated as long as six weeks at a time, in
closed burrows in the bank of this miniature pond.

BIBLIOGRAPHY

1. Abbott, ©. C. 73. Notes on the Habits of Certain Crayfish. AM.
Nart., Vol. 7, pp. 30-34.

2. Abbott, C. ©. ’85. How the Burrowing Crayfish works. Inland
Monthly. Columbus, Ohio, Vol. 1, pp. 31-32.

3. Andrews, E. A. ’95. Conjugation in an American Crayfish. Am. NAT.,
Vol. 29, pp. 867-873. L

4. Andrews, E. A. ’04. Breeding Habits of Crayfish. AM. Nat., Vo

38, pp. 165—206.
04. Crayfish Spermatozoa. Anat. Anz., Vol. 25, pp-

on
3
Qu
@

i

©]
D>

705. nee Sperm Receptacle of Cambarus. J. H. U.

A.
Circ., No. 178, pp.
. Andrews, E. A. 06. pasa of the Annulus Ventralis. Biol. Bull.,

Vol. 10, pp. 122—137.

a
>
B
Q
|
á
A
ie]

~q

292 THE AMERICAN NATURALIST [Vou XLVI

8. Andrews, E. A. ’06. Partial Regeneration of the Sperm Receptacle
in Crayfish. J. Exp. Zool., Vol. 3, pp. 121-128.
9. Andrews, E. A. ’06. Egg laying of Crayfish. AM. NAT., Vol. 40, pp.

343—356.
10. Andrews, E. A. ’06. The Annulus Ventralis. Boston Soc. of Nat.

11. Andrews, E. A. ’06. The Futuri of the Crayfish Industry. Science,
N. 8., Vol. 23, pp. 983-986.

12. Andrews, E. A. ’07. The -o of the Crayfishes Astacus and Cam-
arus. Smithsonian Cont., 5, 79 p

13. Andrews, E. A. 710. Co ae in the Crayfish Cambarus affinis.
J. Exp, rete Vol. 9, pp. 235—264.

14. Bateson, W. 87, Notes on re Senses eey o of some Crustacea.
J. Mar. Biol. Assoc. Un. King., Vol.

15. Bell, J. C, 706. The fe of the faa to Chemical Stimuli.
J. Comp. Neurol. and Ps., Vol. » pp. 299-326.

16. ae J.C. 706. The Reactions of re Crayfish. Harvard Ps. Studies,

. 2, 5.

P-
17. Broce, P. PR de Zoologie Agricole. (Ecrevisse.) Pp. 702—720.
18. Bunting, M. ’93. Ueber die Bedeutung der Otolithen- -organe fur d.
geotropischen Funktionen von Astacus fluviatilis. Pfliigers Archiv,
Bd.. 54, 8. 531.
19. Chantran, S. ’70. Observations sur l’histoire naturelle des ecrevisses.
ompt. Rendu, t. 71, pp. 43—45. -
20. Chantran, S. - Sur la fecondation des ecrevisses. Compt. Rendu, t.
4 01-202.

bartonius bartoni. AM. NAT., Vol. 42, pp. 710-716.

24. Faxon, W. ’85. A Revision of the hannah. Pt. 1: The Genera
Cambarus and Astacus. Mem. Mus. Comp. Zool., Vol. 10, pp. 1-186.
25. Forel, A. (Tr. by Yeardley, M, ’07. ) The Senses of Insects. Lon-

is
26. Garman, ’89. Cave Animals from Southwestern Missouri. Bull. Mus.
Comp. Zool., Vol. 17, pp. 225-259
The

27. Gulland, F. 06, Sense of Touch in Astacus. Proc. Royal Soe.
Edinburgh, Vol. 9, pp. 151-179.
28. Hay, W. P. ’05. Instances of Hermaphroditism in Crayfishes. Smith- `

sonian Mise. Coll., Vol. 48, pp. 222-228.
29. Herrick, F. H. 295. The Ameda Lobster. Bull. U. S. F. C., pp-
1

25

30. Holmes, S. J. 03. Death Feigning in Terrestrial Amphipods. Biol.
Bull., Vol. 4, pp. 191-196.

32. Holmes, S. J. 03. Sex Recognition Among Amphipods. Biol. Bull.,
Vol. 5, pp. 288_292.

Ha
©

A
a

A
~q

a
Cc

. 545] THE BIOLOGY OF THE CRAYFISH 293

. Holmes, S. J., and Homuth, E. S. ’10. The Seat of Smell in the

Crayfish. Biol. Bull., Vol. 18, pp. 155-160.
Huxley, T. H. ’78. Ön the Classtfiestion and Distribution of the Cray-
fishes. Proc. Zool. Soc. London, pp. 752-788.

» duxley, THE ’80. An PRP to the Study of Zoology, Illus-

trated by the Crayfish. Pp. 1 D. Appleton & Co., New Yor
1895.

. Jennings, H. S. ’06. Behavior of Lower Organisms. Macmillan, 1906.
y E. P

on, . 799. Contribution to the Comp. sig of Compensatory `
Movements. Am. J. Physiol., Vol. 3, pp. 86-1

. Ortmann, A. E. The Ga. of the State of We Memoirs
43-52

of the Carnegie Museum, Pittsburgh, Pa., Vol. 2, pp.

. Ortmann, A. 705. The Mutual Affinities of the asain of the

E.
Genus Cambarus and their Dispersal over the United States. Proc.
Am. Phil. Soc., Vol. 44 6.

, pp. 91-

. Pearse, A. S. ’09. Observations of. Copulation among Crawfishes with

pen Reference to Sex Recognition. Am. Nart., Vol. 43, pp. 746—

5 Pe C. W. ’01. The Otoeyst of re a Crustacea. Bull. Mus.
6

Comp. Zool. Harvard Univ.,

- Souberain, Leone. 65. fas. l’histoire Sonar et l’education des

ecrevisses. Comptes Rendus, t. 60, pp. 1249-1250.

- Tarr, R. S. ?84. Habits of a Crayfishes in the United States.
128.

Nature, Vol. 30, pp. 127—

- Washburn, M. F. ’08. The Animal Mind. Macmillan, N. Y., 1908.
- Weed, H. E. Carbon Bisulphide for Crayfish. Proc. Seventh Annual

Meeting of Assoc. be Ec. Entom. U. S. Dept. of Ag., Div. of Entomol.,
Bull. 2, N. S., p. 100.

- Wheeler, W. M. ’10. Ants, their Structure, pae and Behavior.

Columbia Univ. Press, New York, 1910. (Pp. 509-5

15.)
. pera E. B. ’07. A Collecting Trip North of Sault Ste. Marie,

Ohio Nat., Vol. 7, pp. 129—

. S R. M., an ad Huggins, G. E. ’03. Habit Formation in the

Crawfish Cenir affinis. Harvard Ps. Studies, Vol. 1, pp. 565-
577 ;

PRESENT PROBLEMS IN SOIL PHYSICS AS RE-
LATED TO PLANT ACTIVITIES?

PROFESSOR BURTON E. LIVINGSTON
THE JOHNS HOPKINS UNIVERSITY

Tr is from the point of view of the physiologist and not
from that of the analytical physicist that I propose here
to consider some of the most obvious and insistent of the
non-chemical problems of the soil. We shall thus be
interested not in the physics of the soil, but in the rela-
tion of some of its physical properties to certain plant
activities. This is a somewhat unusual point of view, for
most soil investigators have studied the soil largely to
the exclusion of the plant, bringing refined chemistry
and physics to the statement of one member of the equa-
tion and stating the other member largely from the
standpoint of the unscientific man. This generalization
applies to studies upon both the physics and the chem-
istry of the soil, but, owing to the majesty of the great
chemist Liebig? and to the multitude of his followers,
soil physics has nowhere received the attention which it
deserves, and the relation of the physical condition of
the substratum to plant activities remains perhaps the
most fundamental and at the same time most neglected
of all the various environmental relations.

Since we are certain that the water relation is of ex-
ceedingly great importance in the control of plant proc-
esses, and since so many other physical soil conditions
depend largely upon soil moisture, I shall consider here
primarily only the water relation of terrestrial plants

Par in the Symposium on Problems of the Soil, before Section G,

. A. A. S., at the Washington meeting.

Speke te in regard to the present contention, the title of Leibig’s monu-

mental work, ‘‘Die Chemie und ihrer Anwendung auf Agrikultur und Phys-
iologie,’’ 1846.

294

cae soe eon fs

No. 545] PRESENT PROBLEMS IN SOIL PHYSICS 295

below the soil surface. But, as will shortly appear in
some detail, to appreciate the problems before us it will
be necessary, not only to deal with the internal condi-
tions of the root system together with the external ones
of the soil, but also to bear constantly in mind certain
relations which obtain above the soil. I shall begin with
a brief treatment of the water relations of the plant,
with special reference to the physical conditions of its-
subterranean environment.

In order that the water content requisite for the
various physiological processes may be maintained, the
condition must obviously be fulfilled, that the ratio of the
rate of water income to that of water removal must never
fall below unity. Now, the removal of free water from
the physiological system of the plant occurs in three
general ways: (1) the fixation of water by growth, etc.,
(2) the excretion of liquid water at the periphery and (3)
the loss of water vapor by transpiration. The first two
of these are usually negligible, and the prime aerial con-
dition of plant activity—as far as the water relation is
concerned—is the rate of loss of water vapor. This loss
is to a variable extent controlled by conditions within
and without the plant, but we do not need to give these
attention now. The main point for us to bear in mind is
that, for the activities of the majority of terrestrial
plants, it is requisite that the entrance of water through
the roots must equal its rate of exit through the leaves
and other aerial parts.

Of course water will not, in general, enter through the
roots faster than it is removed from the plant body or
fixed therein by growth and metabolism, and the critical
consideration in respect to the soil water relation is not
the actual rate at which water is entering (this depend-
ing upon the internal conditions of the plant as well as
upon the soil), but the maximum possible rate at which it
May enter if the prerequisite internal conditions arise.
In this respect, then, that soil is best suited to continued
physiological activity, which possesses the highest power

296 THE AMERICAN NATURALIST [VoL. XLVI

of supplying moisture to the absorbing regions of the
plant.

It would seem, a priori, that a flooded soil should offer
the least possible resistance to water movement, but such
a soil appears indirectly to reduce water entrance in
many forms by influencing (probably directly or indi-
rectly in a chemical way) the internal conditions of the
plant, and it is only with a soil in a considerably drier
condition than the flooded one, that we find the optimum
subterranean environment for ordinary plant processes.
As the soil becomes drier, its direct resistance to water
intake by the roots increases, slowly at first, then rapidly,
and at a certain stage (for any given complex of aerial
conditions, and hence for any given transpiration rate)
the combined resultant of the movement of soil moisture
to the root surfaces and that of these surfaces through
the soil (by growth) falls to a magnitude so low that the
processes of transpiration and of growth, etc., remove
water from the tissues more rapidly than it enters below.
This condition of the substratum is approximately what
is usually termed the wilting point, and the remaining
water in the soil is said to be unavailable for plants.

In researches which have yet to be published, my asse-
ciates and I have shown that this wilting point is not the
constant which it has been supposed to be, for either soil
or plant. It is possible to cause the lower limit of ‘‘avaul-
able’’ water in the soil to assume almost any magnitude,
within a broad range, for any given plant, merely by
altering the rate of transpiration,—through proper
changes in the evaporating power of the air and the
intensity of the impinging solar radiation. The wilting
point thus ceases to have any meaning at all, unless the
corresponding rate of transpiration is known, or unless,
indeed, the aerial environment is known to be the same
throughout any series of cultures the data from which
are to be compared.

The primary problem, then, which must be quantita-
tively solved if we are to place the soil water relation in

No. 545] PRESENT PROBLEMS IN SOIL PHYSICS 297

a way that may lead to a scientific foundation, is con-
cerned with the maximum rates at which various soils
may furnish moisture to the root systems of whatever
plant forms with which we may be dealing. To such an
end, our knowledge of the physiology and ecology of
roots must be enormously increased, but with this phase
of the matter we need not here concern ourselves. It is
obvious, however, that the really crucial question with
regard to any soil, the properties of which we wish to
study with reference to plant behavior, is this: at what
rate, and for how long a time, can it deliver water to unit
area of a water-absorbing surface? This is a purely
physical question and one for which it ought not to be
very difficult to find adequate methods of attack. Indeed,
the method of studying evaporation from soil surfaces
already offers approximate results in this direction.
This maximum rate of delivery per unit of cross sec-
tion must be related in some manner to the soil charac-
ters which are now often measured; the power of water
delivery will vary with the percentage of water content
for any particular soil, and its graph will most likely ex-
hibit a critical point under about the same conditions as
those which accompany the critical points for evapora-
tion from the soil, the apparent specific gravity of the
latter, its penetrability (as recently brought out by Cam-
eron and Gallagher), its critical moisture content and
its moisture equivalent (as brought out by the centrif-
ugal method of Briggs and McLane). That the critical
point in maximum rate of delivery of moisture will be
found to correspond to the ordinarily observed optimum
water content for many plants is also to be expected, but
the physiologist will not make the mistake of supposing
that this optimum water content will not vary largely
with the nature and condition of the plant and also with
its rate of transpiration. That this critical point, with soils
of varying water content, will be found to be related to
the size, nature and arrangement of the soil particles is
likewise fairly certain, and it may confidently be ex-

298 THE AMERICAN NATURALIST [Vor. XLVI

pected that this point will exhibit some definite relation
to the heat of wetting (as this property has been devel-
oped by Mitscherlich), and perhaps also to the commonly
determined water capacity or water-retaining power of
the soil. The last named is a property which, as I have
previously pointed out, seems especially worthy of inves-
tigation by ecologists who are seeking some easily de-
termined soil characteristic for use in studies on plant
distribution.

In this connection it is well to call attention to the
apparent futility of the method of mechanical analysis,
which is resorted to so extensively—and so expensively
—in attempts physically to describe the solid portion of
the soil. I think I do not exaggerate when I say that the
physical analysis has shown itself to be practically worth-
less for any physiological purpose. It assuredly does
furnish a means of describing a given soil sample with
considerable accuracy, and if two samples could ever be
found to exhibit exactly the same proportions of the
different sized particles, it might plausibly be supposed
that, ceteris paribus, these should possess the same rela-
tions toward water and toward plant roots, but the con-
verse of this statement is not at all true. This method
furnishes a mass of data from which no one has yet been
able to derive any single comprehensive summation that
will express anything definitely as to the possibilities of
the given soil as a substratum for plants. Undoubtedly
the size of the component soil particles plays a large part
in determining how the water conductivity varies with
different conditions of soil moisture, etc., but we need to
seek some feature which may be much more readily meas-
ured for the soil as a whole than merely the proportions
of various-sized particles.

Should we be able to find out the relations which obtain
between the maximum rate of water delivery and the
other soil characters that I have mentioned, it might at
length become possible physically to assay a given soil
by the determination of one or more of the latter, sub-

No. 545] PRESENT PROBLEMS IN SOIL PHYSICS 299

sequently passing to the real point of interest by means
of an interpretation, but such a possibility is at present
so far removed from actuality that it seems highly desir-
able to begin with attempts to measure the soil property
which directly influences plants. In any event, it can not
be too strongly emphasized that such soil studies as I
am suggesting must always be carried on simultaneously
with studies on the behavior of plants, and also with ade-
quate determinations of the water-extracting power of
the aerial environment. It seems quite likely that we
shall be able empirically to determine some highly im-
portant principles bearing upon the water relations which
exist between plants and soils, without having yet suc-
ceeded in analyzing the mode of manifestation of these
into its elementary physical propositions—just as it has
recently been possible to work out exceedingly valuable
principles with reference to the relation of plants to
evaporation, without any one’s having yet succeeded in
determining the quantitative dependence of this climatic
factor upon its components, water and air temperature,
air humidity and air movement.

When a little headway has been gained in the dynamic
study of the soil in relation to plant processes, we shall
probably begin to be able to interpret and correct, and
place upon a proper quantitative basis, some of the eco-
logical classifications of plants and the physical classi-
fications of soils, which already occupy so much of our
literature.

Another aspect of this whole question of the water
relations of the subterranean parts of the terrestrial
plant may be worthy of attention. The majority of the
physical soil studies which have so far been made depend
upon the removal of the soil sample from its natural
position, with consequent and usually profound altera-
tions in the arrangement of its component grains, upon
which arrangement assuredly depend some of the most
fundamental of the soil qualities which we need to know
about. Various methods have been devised aiming to

300 THE AMERICAN NATURALIST [Vou XLVI

avoid this difficulty, but all are exceedingly cumbersome
in the operation and are at best of somewhat doubtful
efficiency. Here is suggested a line of work which has
already been attempted by a number of enthusiastic
students, many of whom have afterward given up in
despair without even publishing their experience. The
director of one of the great European experiment sta-
tions told me of a somewhat elaborate apparatus which
he once constructed for determining soil moisture in situ.
He concluded with the remark, ‘‘the principle was cor-
rect enough, but the method proved useless.’’ Jam sure
that he is not alone in his experience. But the problem
of soil instrumentation will not be dropped; I am confi-
dent that the future will develop methods in soil physics
which will not necessitate any alteration in the soil at
the time a determination is made. Studies upon the soil
properties in the light of their rôle in plant environment
and accompanying studies on the physics of plant activi-
ties will do much toward furthering our science in this
direction. The actual accomplishment of this end may
not be very far off; we may take heart from such facts
as this, that a single decade has sufficed to bring aerial
navigation from the limbo of scoffed-at impossibility (in
the minds of all but a very few scientists) into the cate-
gory of accomplished fact. And the importance of ade-
quate methods for the study of problems of the soil is far
greater, and probably will ever remain far greater, than
that of any problem of transportation.

To summarize my suggestions:

1. The soil water relation is of fundamental impor-
tance if we are some time to know about and be able to
predict and control plant processes.

2. The moisture of the soil, as well as its other fea-
tures, is most profitably to be studied as plant environ-
ment, the relations which obtain between plant activity
and soil phenomena comprising a fundamental and
primary requirement for the scientific advance of our
knowledge.

No. 545] PRESENT PROBLEMS IN SOIL PHYSICS 301

3. The physical nature of the subterranean environ-
ment of terrestrial plants is effective in controlling plant
activities, mainly with regard to the possible rate of de-
livery of water by the soil to unit area of absorbing roots.

4, It is highly desirable to study this power of water
delivery with reference not only to the growth of plants
but also to other soil characteristics, some of which are
already commonly measured.

5. The whole problem of the physics of the subter-
ranean surroundings of rooted plants awaits the develop-
ment of an instrumentation which will not necessitate the
preliminary destruction of some of the most important
soil properties before the soil can really be studied.

SHORTER ARTICLES AND DISCUSSION

FURTHER NOTES REGARDING SELECTION INDEX
NUMB

RS!

THE purpose of the present communication is to correct and
extend a former paper from this laboratory? dealing with the
use of index numbers in mass selection operations. In the cor-
respondence which the writer has had with various workers re-
garding that paper it would appear that a point which it was
intended should be emphasized has been rather overlooked.
This is that the examples of index numbers therein given for
sweet corn and for poultry were intended merely to illustrate
the principles involved. They were not put forward as the best
formule which could be devised, even for the organisms dis-
cussed. It was pointed out that the particular formula to be
used should be devised by each worker to fit his particular needs.
Apparently a number of workers have adopted without change
the formule given in our first paper. I wish again to empha-
size that unless these happen to meet exactly the particular needs
of the breeder, it is highly desirable that he develop formule of
his own, involving the same general principle, but adapted to
his special conditions.

I. CORRECTION OF AN ERROR IN THE FORMULA OF A SELECTION
INDEX NUMBER FOR CORN
a

In our first paper there is an error in one of the equations for
the selection index for sweet corn (loc. cit., pp. 397-399).

This error has given trouble to some workers desiring to use
this index number in breeding work with corn, and may cause
confusion in the future. Doubtless some of those who have used
the index in their work have, like the writer, made for them-
‘selves the somewhat obvious correction. Nevertheless, to insure
that there may be no further confusion it seems desirable to pub-
lish a formal correction.

*Papers from the Biological Laboratory of the Maine Experiment
Station, No. 35.

? Pearl, R., and Surface, F. M., ‘‘Selection Index Numbers and Their
Use in Breeding.’’ AMERICAN NATURALIST, Vol. XLIII, pp. 385-400, 1909.

302

No. 545] SHORTER ARTICLES AND DISCUSSION 303
The corn index number has the following formula

fa A-+3B 4.20
~ D+ ue
The definition of the variable C given on p. 393, by an un-
fortunate slip of the pen, which escaped detection in the proof,
as such things will, gives precisely the inverse effect from what
it should. The equation should read as follows:

100 times the circumference of the cob at middle
Circumference of ear at middle

C= 100 —

The example on p. 399, which was worked out after the text
. Was written, followed the erroneous text with scrupulous exacti-
tude in theory, but with a slip in the arithmetic. The correct
value of J, for the ear used as an example is

190.0 + 70.5 477.6 338.1
he sake S

Experience in the use of this index suggests that in the equa-
tion for C given above it may be advantageous to substitute
‘‘diameter”’ for ‘‘cireumference’’ in each case. The diameters
can be much more easily and accurately measured and they
probably give a better appreciation of the relative kernel depth
than do the circumferences.

II. A SELECTION INDEX NuMBER FOR BEANS

The writer has under way at the present time some breeding
experiments with a very interesting variety of beans, known
locally as the ‘‘Old-fashioned Yellow Eye.” It is a variety ap-
parently scarcely known now outside of northern New England.
Owing to certain defects it has been replaced in most of the bean-
growing sections of the country where formerly grown by the
Improved Yellow Eye, a perfectly distinct and in many respects
inferior type. From the standpoint of experimental genetics
the old-fashioned yellow eye bean promises to furnish material
of great interest and value in the unraveling of such problems
as pattern inheritance, the effect of selection in pure lines, ete.

Aside from the technically biological considerations, however,

304 THE AMERICAN NATURALIST [Vou. XLVI

this bean possesses much economic significance in Maine. It is
esteemed above all other sorts for baking purposes, and if a
strain could be developed which would possess (a) high yielding
qualities, (b) reasonable disease resistance and (c) earliness
and uniformity of maturing it would be of great value to the
bean growers of the state. In connection with the purely bio-
logical studies an attempt is being made to see whether a pure
line possessing these desirable qualities may not be found.

In this specific breeding problem we obviously have the con-
ditions which demand the aid of selection index numbers. Sev-
eral characters (not one only) must be concurrently selected.
An estimate must be formed in each case of the net worth of an
individual plant (or of a biotype), taking into account at least
all of the three factors named. In order to do this impartially
and accurately a selection index number has been devised.

In deriving this bean selection index a general equation of a
slightly different type than that discussed in our former paper
has been employed. In that paper (loc. cit., p. 389) the general
formula suggested is

fa” + by ce es Ee
: reer Ge ee et lb:
In the case of beans (and very likely this may prove true for
other plants and animals as well) it has seemed desirable to form
an index number on the plan of the following type of equation:

axy + bwz + --- + nuv
apronda TAI

In this equation, as before, a, b, c,...n, and a’, b,

n’ are constants, given arbitrary values in accordance ae the
scheme of weighing adopted, and z, y, z, w, u, v, are variables
which measure characters increasing in desirability (from the
breeders’ standpoint) as their absolute magnitudes increase,

while p, q, r, s and t are variables measuring characters which
decrease in desirability as their absolute magnitudes increase.

The variables specifically taken account of in the bean selection
work are:

Y = Absolute yield. The weight in grams of dried shelled
beans per plant.

No. 545] SHORTER ARTICLES AND DISCUSSION 305

V = Relative yield. The percentage which Y is of the weight in
grams of the whole plant. This factor measures the
degree to which the plant transforms its food materials
into seeds rather than into foliage parts.

P = Number of pods per plant.

B = Mean number of beans per pod.

D = Disease-maturity index. The percentage which the number
of perfectly matured beans free of disease (anthrac-
nose) is of the total number of beans originally set in
the pods. This measures the degree to which the per-
formance of the plant in seed production approaches
its promise in that regard. It does not separate disease
resistance from earliness and completeness of maturity,
but from a purely practical standpoint this is not essen-
tial. By making separate counts of diseased and imma-
ture beans it would be possible to take account of each
of these factors by itself. It must be understood further
that the separation of diseased beans is not absolutely
complete. Only those are counted as diseased which
show to the unaided eye evidence of anthracnose infec-
tion. It has not been found feasible as yet to get a
simple and satisfactory measure of the degree of attack
of other bean diseases. Hence, for the present, only
anthracnose is being taken account of specifically in
the selection index number.

` These variables are combined in the following bean selection
index number:

The values taken by this index number for a particular strain
of Old-fashioned Yellow Eyes are shown in Table I.

From the table it is clear that the index may take a rather
wide range of values, depending upon the character of the plant.
Further, the value of the index is obviously not unduly influ-
enced by any particular variable. The high index values seem
clearly to indicate the plants wick e te best, taking all
things into account. This, of course, is the goal sought.

It is of interest to note the values taken by the index in the

ease of a bean of quite different type, namely, a White | field oe

306 THE AMERICAN NATURALIST [Vou. XLVI

TABLE I
VALUES OF THE SELECTION INDEX NUMBER FOR A SERIES OF PLANTS OF
MORSE’S OLD-FASHIONED YELLOW EYE BEAN, TOGETHER WITH THE
VARIABLES ON WHICH THE INDEX DEPENDS

Plan lection| Absolute | riala ‘| MeamNo.| Mean | Total no, | Disease
No. | Ta Yield Index EERDE prr Weight of Pods | a
873 OO. Shen: ma 55 eas 5 60.
85 79 2 36.36 3.90 AT 0 33.33
86 1.24 | 10 41.67 3.43 .50 7 62.50
61 ist | 15 00 3.77 .39 13 | 41.45
98 1.37 41.67 3.20 58 75.00
95 £5) 1:21 56.40 2.73 33 15 57.77
59 1.56 | 13 54.17 3.78 55 58.82
76 173 |12 46.15 3.42 43 12 54
62 00 | 16.1 49.53 4.33 50 9 64.10
84 2.08 50.97 3.58 43 12 67.44
71 2.58 | 24.5 64 3.47 56 15 55.98
2.92 8 59.99 2.87 49 15 69.77
89 3.26 | 20.5 52.58 3.71 50 14 70.37
55 3.43 59.38 3.69 42 16 69.49
69 3.73 1 59.99 3.73 57 11 75.61
92 3.73 | 12 59.99 3.57 57
90 3.83 | 19.5 59.10 3.33 53 12 77.50
56 3.97 | 30.9 59.42 4.33 47 18 58.97
63 447 |23 51.10 4.53 41 17 70.13
75 5.35 | 39 59.99 4.07 47 27
81 5.51 | 27 64.27 3.18 51 17 75.93
7.32 | 27 59.99 3.42 46 19 80.38
70 7.37 | 26 61.16 3.59 46 i See
60 7.56 | 19 55.07 3.38 36 16 87.04
78 8.47 | 29 61.70 3.19 46 21 80.97
8.79 | 30 63.17 4.14 56 14 84.48
67 9.42 | 23 52.88 3.68 34 19 86.59
83 9.50 | 26 59.10 3.14 43 22 84
g2 | 10.09 | 3 61.24 3.67 46 18 98.48
97 | 10.38 | 50.5 44.11 5.42 29 41 57.21
79 15.30 | 3 62.50 3.48 43 31 88.93
73 | 17.04 | 38 60.31 3.89 42 26 86.1

pea bean. Table II gives the index and component aaa:
for a series of plants of such a variety.

The range of values here is large. The extremely high values
are probably much larger than will ever be obtained for a bean
of the yellow eye type, though it is rather risky to make such
a prophecy. Two factors help in reaching such high index
values in the case of this variety. One is the tendency to prolifi-
cacy, there being relatively many pods per plant and beans per
pod. The other is the rather high disease resistance of the

3 Plant injured by cut worms.

* Plant injured by cut worms, but subsequently grew. x

No. 545] SHORTER ARTICLES AND DISCUSSION 307

plants. They mature, apparently free from disease, a large
proportion of their seeds
TABLE II

VALUES OF THE SELECTION INDEX NUMBERS FOR A SERIES OF PLANTS OF
SNow FLAKE FIELD BEANS

; Mean
Plant lecti Absol Yield otal N: i
No. Sever Yield ader Pen por Weight pe a yp a gd
1864 7 31.82 4.13 21 8 87.88
1924 4.42 23 46.95 5.30 16 33 54.42
187 6.5 44.82 5.86 17 7
191 12.25 38 56.72 6.06 24 34 73.30
183 12.86 9 56 4.10 22 10 95.12
190 25.44 55 48.24 5.10 30 42 79.91
188 30.72 6 7.80 5,35 32 40 82.71
84 71.98 50.5 57.14 5.64 19 52 91.81
185 101.12 40 57.14 5.35 22 34 96.71

Of course, the index numbers may, strictly speaking, be com-
pared only among plants of the same variety. The absolute de-
sirability of a variety for a particular purpose depends upon
many other factors not taken account of in the index number.
These numbers can not be used directly and solely as measures
of the relative worth of varieties.

It is hoped that this bean selection index number or some mod-
ification of it may be found useful by other workers. It will, at
any rate, serve to illustrate further the adaptability of the gen-
eral idea of such numbers to a wide range of practical selection
work. In the present instance a selection index number is
applied to the measurement of the relative worth of different
distinct biotypes, rather than in the mass selection of fluctuat-
ing variations, in which latter type of work such numbers were
shown in our former communication to be useful.

RAYMOND PEARL

UNIVERSITY OF MAINE

NOTES AND LITERATURE
PROTOZOA:

THAT so expensive and highly specialized a text-book as this
of Doflein’s should run through a whole edition in less than a
year is a tribute to the excellence of the work and an index to
the scientific activity in this field of biological research. An
indication of the rapid progress now in the making in protozo-
ology may be derived from the fact that every chapter in this
elaborate work has been rewritten or substantially emended
and the number of pages and illustrations increased by fifteen
per cent. in this third edition, the second having been issued
less than two years ago. The main changes include the insertion
of a chapter on the origin of the Protozoa, the conception of
species within the group, and the phenomena of variation and
heredity as revealed by methods of culture and experiment,
especially by the study of pure lines and the results of selec-
tion. Doflein calls attention to the appearance of direct adapta-
tions in parasitic organisms in response to definite environ-
mental factors in the form of chemical substances such as
atoxyl and various compounds of arsenic and of antimony,
unknown in the normal environment of the protozoan organism.
These adaptations result in so-called resistant races and may be
heritable. The possibility of control, the large numbers avail-
able and the rapidity of multiplication of these pathogenic
organisms unite to open an inviting field, thus far too much
neglected by the investigator in experimental evolution.

Considerable additions are made to the discussion of repro-
duction, especially to the maturation of the gametes, in which
homologies to maturation in the Metazoa are becoming increas-
ingly definite. The detailed discussion of the various groups
of protozoa is noticeably extended in the case of the Spiro-
chætes, the Hemosporidia and the Sarcosporidia.

***Lehrbuch der Protozoenkunde. Eine Darstellung der Naturgeschichte
der Protozoen mit besonderer Beriicksichtigung der parasitischen und patho-

en Formen.’’ Dritte stark vermehrte Auflage. Von Dr. F. Doflein.
xii + 1043 pp., mit 951 Abbildungen im Text. Jena, Gustav Fischer, 1911.
M. 26, gb. M. 28.50.

308

No. 545] NOTES AND LITERATURE 309

Doflein is inclined to accept the evidence that Schaudinn’s
account of Entameba histolytica is based in part upon phe-
nomena attendant upon processes of degeneration and suggests
that Viereck’s E. tetragena is probably the most widespread
form causing ameebic dysentery, and that the two are possibly
identical, but that the organism according to the rigid laws of
priority should be called Entamaba dysenteriae (Councilman
and Lafleur).

The doubtful group Chlamydozoa established by Prowazek
for that group of immunizing organisms with a filterable virus,
the supposed etiological factors in such diseases as vaccinia,
variola, trachoma, molluscum contagiosum and epithelioma con-
tagiosum, is still denied admittance by the author to the Pro-
tozoa on the ground that the minute structures described by
Prowazek are not themselves with certainty proved to be living
organisms. Doflein admits, however, that the evidence is con-
Stantly increasing that we have to do in the case of these dis-
eases with parasitic organisms, but thinks they may be more
closely related to the bacteria than to the protozoa.

It is a matter of regret that the non-parasitic groups, such, for
example, as the pelagic Foraminifera and Radiolaria, and non-
parasitic flagellates can not receive in a work of this sort com-
mensurate treatment with pathogenic forms of confessedly great
biological, as well as medical and hygienic interest. The author
expresses the hope that medical research may in the near future
so clear up contested points that less space will be required for
the discussion of pathogenic forms. The present output is,
however, not very promising for a reduction in extent in this
field. The fact is that a six-volume edition of the Protozoa in
Bronn’s ‘‘Thiereich’’ is needed to give anything like an ade-
quate review of the results now achieved in the fields of Pro-
tozoology.

CHARLES ATWOOD KOFOID

UNIVERSITY OF CALIFORNIA

HEREDITY
H. M. Leake’ gives additional results of his studies of inheri-
tance in cotton. The flower color factors found were yellow,
Pale yellow and red, the latter being due to red sap color which
showed not only in the flowers but in stems and leaves as well.
“<< Studies in Indian Cotton,’’ Jour. of Gen., Aug., 1911.

310 THE AMERICAN NATURALIST [Vou. XLVI

Yellow was completely dominant to its absence and to pale yel-
low. Red was incompletely dominant. The very interesting
fact developed that although yellow behaved as an allelomorph to
its absence in crosses with white, it was also allelomorphic to pale
yellow in crosses with the latter. This indicates that pale yellow
is simply a modified form of yellow, a fact in entire accord with
my teleone theory of Mendelian inheritance, and opposed to the
de Vriesian idea of the immutability of the so-called unit charac-
ters. An interesting case of correlation was found. White (ab-
sence of yellow) is hardier than yellow.

In shape of leaf Leake uses as an empirical means of describ-
ing leaf shape a formula which is essentially the ratio between
the length and breadth of the central lobe. The pure races (and
the author took the pains to work with pure races) may be di-
vided into two groups with reference to this ‘‘leaf factor,”
namely those in which it is less than 2, and those in which it is
greater than 3. No cases were found in pure races in which the
value of this factor was between 2 and 3. F, between these
groups gave intermediate leaf factors. F, apparently behaved
as if the cross involved a single gene, but fluctuating variation
obscured the results considerably. Crosses between F, and
either parent form gave only the intermediate and the one par-
ent form, the same difficulty appearing from fluctuation in the
character. This strongly confirms the conclusion that a single
gene is responsible for the difference between these two groups.

Earliness of flowering late flowering proved to be a very in-
teresting study. The author had previously discovered that
types with sympodial secondary branches flower early, while
those having monopodial secondaries are late flowering. This re-
lation had also been noticed by others, the early or late flowering
being a result of the manner of branching. Length of vegetative
period (time between planting and first flower) proved to be
highly fluctuating, varying widely as between different seasons.
F, between the monopodial and the sympodial types was inter-
mediate between the parents, but nearer the sympodial (early)
parent. F, gave a continuous series extending from the early
parent nearly to the late parent, the frequency curve for the
earliness in the F, population being monomodal. While the
author does not pursue the subject further, it may easily be
shown that this is exactly what Mendelian theory calls for on
the assumption that several factors, each alike in effect, their

No. 545] NOTES AND LITERATURE 211

effects being additive, are responsible for the parental differences,
especially when the character in question fluctuates widely as
compared with the differences between the several genotypes
occurring in F,. Thus, suppose three factors, A, B and C, each
alike in effect, and each producing the same average increase in
length of vegetative period. The F, generation of the cross abc
X ABC will consist of the genotypes aabbcc, aabbCC, aaBBcc,
AAbbcc, aaBBCC, AAbbCC, AABBcc and AABBCC and their
crosses. The genotype aabbcc would be similar to the early
parent. Genotypes AAbbcc, aaBBcc and aabbCC would con-
stitute a group one stage later in flowering. AABBcc, AAbbCC
and aaBBCC constitute a third stage, while AABBCC would be
equivalent to the late parent. Thus the four stages resulting
from these three factors tend to be present in the ratio 1:3: 3:1,
which ratio is merely one way of stating the properties of an
ordinary frequency curve. LEarliness being nearly completely
dominant, the norm of this curve would be shifted toward the
early parent, as Leake found was the case. Even if this progeny
were selfed to the tenth generation, by which time heterozygosis
would have largely disappeared, the mixture of the four geno-
types would still give a monomodal curve. The only exception
to this would be cases in which fluctuating variation is not trans-
gressive between the genotypes. It is possible that more than
three genes were involved in Leake’s crosses.

Crosses between pure lines having no leaf glands and those
having leaf glands gave intermediate F,. F, gave evidence of
segregation, but the intermediate and apparently highly fiuctu-
ating character of the heterozygotes rendered positive conclu-
sions difficult or impracticable.

Complete correlation occurred between flower color and length
of petals. White petals were little if any longer than the brac-
teoles, while yellow petals were about twice as long. Interme-
diates did not occur, and no exceptions were found in over 100,-
000 plants.

Red sap color was independent of the size of the petal but
when present it lengthened the vegetative period. This paper is
exceedingly clear and lucid in treatment, and we may expect
much valuable work from the author in future.

Dr. Shull has resumed his interesting studies of Bursa. He

*Dr. G. H. Shull, ‘‘ Defective Inheritance-Ratios in Bursa Hybrids,’’
Verh. d. Naturforsch. Ver. in Brünn., Bd. XLIX.

312 . THE AMERICAN NATURALIST [Vou XLVI

had previously shown that four genotypes of Bursa bursa-pas-
toris are the four Mendelian types corresponding to two inde-
pendent factors (AABB, AAbb, aaBB, aabb). In his paper
above cited he deals with a cross between one of these types
(aabb) with a genotype of Bursa Heegeri corresponding to the
type AABB. The factors A and B in this cross behave in the
usual Mendelian fashion, departures from expected ratios being
explained by variation in dominance in one of the families. But
the factor or factors governing differences in the seed pod of
these two species present departures from expected ratios that
are not fully understood. There is evidence that at least two
genes are concerned in this difference. If only one gene were
concerned the ratio between the two types of seed capsule in F,
should be 3:1; if there are two genes, the ratio should be 15:1,
three genes, 63:1.. The ratios observed in F, were 4.67:1 (in-
stead of 3:1), 15.6:1, 24:1 and 63.5:1. The latter ratio, ob-
tained in a rather large family (129 individuals), suggests three
genes. The first and fourth of the above ratios are rather wide
departures from expected ratios, and their meaning is not yet
clear. The matter is still under investigation. There seems to
be little doubt that Dr. Shull has added another case to the in-
teresting class of Mendelian characters that may be represented
by more than one independent gene, such as those found by Nill-
son-Ehle, in oats and wheat and by East in corn.

A very interesting paper by Gortner,7® giving further results
of his studies on melanin formation, appeared in the December
(1911) number of this journal. He was able to show the color
pattern in the Colorado potato beetle is due to the fact that the
chromogen z secreted only in certain spots, while the oxidizing
enzyme, which is of the tyrosinase type, is present generally in
the elytron.

W. J. SPILLMAN

* Dr. R. A. Gortner, ‘‘Studies on Melanin,’’
AMER. i V,
No. 540, pp. 743 et seq. ee

VOL. XLVI, NO. 546 “ JUNE, 1912

The American Naturalist

MSS intended for publication and books, etc., intended a review should ns
eget toe of THE AMERICAN Dagas Garrison-o
containing researc bearing on the Mansell of organic "ae
tion are cao welcome, and will be given padedami in publication
One hundrea reprints of contributions are supplied to authors free of charge.
Further reprints will be amiei at cost.
Su be sent to the publishers. The

subscription price i yie ote F osta fifty ts and
| price is four a year. oreign postage is cents a
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 as se , April 2, 1908, at the Post Offi t Lancaster, Pa., under the Act of
“ Congress of Mareh 3, 1879.
THE BULLETIN—For bargains in Ethnolograph- TENTH EDITION.

~ ical and Pre-historic Specimens. Books on Natural | THE MICROSCOPE,
History, Science, Travel, Voyages, etc. See Tug | 22introduction bo M erencuphe: Methods sat and to Histology, ,
S BUELERTS post tree tor 3 cont = Ba large octavo pages an GAGE, and above 200 ia ieee

oe ee ee See COMSTOCK pone Ithaca, N. Y.

University of Colorado
Mountain Laboratory

Fourth summer session June 24 to An-
gust 2, 1912. Courses in Field | B
ae and pr Ecology, §

THE
AMERICAN NATURALIST

VoL. XLVI June, 1912 No. 546

A FIRST STUDY OF THE INFLUENCE OF THE
STARVATION OF THE ASCENDANTS UPON
THE CHARACTERISTICS OF THE
DESCENDANTS—I

Dr. J. ARTHUR HARRIS

CARNEGIE INSTITUTION OF WASHINGTON
I. [INTRODUCTORY REMARKS

One need not search widely in biological or agricultural
literature to encounter discussions of the influence of the
conditions to which the ancestors are exposed upon the
characteristies of the offspring which they produce. To
review here the mass of more or less pertinent literature
would lead us too far afield from our present main pur-
pose, which is simply to present the data and state the
apparent conclusions from an experimental and statis-
tical study of the influence of starvation and feeding upon
the characteristics of garden beans. It is sufficient for
the moment to point out that some biologists have attrib-
uted a very important rôle to the environment of the
mother in determining the characteristics of the off-
spring. It is perhaps superfluous to say that others of
equal authority have expressed diametrically opposite
opinions.

The problem is, therefore, a real and an important one.
Unfortunately the serious investigator who publishes in
this field is sure to be between two large and several

313

314 THE AMERICAN NATURALIST [ Vou. XLVI

smaller fires. If after cultures of a few generations he
finds that the offspring of starved parents do not differ
from those which have been well fed, he will be railed at
for having wasted his time in demonstrating what was
obvious in advance. At the same time he will be criti-
cized by others for not having carried out his experiments
‘for a sufficient number of generations to allow the accu-
mulation of small effects of the environment’’ on the
ascendants before deciding against the possibility of some
influence upon the descendants of ancestral environ-
mental conditions. If he finds that there are measurable
differences between series of individuals whose ancestry
has been subjected to opposed conditions, the results are
sure to be dismissed in many quarters as of little impor-
tance because of purely physiological and not hereditary
significance.

The very fact of the inevitability of criticism—what-
ever the results obtained—seems to render it even more
highly desirable to appeal to the facts afforded by a large
and detailed experimental investigation. Naturally such
an experiment can never be so large and so refined as to
be beyond all criticism.

The problem is not merely of wide interest from the
purely biological viewpoint, but it is of first rate impor-
tance from the practical side as well. The biggest pump-
kin, the heaviest bull, and the finest ear of corn are the
resultant of germ plasm and environment—of nature and
nurture, to use Galton’s apt words. But in paying fabu-
lous prices for the seed of prize winners little thought is
given to the question of the proportionate importance of
breeding and feeding in producing this excellence. From
the practical standpoint it seems desirable to know —
whether parents—animals or plants—of as nearly as
possible the same hereditary endowment differ at all in
their capacity for producing high-grade offspring because
of the superior care and feeding which admits them to
the show bench. If it be found that the well-fed mother
produces finer, or poorer, offspring than the starved one,
the practical significance of the result is obvious and the

No. 546] INFLUENCE OF STARVATION 315

further biological problems of the nature and permanence
of this influence will be open for investigation.

Finally it may be said in passing that the work on these
beans was so carried out that data for many other prob-
lems besides those discussed here were secured. That of
the pure line, that of the relationship between the size of
the seed planted and the characteristics of the plant
produced, that of the relationship between the size of
the plant and the fertility of its pod and the size of the
seeds which it produces, that of the relationship between.
the ovule characters of the pod and its fertility, may be
mentioned. These will shortly be made ready for publi-
cation; hence if the reader encounters these series of
beans in several different places he must not assume du-
plicate publication. The mass of data in hand is so great
that it is either necessary to scatter the material in this.
way or to withhold it all for several months or years until
it can be presented in one volume. The former scheme
for several reasons seems the most expedient.

Il. STATEMENT OF PROBLEMS AND DESCRIPTION or MATE-
RIALS AND METHODS

A. Limitation of the Problem

The purpose of this paper is to present the results of a
series of experiments to determine whether plants whose
ancestors have been starved differ from those whose an-
cestors have been well fed.

It might seem to the reader that the first step in such a
problem would be to define starvation and feeding, to list
the factors underlying these conditions, and to ascertain
the weight of each of these factors in determining the
characteristics of a series of plants subjected to them.

This seemed to me in undertaking these particular ex-
periments precisely the course which one should not fol-
low. Physiologists, especially those concerned with plant
nutrition in the agricultural stations, have devoted a
quarter of a century or more to these very problems.

316 THE AMERICAN NATURALIST (Vou. XLVI

But concerning the influence of the feeding or starving of
the parent upon the characteristics of the offspring, we
have little direct experimental knowledge.

It seemed expedient therefore to neglect for the mo-
ment the problem of the various edaphic and metereolog-
ical factors which determine the characteristics of the
individual and to ascertain whether the subjecting of
parent plants (or parents and earlier ascendants) to dif-
fering environmental conditions has any influence upon
the characteristics of the offspring. It was therefore
only necessary to find fields in which the soil barely sus-
tained a given variety and others which produced a luxu-
riant growth. The first would represent for the species
in question starvation fields.

The judgment of the relative richness of the plots by
their actual productiveness is justified by our ignorance
of the nature of soil fertility.

The reader who is inclined to criticize this method of
approaching the problem as very coarse may be reminded
of the following points:

(a) The complexity of the problem of soil fertility is
such as to preclude a trustworthy evaluation of the par-
ticular factors determining the productiveness of any
parcel of ground. For this reason I have purposely
omitted all but the barest descriptions concerning the ex-
perimental plots employed.

(b) Artificial soils or water culture media of known
chemical composition were carefully considered and ruled
ont. In the first place, the technical difficulties seemed
almost unsurmountable. Again, it seemed desirable to
earry on the experiments under conditions as nearly as
possible identical with those to be met with in practical
agriculture. Chemically prepared nutrient solutions are
useful in the physiological laboratory, but they do not
occur in practical farming, while soils which are ‘‘sterile”’

*Soil experts now agree that chemical analyses of soils furnish no sure
criterion of their productiveness.

No. 546] INFLUENCE OF STARVATION 317

and those which are ‘‘productive’’—for what reason we
do not know—do.?

The solution of our problem is to be sought by means
of a series of comparisons which fall into two classes.
The first is designed to test the influence of the environ-
ment upon the characteristics of the individual; the sec-
ond is intended to show what influence, if any, the treat-
ment of the ancestors has had upon the offspring.

The first series of comparisons is essential in that it
brings out clearly the extent to which the ancestors were
modified by the environment to which they were sub-
jected. It affords no evidence whatever as to the factors
to which these effects are due. The second set of com-
parisons is the important one. Our problem, the reader
must distinctly understand, is not to determine why some
individuals are depauperate and others luxuriant, but
whether the rendering of individuals depauperate
through the environment to which they are subjected has
any influence upon the measurable characteristics of their
offspring.

B. Material

The materials upon which this study was based were
furnished by five series of garden beans, Phaseolus vul-
garis. Two of these were the common white Navy. The
third was a strain of Burpee’s Stringless first grown
from commercial seed at the Missouri Botanical Garden
in 1905. The other two were from the seed of the White
Flageolet and Ne Plus Ultra which Dr. Shull had used in
his hybridization experiments.*

2Our great ignorance of the problem of soil fertility is attested by the
words of Professor Hall in a chairman’s address before the Sheffield meet-
ing of the British Association (Science, N. S., Vol. 32, p. 364, 1911). He
said:

d:
‘‘The fertility of the soil is perhaps a vague title, but by it I intend
f

to signify the greater or less power which a piece of land °
in, the causes which make one

small ones, differences which are so real that a farmer will pay three or
even four pounds an acre rent for some land, where he will regard the other
as dear at ten shillings an acre.’’

* Shull, G. H., Science, N, S., Vol. 25, pp. 792-794, 828-832, 1907; AMER.
Nat., Vol. 42, pp. 433-451, 1908.

318 THE AMERICAN NATURALIST [ Vou. XLVI

The two Navy series first came to my attention on the
farms of George A. Harris and Elmer Dille at Mount
Hermon, near Plantsville, Athens Co., Ohio, in the fall of
1907. From the Harris farm 160 plants were taken,
giving rise to 160 ‘‘pure lines.’ These are the Navy H,
or NH series. From the Dille field 550 plants were taken

ND
STARVED

WELL FED

COMPARISON FIELD
1910

Diagram I. Cultural history of the Navy D series. The history of the Navy
the same, and can be expressed by substituting H for D as the first
habitat letter in the formule.

and yielded 550 ‘‘pure lines,’’ designated as the Navy D
series. These two fields furnished, as explained in detail
in a subsequent section, the starvation and feeding tracts
of the experiment.

Dr. Shull’s seeds saved for individual plants of a crop
of 1907 yielded 80 lines of Ne Plus Ultra and 100 of White
Flageolet.

The history of these strains during the course of the

No. 546] INFLUENCE OF STARVATION 319

experiment is shown by the diagrams. The seriations of
number of pods per plant appear in the Data Tables 4,
B and C.
TABLE A
PODS PER PLANT

Series 1| 2 3 | 4 s|6 171819 | 10| 11 | 12/18 | 14| 15 | 16 | Total
| | | | | | | | | Plants

D 55 (229/165, 63| 24| 8| 4| 1 | i a SA a E p 550
DD 46 |107|141| 89| 57, 36/16 |11| 3| 4| 2|—|—| 1/—|—| 513
DDD |10 61| 93 107| 67 55/31 | 24 5 | 4|—| 1} 1|—|—;—|_ 459
HD '235|333 282/192 Wy 8)-7 1 Pt a ee
HDD |49 |172)234'208 204 var ede 34/22| 9| 9f 3| 3! 2| 1) 1,204
USD = |53 |111| 95| 30) 8} 3| —|—|— |—|—|— ||} |-| 312
USDD |25| 64| 42| 34| 33| 19/12) 3| 1) 2} 1/—|—|— | 1| 287
39 100/118 76| 43 26/12 13 —| 1;—|—|—|— -|—| 428

FSDD |13| 52| 98| 91! 64 43\/15| 4 RRE | 1: —+|—| - 387

For convenience of reference I designate the 1907, 1908
and 1909 cultures the ancestral series and the 1910 crops.
the comparison series. The fitness of these terms will be
apparent.

C. Experimental Methods and Collection of Data
Experimental methods may conveniently be explained
under three heads: Selection and Care of Seed, Cultural
Conditions, and Collection of Data. `

1. Selection and Care of Seed

The necessary requirements are two. First, it is essen-
tial that the material subjected to the various environ-
mental factors shall be identical in its hereditary tenden-
cies. Second, it is essential that in the routine of grow-
ing, harvesting and planting no purely physiological (as
contrasted with hereditary, germinal or genetic) sources
of differentiation shall be introduced.

Consider the first requirement.

We have learned from both biometric and Mendelian
researches that it is impossible to know from the simple
inspection of an apparently uniform group of individuals
whether or not they are really identical as to germinal
constitution. It is therefore idle to plant seeds of some
individuals under starvation and seeds of other individ-

320 THE AMERICAN NATURALIST [Von. XLVI

TABLE B
~ NUMBER or PODS PER PLANT
Series H | HH | HHH | DH \ DHH|\ USS | USH| USHH | FSS | FSH | FSHH
1 — ti 4 4&4 | = | [= — 4 | — —
2 — 8 10 10 GN hes 1 3 5 1 6
3 1 10 20 16 9 2 2 6 12 4 5
4 4 TI 36 21 | 12 2 5 Lt 21 4 H
5 3 25 52 20 | 10 3 5 15 24 9 9
6 + 34 62 20 | 26 6 6 24 6 20
7 6 41 78 29 | 39 | 16 | 15 27 25 | 14 26
8 T 55 91 26 | 42 | 17 | 20 38 38 | 12 34
9 10 58 82 36 | 43 | 3 17 28 43 | 22 32
10 12 | 97 43 | 52 | 48 | 40 | 23 | 67 31
11 76 91 | 35 | 51 | 4 17 15 24 33
12 12 78 96 39 | 37 | 54 | 22 12 43 | 22. 42
13 12 94 115 | 34 | 47 | 48 | 28 ri 65 | 22
14 9 74 34 | 40 | 49 | 25 3 47 | 24 34
15 7 2 60 28 | 35 | 52 | 23 3 50 | 25 25
16 T 83 72 35 | 19 | 37 | Z6 3 42 | 23 16
u 9 69 52 35 | 22 | 49 | 18 X 27 | 19 10
18 8 66 39 30: 12 3] B 4 46 | 27 13
19 5 56 29 24°) 17 | 32 | 20 — 31 | 19
20 1 25 | 21 | 11 3i | — | 36 | 17 10
21 i 51 25 16 5118 | 12 — 42 | 19 7
22 3 41 12 9 T 1.22 6 — 99 i 12 2
23 5 | 46 8 | 13 4 | 20 3 2 14 | 23 3
24 3 37 fi il 3 | 12 5 — 12 8 2
25 2 | 28 6 5 6 | 16 4| — ; 11} ll 1
26 2 27 6 8 4 | 10 2 — 16 | 12 1
27 2 19 1 9 2 T 3 — 9 | 12 2
28 4 12 2 Ti e A 5 4 | 13 =
29 1 14 1 10 am 2 1 — yi 7 seeing
30 2 23 2 5|—-{|— 1 8 2 a
31 1 20 = 2 D ire — 5 6 =
32 — 12 3 nA ag 1} — a 2 2 —
33 — 10 — 5i 3 | — -— 2 + ==
34 1 10 2 3 | — TA — 4 3 1
35 — 11 — es BE cae — 2 4 m
36 1 4 = 220 aa ee — 1 1 oe
37 1 7 — 4| — 2 | = — 4 3 Zer
38 no 6 — 3|ļ— | — — |— 1 —
39 = 5 — LC = 1 1 — Le =
40 i 5 -| — U care 1 — — 1 =
41 — 3 — — | — | — |— — — 2 oa
42 = 2 1| — | — = Ile ee
43 — 3 = |t 2 1 —
44 1 — fo pe a 1 -—
45 = 2 == 1 |— — — Iie
46 1 5 — EN Ra Ge nei — | — —
47 — 1 ae ee ee ee See lire =
48 — — Tie 2) — Lop = SEE
49 — 1 ne fee pe Gas MEE f pois
50 — ous een ee ee eg ae os La ean pie
51 — 1 Ga Gees A ome oe Wee aoe
52 — 2 a Kan EA ha E See —
54 — 2 L aad a aa Ea a T pie
55 = 1 a E E AT Baie ei set
56 — t oo ech ae f coe eh ek oe aes ae
67 1 se ER a, Cee a Pe gs. ==
Totalplants| 160 |1,484 |1271 |670 | 565 | 680 |361 | 224 |868 |475 | 429

No. 546] INFLUENCE OF STARVATION 321

uals of apparently the same uniform variety under feed-
ing conditions. The only certain method of securing the

TABLE C
NUMBER OF PODS PER PLANT
| |
Series Hea kE | ee beste | |
ME stale mi
HHC 8! 816 26 35 30 25/31/30 46 25/23 23 21 28 14|18| 5| 9| 8|12/12|11| 4| 4 23
HHHC 8 13 20'34 41/48 38 35 43 37 2636|25 24/19 17/10/14/15| 810| 8| 5| 3| 4 2| 3
HDC 6/10 2526|29 30/26 28 25 14/12|11/16 23 16] 9/10/11) 5| 5| 7| 5| 4| 4 31. 3\—
HDDC (11115/233536 3532/43 28/34/19|29/32/17,17|1019|13,15) 6| 5| 4| 4| 4) 1| 3| 2
DDC 616 12 12/20 22/26 14 6 8j16|15/15| 8| 7| 4| 5| 6| 1| 2) 1| 1—| 1) 1|—
DDDC 6/10/14 23/30, 27/21 22 21 161 si : me 7| 2| 1| 2| 4| 2| 1—
C 921/193 41,94132185i20.30, p3 18 1513 16 1 1113| 7| 5| 1| 7| 2| 2| 1}—
DHHC 7 17|30 38 45 4052 40 27 271911 1719 a n 9110| 6| 1| 4i 4| 4| 2| 4
SC 3| 5/15 22/2 25 31 a2 609 60 a7 28 28 31 30 29 15/10] 6| 6| 5| 3| 2| 1| 2) 1 1 2
USSC 1—| 5 $10 212931233143 24 2 19| 6| 8| 8| 6| 3| 3| 11/—| 3| 1
SHC 1) 1) zii 19 28/31/38 18 39 22/22/25 1 8| si 71 7| 3| 4| 1| 2) 1-H
USHHC |2|7 7 13 31 3131|36 3541 soe pel 12) 6/11] 9| 5| 7| 2| 1—| 1-1
SDC 510) 921 42/43 43 44 19 2 13115111| 8! 5! 2, —! 1— Em
USDDC |4| 117 10114222728 36136137 28/23/15|13/11| 9| 7| 1| 2| 1⁄1—| 1 —|—|—i—
SC 1| 6| 9 14/17/29 0 29 26 30 45 3831 21 31 29 27/13 1715181912 14/10| 7 6
FSSC 2| 6| 810/132 | 21308 35 26 25 25 = 3/17|10|10| 8| 4| 3| 2| 4
FSHC |—| 2| 9 18/20 33 302834 21 35/31|19/24/17|14 14| 91114) 7| 7| 8 6 5| 3 —
FSHHC |—| 7/15 20 36 3531 pae 7 41/39/30 21,19 aclauing 17/11) 8|10/12)10 7
FSDC ola es 6 9 14 24 17 21 23 20 26 2818) 9 16| 6| 7| 8| 9111] 4| 4| 1| 3| 3| 2
FSDDC |1| 2| 8110 14|23|25 3132/3024140 ae aolaaiasiis\isiolasiiel 9|13|15|10| 9 5
i LA e ehoketa at Total
Series itd ee |35] 42|43|44|46|47 49|51|52|55/79| piants
| i | | ENE cere ERTA Meron Scam
HHC agli 24 Li yg fe AT
auae | 1a hili HHA H HH HH 5
DC -4a 111132 Hi likre 376
HDD |——}2 i1 2 4-2 eee 498
DC 11 —— ———— —| 255
DDDC 3\—|—| 2 | 3 i—i i—mar 331
HC —|3 |1 |1 |2 | ——|—i 1 |2 |---|! —|—| 452
DHHC əl—| 1 |3 |1 1 i i k 598
USC al N EA LEI 580
USSC | ah saint: BSD
SHC | —|————— =l S21
USHHC |—|1 | 1 —i—|——— 399
SDC ik Fle TS ie Hon SS IE a ES = $50
USDDC | Fs es E Vat SI en A =| 938
FSC 2al3 6l3l6l1l1l2l1 2—1 H+ 586
FSSC iilh in at ee 50
FSHC 3/3 — 2 |2 |1 |2 |——————— iced’ a
FSHHC holle sls lalalala =s —|12-—--+ +I eS
FSDC 2\— 3 —|—|—-}1 | 1 |---| 1 = 307
FSDDC 2i3i3i6l2i-ti3eiie eee 538

desired result is to divide the seeds of individual plants.
If this be done and if all the lines* be represented —

‘Line or pure line is used merely in the genealogical sense.

San THE AMERICAN NATURALIST [Vou. XLVI

throughout the experiment by approximately the same
number of individuals, we shall not only be sure that like
hereditary tendencies went into all branches of the ex-
periment at the beginning, but can feel confident that no
material source of error is introduced by a change in the
mean hereditary tendencies in either branch of the ex-
periment through the selective elimination (by reason of
relative unfitness for the chosen habitats) of certain (dif-
ferentiated) lines.

These are ideal conditions, quite unattainable among
the innumerable difficulties of practical experimentation.
Omitting all particulars, I believe we may with reason-
able security consider the seeds which went into the orig-
inal starvation and into the original feeding series ran-
dom samples from the same individual plants. These
lines were maintained with moderate success throughout
the experiment.

The following details concerning the methods of ma-
nipulating the material may not be irrelevant.

Every seed was, so far as could be determined by in-
spection, perfectly formed and developed.” No seeds in
which the coats were sensibly wrinkled were included,
since this might indicate either a premature drying of the
seed in the pod, or a subsequent wetting.’

In harvesting, the plants were left in the field as long
as possible to allow the pods to ripen. They were then
gathered, and wrapped intact in newspaper to permit any
possible translocation of remaining plastic materials
from the stems or the pod walls to the seeds."

* Every seed was examined at least once. Unfortunately this can not
preclude the possibility of a seed containing a weevil which had not emerged

up to the time of planting. A large proportion of the seeds planted in these
experiments was also weighed individually for use in pure line and other

D, US, FS and BG individuals—

“I have carried out no experiments to determine what the real causes
are of this wrinkling.

* This precaution applies to only two of the original series, to NH and
ND, but not to FS, US and BG.

No. 546] INFLUENCE OF STARVATION 323

2. Cultural Conditions

Having prefaced that the purpose of this study is not
to determine what chemical and physical factors produce
in the individual the effects which we designate as star-
vation, we are free to choose for the ancestral series any
plots which present reasonably extreme conditions of
starvation and feeding.

The two fields in southeastern Ohio seemed perfectly
adapted to the purposes of the experiment. Their crops
of the common Navy beans presented the most diverse
appearance. The H field—that grown by Mr. Geo. A.
Harris bore a moderately heavy crop. The D field—
grown by Mr. Elmer Dille—seemed to have almost if not
quite as good a stand, but the plants were exceedingly
small.

The differences were apparently not due to variety,
for both were, in so far as could be seen, identical. They
were obviously not referable to cultivation, for both had
been equally well tended. The differences seemed en-
tirely attributable to the exceedingly poor soil of the D
field.

Minute description of these two fields is quite unneces-
sary. They were about a mile apart, and hence under
the same general conditions of climate. Neither was
level. Field H was much longer than wide and sloped
from the ends towards the middle, where the ground was
apt to be a little too damp. Plot D was situated on an
exposed ridge where practically all the surface soil had
washed away.

The plants originally growing upon these fields formed
the starting point for the starvation-feeding comparison.
This was in the fall of 1907. In 1908 transfers were
made, in order that we might be sure that genotypically,
as the pure linist would have it, the plants cultivated on
both fields were the same. Other varieties were also
added in 1908. These points are made quite clear by the
diagrams.

The comparison furnishing the test of the influence of

324 THE AMERICAN NATURALIST [Vou. XLVI

the D and H conditions upon the offspring should not
be made on either of these fields.®

Three fields? under control of the Station for Experi-
mental Evolution at Cold Spring Harbor were chosen
for the comparison. All the Navy series were tested

USH
T FED

USS
WELL FED

TA
ao FED

COPPA FIELD

Diagram 2. Cultural history of the Ne Plus Ultra ee series. The White
Flageolet (F) series was subjected to an identical treatmen

on one field and all the Ne Plus Ultra and White Flageo-
let on another. The third field was devoted to the

*A number of breeders hold that among plants there is a gradual adap-
tation to the substratum; when plants are transferred from one locus
to another there is a ‘‘new place effect.’? Tf, now, the series grown OP
the starvation field for two years should from some such process of adap-
tation be better able to thrive under these conditions than a series newly
transferred there from a rich soil, or vice versa, the comparison would be an
6 oan unfair one.

The facts bearing T this point derivable from our material will
probably be discussed lat

Three were selected Pae accidents of season and culture do occur
and it is as unwise to plant all one’s experimental seed on a single field
as to carry all one’s pedigreed eggs in one basket.

No. 546] INFLUENCE OF STARVATION 325

fourth variety, BG, which must be reserved for a later
paper. .

The following method was adopted for counteracting
the possible hetereogeneity!® of the fields upon which
the plants were grown.

The different strains must be subjected to as nearly
as possible a random sample of the conditions afforded
by any plot. This end is secured by labeling each seed
individually and then scattering those of a particular
series quite at random over the field. If, then, certain
spots are somewhat more fertile or slightly moister
than others, all lines will have equal chances of being
represented there. If this were not done an undetected
differentiation in the substratum might induce quite de-
ceptive differences in the crops.

In these experiments I did not, unfortunately, work to
quite this degree of refinement. For technical reasons,
it was desirable to have each of the varieties planted in
separate rows. Each seed was placed in an individually
labeled envelope and the envelopes of a series thor-
oughly shuffled. The series were then planted in rows,
which were scattered as nearly as possible at random
across the field. By this means an almost but probably
not quite random distribution was secured.

3. Collection of Data
The recording of the data from the mature plants was
an onerous but relatively simple process.
As noted above, the plants were wrapped individually
at harvest time when as nearly dry as they could be
* Conditions were sp worse than those under which much of the experi-

mental evolution work has been done. At the same time I must frankly
confess that to the biometrician the comparison fields left much to be
d

available. In stating that conditions are in defect of those desired by the
biometrician, we may perhaps remember that they have the advantage K
presenting no experimental artificiality, but of being precisely the so
which would be met in ordinary agriċultural practise.
Entirely too little attention has been paid to these matters by experi-
mentalists. Compare,‘for instance, some suggestions in AMER. NAT., Vol.
6.

wan. THE AMERICAN NATURALIST [Vou. XLVI

allowed to become in the field, and after thoroughly dry-
ing stored until they could be studied. They were then
placed in a saturated atmosphere for a few hours until
the pods could be handled without snapping open, and
records made of the number of pods per plant and num-
ber of ovules and seeds per pod. The seeds were then
stored until thoroughly dried at laboratory temperature
and humidity, when they were looked over for weighing.
Particulars concerning the various characters will be
given in the special sections. }

D. Methods of Analysis of Data
1. Pertinent Comparisons

The possibility of an influence of ascendant starvation
upon descendant characters is to be tested by a series of
comparisons. The number which might be made, and
with profit, is so great that space requirements impose a
stringent limitation.

A first restriction is effected by basing the compar-
isons upon the simplest of the statistical constants.

A second limitation is effected by the exclusion of all
comparisons showing the relative influence of environ-
mental conditions on different varieties. Possibly this
question will be considered in another place. Such inter-
racial and inter-varietal comparisons are in this paper
quite incidental to those which are strictly intra-racial
and intra-varietal.

Finally the comparisons within the varieties must be
limited"! to those which seem absolutely essential to our
purposes. The constants for the 40 series are given so
that the reader may make any comparison he deems
desirable.

In the dichotomous system adopted for these experi-
ments, one branch of the stem material was subjected to

“In all we have three distinct varieties represented by 40 series of
material—18 of Navy and 11 each of White Flageolet and Ne Plus Ultra.
If all the 4n(m—1) comparisons within each variety were made for the

three constants, A, e, and CV, 789 differences and their probable errors
would have to be calculated for each character observed.

No. 546] INFLUENCE OF STARVATION 327

starvation and the other to feeding. Both ancestral and
comparison series allow of two kinds of comparisons,
intra-ramal and inter-ramal.

In the first case the comparisons will be made within the
same branch of the dichotomous system, i. e., the off-
spring of the starved parents and starved grandparents
will be compared with plants whose parents only were
starved, both parents and grandparents being in the
direct line of descent.??

In these tests the individuals grown on the comparison
field bear to each other the relationship of ‘‘aunts’’ and
“nieces.” Such comparisons are possible where the seed
retains its vitality for a number of years. They are
open to criticism unless it be known that the age of the
seed has no influence upon the characteristics of the
plants developing from them.*®

In the second class, the inter-ramal, are those compar-
isons between points on different branches of the dichoto-
mous scheme. Here two subclasses may be recognized.
In the one the comparisons are between strictly homolo-
gous points on the starved branch and on the well-fed
branch. The effect of one generation’s starvation will
be compared with the effect of one generation’s feeding.
In this case comparisons will be made between ‘‘first’’
and ‘‘first’? cousins. Or Mendelianwise, all individuals
compared will be F,, F, or Fx. Such comparisons will-
be called direct inter-ramal comparisons.

In the second subclass, the comparisons will be made
between different points on the two branches; all com-
” 32Tn the same manner any one who desires may compare plants whose
parents and grandparents were well fed with those whose parents only were
well fed. This is not done here for the simple reason that I do not know
that the well fed series were grown at an extreme of feeding at all com-
parable with the extreme of starvation which was possible in these experi-
ments. If they were not, one would expect to find a smaller influence, if
any, upon the offspring.

18 It may have occurred to the reader that a valuable comparison for our
purpose could be made within the starvation series by determining, e. g.,
whether USDD whose parents USD had been starved, had a lower value for
any character than USD whose parents US were not grown under starva-
tion conditions. Such tests are, however, useless because both edaphic and
meteorological conditions may differ from year to year. —

®

328 THE AMERICAN NATURALIST [Vou. XLVI

parisons will be between ancestral individuals or their
offspring belonging to different generations. Such will
be called cross inter-ramal comparisons.

The most crucial test is that afforded by the direct
inter-ramal comparisons. Both the intra-ramal and the
cross inter-ramal comparisons have the disadvantage
that the (possible) seed age factor is not excluded.
Again, atmospheric (meteorological) factors play a much
larger part where different seasons instead of a single .
season are involved.

Turning to our own available data, we note the fol-
lowing points concerning the comparisons:

Only such comparisons as can be made on the basis of
both ancestral and comparison series are discussed, al-
though data for some others, e. g., NH, ND, US, FS, USC,
FSC are given.

In all cases the differences are taken

Starvation
less
feeding

so that when starvation tends to reduce a character the
difference bears the negative sign.

If we continue our attention strictly to those within
the strain, we have the following inter-ramal compar-
isons:

Direct Cross
HD-HH HD-HHH
HDD-HHH . HDD-HH
DD-DH DD-DHH
DDD-DHH DDD-DH
USD-USS USD-USHH
USD-USH USDD-USS
USDD-USHH USDD-USH
FSD-FSS FSDD-FSS
FSD-FSH FSDD-FSH
FSDD-FSHH FSDD-FSH

If we go beyond the limits of the populations formed
by splitting the seeds of the same individual into two

No. 546] INFLUENCE OF STARVATION 329

lots, and consider the Navy D and Navy H comparable,
we get:

Direct ross
DD-HH DD-HHH
DDD-HHH DDD-HH
HD-DH HD-DHH
HDD-DHH HDD-DH

2. Statistical Formule Employed

Methods ample for all the needs of this study are fur-
nished by the simplest of the Pearsonian statistical for-
mule. The comparisons in the main are restricted to
those based on the mean, standard deviation and coeffi-
cient of variation.

These do not fully describe a population, but they fur-
nish more information concerning it than do any other
three simple constants, and are sufficient for our pur-
poses. The methods of calculation are now familiar
or readily accessible to all biologists. The original data
are available for any other comparison, e. g., that based
on skewness.

The chief possibility of untrustworthiness in the sta-
tistical constants seems to me to lie in a possible biolog-
ical source of error introduced by growing the compar-
ison series in rows instead of mixing all the individually
labeled seeds together and scattering them quite at ran-
dom over the entire field.'* If because of the irregular-
ity of the fields, some of the rows were subjected to
slightly better and some to slightly poorer conditions
than the average, and if the rows of an individual series
were not distributed over the field in a perfectly random
manner, a slight source of differentiation quite undetect-
ible by the statistician’s simple probable error would be
introduced. I suspect this to be the case, and conse-
quently our probable errors are perhaps too low as cri-
teria of the existence of differentiation due to the treat-
ment of the ancestry. i

Fortunately we are not limited to a single comparison,

“ As an extra precaution half rows were frequently used.

330 THE AMERICAN: NATURALIST [ Vou. XLVI

but- have- several . pairs. Any one of these might be
wrong in its indication of the influence of ancestral en-
vironment because of uncontrollable factors making for
heterogeneity on the comparison tract, but as long as
these factors differ from series to series in a purely ran-
dom manner, we shall expect to get trustworthy values
by averaging the results for the several comparisons.

This averaging may be done in one or both of two ways.
Most easily one'may simply note the number of alterna-
tive cases, above zero and below zero, and calculate the
probable error of either class by the formula

67449 VN X OX 5

since, unless. there be an influence of the treatment of
the ancestors, the probabilities of differences lying above
and below zero are equal. In the second case, the true
mean and standard deviation of the series of differences
may be obtained and the probable error of the mean
difference calculated by the familiar formula

E = 67449 —
Hue OV

It only remains to say that, except when specified,
Sheppard’s modification was not applied in the calcula-
tion of the moments.

HE PRESENTATION oF DATA AND COMPARISON OF
CONSTANTS

i Number of Pods per Plant in Navy, White Flageolet
and Ne Plus Ultra Beans

The purpose of this section is to present the data for
number of pods per plant in three varieties, represented
by 40 series and over 21,000 individuals, and to draw the
comparisons which may profitably be based upon them.
The other characters for these varieties and all of the
data for still another variety are reserved for later treat-
ment. This character, which is the most easily deter-

No. 546] INFLUENCE OF STARVATION 331

mined of any, is also subject to considerable possibility
of error. It is impossible to know from an inspection of
the matured plants that some of the pods have not been
lost by accident. Another difficulty is introduced by the
fact that some varieties of beans have a tendency to make
a ‘“‘ second growth’’ when they are allowed to stand in
the field after they are completely ripe. Unless frosts
are very late these second growth pods rarely mature.
If the plants be allowed to stand in the hope that they
will ripen these second growth pods, the normal crop of |

re i
90 g ee a8 Ga

ft -----0--—-0m | NDD f

ae ck
80 j 7 Lad i a L

ji : ——e——o =| NH jt
To oi Ai

HAA
ode
i /

ite

( 6 WA U Mo S H WD 35 w 6S
Diagram 3. Number of pods per plant in NDD, ND D, NHH series.

All series are reduced to a pércentage basis and the relative frequencies | summed

from the ing. The influenc ce a starvation in the reduction ‘ot the number

of pods is very conspicuous.

‘ara at eee a ee ee) ee ee

j

4

er may either lose their seeds, if the weather be E
or decay if the weather be wet. All that can be done is
to watch the plants carefully, to harvest as soon as prac-
tically-all the pods that are ripe, and to pull off any sec-
ond growth sprouts. This apparently introdùces a con-
siderable personal equation into the work, but even if true
it is unavoidable. I do not believe that a palpable source
of error was introduced since (a) a large proportion of
the plants do not show the second growth at all; (b) when

332 THE AMERICAN NATURALIST [ Vou. XLVI

they do, a person with a little practise will make very
few mistakes; (c) even if errors are made, the treatment
is the same for all series.

All the pods counted had at least one matured seed.
This specification is necessary since, especially in the
autumn, some plants produce quite a number of half de-
veloped and completely sterile pods. If these were in-
eluded there would be no point where a line could be

WHITE FLAGEOLET F NE PLUS ULTRA
FSD AND FSH USD AND USH
STARVED, 1908 STARVED, 1908

| : FED, 1908 FED, 1908

L L n p aa

0 410 20. Jo 40 50 0 10 20 ag 40

ams 4 and 5. Percentage frequency of number of pods per plant under
sta and feeding conditions for White’ Flageolet and Ne Plus Ultra series,

drawn between the number of flowers and the number
of pods produced by an individual.
The record forms do not interest the general reader.
The original data are given in Tables 4A—-C. The phys-
ical constants appear in Tables I-III. :
The extreme sensitiveness of the number of pods per
plant to environmental conditions is seen at once from an

No. 546]

INFLUENCE OF STARVATION

333

inspection of the tables of raw data, and better still from
the three graphs, diagrams 3-5, for the number of pods
per plant in the 1908 series.1®

We may now summarize as briefly as possible, and
largely by diagrams, the results which may be gathered

TABLE I
Series Mean and Probable Standard Deviation (Coefficient of Variation
Error and Probable Error and Probable Error
NH 15.2375 = .4042 7.5800 + .2858 49.746 = 2.294
NHH 16.9919 + .1518 8.6696 = .1073 51.022 = 0.781
NHHH 11.9308 = .0977 5.1652 + .069 43.293 + 0.679
HD 3.9682 -0348 1.9433 + .0246 48.972 = 0.755.
NHDD 5822 .046: 2.3756 + .0327 51.844 + 0.884
D 2.6782 + .0335 1.1662 + .0237 43.545 + 1.040:
DD 3.5926 = .0563 1.8917 + .0398 52.657 1.384.
NDDD 74 1.9329 = .043 43.855 1.149
DH 14.6179 = .2148 8.2422 = .1519 56.385 = 1.330
NDHH 11.8265 = .1408 4.9595 = .0995 41.935 = 0.978
NHHC 11.9597 = .2212 7.3010 = .1564 1.047 + 1.
NHHHC 10.6498 = .1788 6.2390 = .1264 58.583 = 1.541
DC 10.9362 = ‘2747 7.8970 = .1943 72.210 = 2.539
NHDDC 10.2851 + .1 6.1042 = .1305 350 =
NDDC 9.3098 = .2259 5.3470 = .1597 57.434 = 2.210
NDDDC 9.9819 + .2360 6.3673 = .166 63.789 = 2.252
DHC 9.9801 + .2079 6.5532 = .1470 662 d
NDHHC 9.9851 + .1827 6.2839 = .1292 62.933 = 1.732
TABLE II
Series Mean and Probable Standard Deviation (Coefficient of Variation
Error and Probable Error. and Probable Error
USS 15.7382 = .1562 6.0379 = .1104 38.365 = 0.798
USH 14.0416 + .1972 5.5542 + 94 39.555 = 1.138
USHH 8.4375 = .1462 3.2439 = .1034 38.446 = 1.250
SD 2.5929 04 1.2265 =. 1 47.300 = 1.536
USDD 3.6203 .0919 2.0986 =. 57.970 + 2.322
USC 10.1434 = .1285 4.3846 =. 43.226 + 1.050
USSC 11.7068 + .1594 4.6188 = .1127 39.454 + 1.102
USHC 9.9844 + .1541 4.0936 = .1090 41.000 1.262
USHHC 10.1 .1564 4.6299 = .1106 45.794 1.
USDC 8.4474 = .1331 3.8477 = .0942 45.549 + 1.326
USDDC 10.1231 + .1426 3.8579 + .1008 38.110 + 1.132

3 The 1908 instead of the 1907 series was chosen for these graphical
comparisons, since the number of available series is cnr HR as com-

pared with two—and since the num
giving much epe results.
eader may care to make.
quite comparable, they have been reduced to a percentage In the
first where data for four series are laid side by poal the per-

mn t

ber of individuals is
The data are available for any ep A com-
nder. the results for all series

To re

uch greater,

basis.

centages have been summed from the beginning for each pod class. In the
ney of each cae of

second

tage freque
pods per plant is represented by the height of a line.

and third

the percen

334 THE AMERICAN NATURALIST [ Vou. XLVI

TABLE III
Serica Mean and Probable Standard Deviation | Coefficient of Variation

Error and Probable Error and Probable Error

FSS 15.0265 + .1697 7.4134 = .1200 49.335 = 0.974
SH 17.2947 + .2456 7.93 = .1736 45.889 = 1.197
FSHH 11.8415 + .1562 4.7959 = .1104 40.501 = 1.075
SD 3.4252 + .0552 1.6929 + .0390 49.424 + 1.390
FSDD 4.0362 + .0593 1.7294 = .0419 42.848 + 1.215
C 14.2218 = .2056 Ta = 1454 51.895 = 1.268
FSSC 12.9562 + .1856 6.1678 = .1313 47.605 + 1.222
FSHC 12.2431 += .2068 6.3729 = .14 52.053 = 1.483
FSHHC 14.1505 + .2115 7.9996 += .1495 56.532 = 1.353
DC 9055 = .2616 6.7949 + .1850 52.652 = 1.787
FSDDC 14.4981 += .2090 7.1885 = .1478 49.583 = 1.245

from the tables of constants. As already emphasized
the comparisons between the ancestral series are of in-
terest for our present purposes only in so far as they
furnish proof that the parents of the comparison series
were conspicuously differentiated in type and variability
because of the environmental conditions to which they
were subjected. The reader must always keep in the
foreground the fact that our problem is not to determine
in detail what the causes of this differentiation are, but
merely to show that a conspicuous differentiation exists
and to ascertain whether it has any weight in determining
the characteristics of the offspring.
he differences between the starved and well-fed an-
cestral series are so well marked that constants are best
represented by graphs for all the series. In diagrams 6
and 7, which embody data for all possible comparisons
for A and v, roughly made, the key number of the va-
riety is given along the left-hand margin. The value of
the constant for the ancestral series is indicated by the
position of a solid dot when the series is a starved one,
and by the position of a circle when it is a well-fed one.
The value of the constant for the offspring of each of
these ancestral series grown upon the comparison field
is shown by the position of a solid square under a sep-
arate scale. Thus the key to the comparison series is
given by adding C to the formula for the ancestral series.
The graphs for the mean number of pods per plant and
for the standard deviation of number of pods per plant

No. 546] INFLUENCE OF STARVATION 335

brings out with great force and clearness the following
facts:

(a) The difference between the ancestral series sub-
jected to the S and H conditions and those subjected to
the D environment is very great. In all cases means and

SCALE OF MAGNITUDE FOR
ANCESTRAL SERIES

13 04 15 86 17

SCALE OF MAGNITUDE FOR
COMPARISON SERIES
9 0

| arn een ER. reo ae ee 412 68 as

BER RO. etek: |

STRAIN AND SERIES OF PLANTS

6

----4

---"

Diagram 6. Mea
Subjected to various

n number of pods
conditions of

sta

per plant in ancestral, or as

cendatit, seri
. the oi

to environmental conditions. The comparison series, however, show much smaller

ow
> ny
differences, and no clear indications of -an influence of the ancestral conditions.

standard deviations are conspicuously higher when the

plants are well fed than when they are starved. _
(b) There are considerable differences ‘between a

strain grown on the same field in different years.

336 THE AMERICAN NATURALIST [Vou. XLVI

Season is evidently a large factor in determining number
of pods per plant. This is most striking in the means
but it is also detectible in the standard deviations. For
the means we note that in each of the four strains the
average was conspicuously lower in 1909 than in 1908 on
the H field and slightly higher in 1909 than in 1908 on the

SCALE OF MAGNITUDE F SCALE OF MAGNITUDE FOR
ANCESTRAL is COMPARISON SERIES
o e428 E k o 2s hee ee
NAVY 4
NHH =
NHHH E |
NHD tr a
NAVY
w s
a NDD a
z L
z NDDD -4-- a
= P E arcane Oe le ade sa E]
n
“ULTRA
faal
m uS
=] USS O m
a =
< USH Pe
zZ
< USHH a
x i
Ps UsD z
FLAG.
FS |
FSS x a
FSH t= a
FSHH a
FSD 4 =e

Diagram 7. Standard deviation of number of pods per plant in ancestral
and comparison series. Compare the explanation of diagram 6.

D field. The standard deviations show precisely the
same results for the H field, but the differences between
1908 and 1909 for the D crops are relatively small.

(c) The differences between the comparison series are
considerable, but it is impossible to be certain of any in-
fluence of the treatment of the ancestors.

No. 546] ` INFLUENCE OF STARVATION 337

(a) and (b) are facts to be expected from the common
experience of all those who have occupied themselves ex-
tensively with the growing of plants; they are sum-
marized here merely because it is idle to discuss (c)
unless the results for (a) are clean cut.'®

Turn now from diagrams to physical constants. Con-
sider first the intra-ramal comparisons, those cases in
which individuals whose ancestors have been starved for
a longer period are contrasted with individuals in the
same line of descent whose ancestors have been starved
fora shorter period of time. The necessary constants
appear in Table IV.

TABLE IV

i R Ancestors Starved _ Ancestors Starved 1r
Description of Material irs Goneri on hree pa ack

Ancestors starved for one generation:
USDC series:

Standard deviation............. + 0.0102 + 0.1378
aise of papain epee acres — 7.439 = 1.743
FSD FSDDC series:
i E AE Mune eel e bo Wine es alee + 1.5926 + 0.3348
rrp Vee Covina. 2. cis Corah 0.3936 + 0.2369
ə _ Coefficient of variation.......... — 3.069 + 1.178
NHDDC series:
PE EA Bho rb 6 views 0.6511 0.3309
Standard deviation............. — 1.7928 = 0.2341
Coefficient of variation.......... —12.960 = 3.031
Ancestors starved for two generations o
: se
CANE ARE E ee + 0.6721 = 0.3266
andard deviation............. + 0.9203 + 0.2311
Coefficient of variation.......... + 6.355 = 3.155

Two of these means seem to be significant in compar-
ison with their probable errors, and both of these indicate
that starvation of the ancestry for two as compared with
one generation, increases the number of pods on the off-
spring plant. But, it must be remembered that the seed
is necessarily a year older for a single generation of
starvation only. Furthermore, the series are too few
and the differences are entirely too small—only 1.6 pods
—to lay particular stress upon it.

The second set of comparisons, the inter-ramal, those

* Those noted under (b) may be treated more fully later.

338 THE AMERICAN NATURALIST [Vou. XLVI

between individuals whose ancestors had been subjected
to distinctly unlike treatment is made in Tables V-VIII.

Consider first the means. Altogether there are 28
inter-ramal comparisons, direct and cross. The number
of pods is smaller in the plants whose ancestors had been
starved in 16 out of 28 cases. If there were no relation-
ship between the conditions to which the ancestors were
subjected and the number of pods which their offspring
produced, one would expect 14 to be negative, providing
the errors of random sampling had not to be allowed for.
But the probable error is

6145V 28 «5x 5179.

Clearly a difference of 2+ 1.79 has no significance.

If now we restrict the comparison to differences signifi-
cant with regard to their probable errors, and consider
Diff./Eairr. > 3 to be significant, we note that only 11 out
of the 28 differences may be regarded as statistically
trustworthy. Of these, 9 have the negative and only 2
the positive sign. Certainly this looks as though there
were a very slight effect of the starvation of the ances-
tors, but nevertheless an effect quite detectible by the
statistical methods.

This point may be tested further by takin ‘the aver-
ages, regarding sign, of the pertinent differences for the
series of the three varieties. To make sure that slight
racial differences between ND and NH do not obscure the
results we recognize two classes of comparisons, within
the strain and between strains. The results are:

Navy, Within Strains, A == — .515
etween Strains, A = —. .515
General Average, A =— .515

Ne Plus Ultra, A= — 1.315
White Flageolet, A= + .585

In all cases except the White Flageolet aie! the
number of pods is s slightly lower when ihe ancestors have
‘been starved. |

* Note also that the two cases of significantly positive differences occur
in the White Flageolet variety.

No. 546]

INFLUENCE OF STARVATION

TABLE V

Description of Material

| Ancestors Starved for

One Generation
NHDC

339

Ansira Stirred foi
Two Generations
NHDDC

Ancestors well fed for one generation:
NDHC series:

Ancestors well rg for two generations: |
NH s = serie

Cat Eels. EE A E

MOS n aa eres ERS

Standard Pa aE a s A See

ent of variation

ine well fed for three generations:
NHHHC series

Standard oii :
Coefficient of variation..........

Lak 0.9561 + 0.3445
| + 1.3438 + 0.2437
| + 6.548 + 3.238

|

| — 1.0235. =-0.3527

| + 0.5960 = 0.2494

| +11.163 = 3.071

| + 0.9511. + 0.3298
+ 1.6131 + 0.2332

| 49.277 = 3.073

.2864 = 0.3277
| + 1.6580 + 0.2319

+ 13.627 = 2.970 `:

+ 0.3050 + 0.2780
+ 0.4490 + 0.1965
— 6.312 + 2.604

— 1.6746 + 0.2881
— 1.1968 = 0.203°
— 1.697 + 2.393 -

+ 0:3000 + 0.2596
— 0.1797 + 0.1836
— 3.583 = 2.396

— 0.3647 = 0.2569
— 0.1348 = 0.1817
+ 0.767 + 2.262

TABLE VI

Description of Material

| Ancestors Starved for
| Two pene Ws ations

Ancestors Starved for
Three pen rege
NDD

Ancestors well fed for one generation:
NDHC series:
MGatis i046 hia eee vis
oefficient of variation
Ancestors well fed for two generations:
NHHC series:

Standard deviation.............) =

NDHHC series

Oo He oe E TET E a Se ee a oc ie

‘Coefficient as Variation....54% ini.

Ancestors well fed for Tee generations:
C series

en soe sae ATE moe N

CoeMeient of Saini Pepe PRI

|
— 0.6703 = 0.3071
— 1.2062 = 0.2170
— 8.228 = 2.98
S z pose + -on
2236
> 3. ory + ae 805
— 0.6753 + 0 .2905
0.2054

— 0.9369 +
— 5.499 = 2.807

‘— 1.3400 + 0.2881
— 0.8920 + 0.2037

+ 0.0018 + 0.3145
— 0.1859 + 0.2225
— 1.873 + 3.019

— 1.9778 = 0.3234
— 0.9337 =
+ 2.742 + 2,838

— 0.0032 =
+ 0.0834 = O10
-+ 0. ps .841

— 0. .2961
+ 0.1283 + 0.2093
+ 5.206 = 2.729

= 1149 + 2.694

Consider now only the ten direct inter-ramal ‘ind the
ten cross inter-ramal, forming the twenty possible intra-

varietal comparisons.

comparisons which are availabl
seven have the negative and three the positive sign. |

Of the ten direct intra-ramal
e from the four series,

Tn

two cases only is Diff. /Baice > 3, and in one case i

340 THE AMERICAN NATURALIST [Vou. XLVI
TABLE VII
Ancestors Starved for | Ancestors Starved tor
Description of Material One Generation Two Generations
USDC USDDC
Ancestors bal fed for one generation:
— rie
E E oo E re ee — 3.2594 = 0.2076 | — 1.5837 = 0.2138
Standard GOVIRLION pe. E a — 0.7711 = 0.1470 | — 0.7609 = 0.1513
pba of VOR, oo E. + 6.095 = 1.724 | — 1.344 = 1.579
wahe
Ste ae ey cn E eee — 1.5370 + 0.2037 | + 0.1387 = 0.2100
Standard dorato a cae eo, — 0.2459 = 0.1442 | — 0.2357 = 0.1483
Coefficient of injec Bogie acs + 4.549 = 1.830 | — 2.890 = 1.695
Ancestors well ee rations:
USHHC se
Maer a yp a i eee es — 1.6629 + 0.2054 | + 0.0128 = 0.2117
Standard deviation............. — 0.7822 + 0.1453 | — 0.7720 + 0.1497
Coefficient of veins E T E — 0.245 = 1.859 | — 7.684 = 1.726
TABLE VIII

Description of Material

Ancestors Starved for
ne Generation
FSDC

Ancestors Starved for
wo Generations
FSDDC

Ancestors well fed for one generation:
ries:

ee ee eee reer eset ete eee tees

ee ee R ee AS

FSHC
wane oe déviation. o r.a
Coefficient of ane PEE es
Ancestors rane fed for two generations:
FSHHC series
I sires Bare ier ae mew i

All these are negative.

— 0.0507 + 0.3208
+ 0.6271 + 0.2269
+ 5.047 = 2.165

+ 0.6624 + 0.3335
+ 0.4220 + 0.2358
+ 0.599 = 2.322

— 1.2450 + 0.3365
— 1.2047 + 0.2379
— 3.880 + 2.241

+ 1.5419 + 0.2795
+ 1.0207 + 0.1977
+1.978 + 1.744

+ 2.2550 + 0.2939
+ 0.8156 = 0.2078
— 2.470 + 1.936

+ 0.3476 + 0.2973
— 0.8111 + 0.2102
Bae 83

There are no statistically signifi-

cant positive differences, the actual values being .013
+ .212, .348 + .297, and .662+ .334. The two larger of
these occur in the White Flageolet series.

the ten direct comparisons is — .589 pods.

The mean for

Of the ten cross inter-ramal comparisons, five are neg-

ative and five are positive; six are significant with regard _
to their probable error, four with the negative and two
with the positive sign. In both cases of positive differ-
ences (i. e., where the seeds from starved ancestors pro-
duced more pods than those from fed ancestors) the seed
from the fed plants was a year older than that from the
starved plants. The average*for the cross comparisons
is —.262 pods.

No. 546] INFLUENCE OF STARVATION 341

Consider the standard deviations.

As already noted, and as is clearly to be seen from the
graph, the standard deviations for the starved and fed
ancestral series show differences agreeing in general with
those seen in the means. This is to be expected, since
A and e are generally closely correlated. For this rea-
son it is idle to discuss the influence of starvation or
feeding upon variability on the basis of the standard
deviation alone.

Turning to the comparison series, we note that of the
28 differences, taken altogether, 17 are negative and 11
positive. The deviation from expectation is therefore
3+ 1.79, and can not be asserted to be significant.

Again taking Diff./Hair. > 3 as indicating differences
significant with regard to the errors of sampling, we note
that 17 cases out of 28 are statistically significant. Of
these 17 cases, 12 are negative and 5 are positive. Con-
sider averages as before:

Navy, Within Strains, A = — .165
Between Strains, A = — .053
General Average, A = — .109

Ne Plus Ultra, A= — 595

White Flageolet, A= + .145

Again limiting comparisons to the strictly intra-
varietal, and segregating into direct and cross inter-
ramal comparisons, we find that of the ten direct com-
parisons possible in the four lots, six are negative and
four positive in the sign of the difference. Only four are
statistically significant, i. e., Diff./Eart:. > 3, and all are
negative. The average is —.221. Of the ten cross inter-
ramal comparisons, seven are negative and three are
positive in sign; with regard to their probable error, eight
are significant; of these five are negative and three are
positive. The mean for the series is —- - 181.

Note the following points elative varia-
bilities as expressed by the coefficients. of variation.

Taken altogether, fifteen differences are negative and
thirteen are positive in sign. Accepting a difference of

342: THE AMERICAN NATURALIST [ Vou. XLVI

three times its probable error as statistically significant,
we note that altogether only six out of the twenty-eight
differences may be regarded as trustworthy. Of these
four are positive and two are negative in sign. Taking
means as for the two preceding constants, we find:

Navy, Within Strains, 4 = -+ .912
Between Strains, A = -+ .912
General Average, A = + .912

Ne Plus Ultra, A = —.,253

White Flageolet, A= — 946

With mean differences as slight as these, one certainly
can not argue that the starvation of the parents has had
any pronounced influence upon the relative variability of
the offspring.

j PROVISIONAL SUMMARY

1. The foregoing pages are devoted to a statement of
problems, description of methods and the presentation of
a first part of the data secured in a biometric investiga-
tion of the influence of the starvation of the ascendants
upon the characteristics of the descendants in garden
beans. Since several months will necessarily elapse be-
fore all of the materials can be worked up, it has seemed
undesirable to withhold the constants already calculated
and checked, viz., those for number of pods per plant in
three varieties represented by forty series comprising
altogether about 21,000 individuals. The publication is
therefore partial but in no sense preliminary. Several
questions that might be discussed on the basis of the data
presented are passed over until more series of material
can be lined up. The conclusions drawn—even for num-
ber of pods per plant—are provisional merely. =

2. The purpose of this research was not to ascertain
the physico-chemical factors to which starvation is due,
but to determine whether such artificial depauperization
of the ancestors has any influence upon the characters of
the offspring. Such ordinary ‘‘fertile’’ and ‘‘sterile’’ or
‘‘good’’ and ‘‘poor’’ agricultural land was therefore

No. 546] INFLUENCE OF STARVATION 343

taken for the ancestral series as would produce moder-
ately extreme conditions of depauperization and luxu-
riance in the crops.

3. The influence of from one to three generations star-
vation of the ascendants upon the characteristics of the
adult descendants is not conspicuous, in fact hardly to be
detected by the eye in the field. Statistical constants
seem, however, to show a slight yet unmistakable influ-
ence of the treatment of the ancestors in the form of a
slight decrease in the number of pods per plant.

4. The published data are as yet insufficient to justify
any discussion of the question of the cumulative influence
of the starvation conditions, or of the mechanism through
which the characters of the offspring plants are modified.
Evidence on these and various other pertinent questions
are being gathered as rapidly as possible.

MENDELIAN PROPORTIONS AND THE IN-
CREASE OF RECESSIVES

PROFESSOR FRANCIS RAMALEY
UNIVERSITY OF COLORADO

REcENTLY in working over some data! on the inherit-
ance of lefthandedness, certain questions came up,
which seem of considerable interest, as: Does the pro-
portion of lefthanded people remain the same from cen-
tury to century or does it diminish or increase? In any
case, how does the result come about? Although well
aware of the present-day aversion to arm-chair biology,
it yet seems that these problems can hardly be attacked
from the experimental side, and that a theoretical dis-
cussion may be of some value.

It may be stated at the outset that I consider left-
handedness to be a true Mendelian recessive,? and also
that there is no selective mating with reference to the
character. There seems also no reason to suspect that
lefthanded people exhibit less fertility? than normal in-
dividuals. If these suppositions are correct, the con-
dition offers a happy opportunity for study, since most
human characters thus far examined are such as might
be likely to be affected by selection.

Concerning the first question asked above, no positive
answer can be given, for there are no statistics. It is
probable that the affection is a very old one, and not of
recent origin. If it tends to increase it might be expected
that a very considerable part of the population would
now show the condition, while if it is decreasing, we

*As yet unpublished.

* Evidence for this view is shown by Professor H. E. Jordan in the

Breeders’ Magazine, Vol. II, pp. 19-29 and 113-124. My own observations
confirm this belief.

* My own records even suggest the opposite condition, but this is prob-
ably merely chance due to the small numbers studied.

344

No. 546] MENDELIAN PROPORTIONS 345

should, after all the long period of its existence, find
only a few persons with the condition.

_ On the ‘‘presence and absence’’ theory lefthandedness
is due to the loss or ‘‘dropping out’’ of the factor or de-
terminer for righthandedness. If such loss could occur
in the past, why not from time to time now? If so, why
would not the proportion of lefthanded people continu-
ally increase?

In a population where the dominant, the heterozygote
and the recessive have a certain proportion, slight
changes in the relative numbers seem to have no perma-
nent effect. There is a tendency to stability and unless
a certain point is passed because of the appearance of an
unusual number of recessives, there will be a return to
the usual ratio. The mathematical features of the case
have been discussed by Dr. W. J. Spillman,‘ and by Mr.
G. H. Hardy. The earliest clear statement of the case
which I have seen is by Dr. George H. Shull*® in his dis-
cussion of elementary species in Bursa bursa-pastoris.

The fact of the stability of certain ratios and instabil-
ity of others can be readily comprehended froni the fol-
lowing tables (I, II, III). We may let RR represent a
pure righthanded individual, Rr, a heterozygous right-
handed individual, and rr, a lefthanded individual. As
a first example it may be supposed that a large popula-
tion exists in which the various types occur in the fol-
lowing proportions, viz. :

1 pure righthanded:2 heterozygous righthanded:
1 lefthanded, or,
RR:2Rr: rr.
Random matings would occur in such fashion that mem-
bers of each group would mate with those of their own
group or with members of other groups. The various
possibilities are represented in Table I. The filial gen-
* Science, x S, Vol. XXVIII, pp. 252-254, 1908.

* Ibid., pp. 50.
« olenos, x a. Vol. XXV, pp. 590, 591, 1907.

346 THE AMERICAN NATURALIST [ Vou. XLVI

eration derived in Table I is composed of the three types
in the same ratio as in the parental generation.‘

TABLE I

MATINGS AND OFFSPRING IN A LARGE POPULATION HAVING THE COMPOSITION
1 RIGHTHANDED: 2 HeETEROZYGOUS RIGHTHANDED: 1 LEFTHANDED
E .

- Offspring .
Matings eee ed
RR Rr rr
HR SR eee ee 1

Pile Ct eo ei Ce ks ae eet cers 1 1

More See N ree 1

IRT X RR. osa es ee 1 1
OUr A Rr oa bas ook E 1 2 1
Bier KR PR ee ee 1 1

0 RR oe ae ek ee eee 1
i ARF ee ee 1 1
a OE es E ek ay 1
prem E ee a ean 4 8 4
a ye a a 1 2 1

As a second example, the ratio 2:2:2 may be taken.
This will give the results shown in Table II. The ratio
2:2:2 is not constant, but stability is reached in Fj,
which is found to show the same ratio as our previous
ease, Viz, 152+ 1,

TABLE II
MATINGS AND OFFSPRING IN A LARGE POPULATION HAVING THE PROPORTION

2 RIGHTHANDED: 2 HETEROZYGOUS RIGHTHANDED: 2 LEFTHANDED
(2BR : 2Rr : 2rr)

ffs
SO Offspring pee ee
RR Rr Tr
E
IRR x IRR yc, baa 4
SRR x2r 2 2
3RR x ar 4
R IRR a a 2 2
Gy x SB is a ae 1 2 1
Be Me. ea 2 2
Oy MSR. oe 4
Ber SCOR a a 2 2
MP ee a 4
Toa a ee cere OG Ie ae tad 9 18 9
OP ee 1 2 1

TI am indebted to my colleague, Professor Saul Epsteen, of the Depart-
ment of Mathematics, for checking my method of analysis.

No. 546] MENDELIAN PROPORTIONS 347

In Table III a few possible matings are shown with
the percentage of recessives and the constant ratios
which appear in the next generation. It is apparent
from examination of the figures that if a disturbance
takes place there is often a partial return in the next
generation to the original condition. Sometimes this
return is complete, as would occur if, beginning with the
ratio of 1:2:1, some disturbance should make it 1:1:1.
In the next generation there would be a return to the
1:2:1 condition. There were 25 per cent. of recessives
at the beginning. This became changed to 33.3 per cent.
by mutation but returned in the next generation to 25
per cent.

TABLE III
EXAMPLES OF RATIOS. OF DOMINANTS, HETEROZYGOTES AND RECESSIVES,
iy

TOGETHER WITH THE CONSTANT RATIOS DERIVED IN THE
FILIAL GENERATION. THE CONSTANT RATIOS ALL HAVE THE FORM
+ BP

A? + 2AB

Original Ratio Per Cent. Recessives | Constant Ratio in F, | Per Cent. Recessives
Ba 5 We i 33.3 25.0
Eh 2 50.0 o: Fa a 39.0
Gn Se 66.6 6: 9 56.2
L235 25.0 Constant Constant
£3331 20.0 25.0
L<t6:<2 33.3 25:70:49 34.0
1:4:4 44.4 Constant Constant
IRR Gey 40.0 1 x
4:4:1 11,2 (Constant Constant
4:6:1 9.0 766 :16 13.2

But there is not always a diminution of this kind. If
the original ratio be 4:4:1, which is constant, and this
be changed by mutation to 4:6:1, the original 11.1 per
cent. of recessives is at first reduced to 9 per cent., only
to rise in succeeding generations to 13.2 per cent. It is
very easy to overestimate the importance of the tend-
ency to return to a stable ratio, since such stable ratios
are practically without limit, and instead of returning
to the same ratios, it is easily possible to reach stability
in a new ratio somewhat different from the original.
This may be shown by an example (Table IV).

348 THE AMERICAN NATURALIST [ Vou. XLVI

TABLE IV
FIGURES SHOWING THE FORMATION OF A NEW STABLE RATIO IN A POPULA-
TION HAVING AT FIRST THE COMPOSITION 4:4:1 WHEN A SMALL

NUMBER OF HETEROZYGOTES BECOME RECESSIVES THROUGH ordan

Ratio in the original population: 400 : 400 : 100 — 11.1 per cent. recessive.

Ratio after mutation: 400 : 396 : 104 — 11.5 per cent. recessive.
Ratio of the next generation:

89,401 : 90,298 : 22,801 — 11.2 per cent. recessive.

The change i Im any case is not so simple as mathemat-
ical study would at first suggest, for many of the indi-
viduals mate with members of earlier or later genera-
tions. When a considerable number of recessive mutants
appear in a given generation this excess is, in part, re-
duced by matings with members of the normal genera-
tions before and after them. Hence a return to the nor-
mal or usual condition is made easy. These considera-
tions naturally suggest a reason for the comparative
stability of species within a genus or of elementary
species within a Linnean species.

From what has been said it will be seen that unless a
recessive character is arising by mutation at some ap-
preciable rate there will be little or no increase in the
proportion of individuals exhibiting this character. On
the other hand, whenever a considerable number of mu-
tants do appear the normal condition: of equilibrium is
disturbed and a new stable ratio becomes established in
which the recessives are in larger proportion.

There are probably many recessive characters which
are not the objects of natural selection by the environ-
ment nor of sexual selection in mating. Lefthandedness
may well be such a character. It certainly seems that
the proportion of individuals with such characters will
increase slowly through long periods of time. Let it
supposed, for example, that the proportion of dominants,
heterozygotes and recessives in the general population
is 9:6:1. Now it is probably far easier in mutation for
a given factor to drop out than to be added. If, however,
the dropping out and addition could take place with

No. 546] MENDELIAN PROPORTIONS 349

equal facility, there would still be much in favor of the
recessive. In the supposed ratio 9:6:1 it is possible for
any one of the dominants or heterozygotes to lose a de-
terminer. In other words, on the average fifteen out of
sixteen individuals have an opportunity to vary toward
lefthandedness; that is, 93.9 per cent. On the other
hand, only the heterozygotes and the recessive can vary
toward righthandedness; seven out of sixteen, therefore,
have this possibility, or 43.7 per cent. of the population.
If, therefore, the addition of the factor were as easily
accomplished as the loss of the factor—which is probably
never true—loss of a factor would be more than twice as
likely to occur as gain. With the ratio 1:2:1 opportuni-
ties are of course equal for upward and downward muta-
tion. In a population with more than 25 per cent. of
recessives, the number of heterozygotes plus recessives
is greater than the number of heterozygotes plus domi-
nants, as, for example, with the ratio 1:4:4, in which the
recessives make up 44.4 per cent. of the population.

It would seem from the above that there is a constant
tendency for the proportion of individuals showing non-
important recessive characters to increase, unless indeed
mutative changes occur so slowly that the population is
constantly held back to a stable ratio. We are led natur-
ally to the thought that for species in general the indi-
viduals showing recessive mutations may be expected to
increase at the expense of the original type, unless the
characters lost are of some real importance to existence
or to mating. Just this sort of thing has probably taken
place with the shepherd’s purse, Bursa bursa-pastoris.
It can hardly be doubted that the numerous species de-
scribed by Almquist! are derived from a single original
species. If the total number of Bursas in the world re-
mains the same from century to century and these ele-
mentary species appear in some number from time to

* Noted by Dr. George Harrison Shull, ‘‘ Advance Print from the Pro-
ceedings of the Seventh International Zoological Congress, Boston Meeting,

350 THE AMERICAN NATURALIST [Vou XLVI

time, they must be crowding upon the original form.
Other examples in great number will occur to any one.
Thus there can be little doubt that primitive man was
dark-skinned and dark-eyed. The various modern races
were produced—so far as skin-color and eye-color are
concerned—by dropping out the factors (determiners)
for these characters. In our own Caucasian race, it
seems not unreasonable to suppose that we are tending
toward a condition of blondness, for there seems to be no
natural or sexual selection against this character.®

I am not unmindful of the fact that an apparently un-
important character such as righthandedness may be
bound up with other characters of great consequence.
In such case lefthandedness might mean, for example, a
distinct inferiority in metabolic activity..°. Unless the
recessiveness is associated with weakened nutrition or
other enfeebled condition, it would seem most natural
for elementary species, when once originated, to increase
in number of individuals by continued mutation.

Dr. Shull points out in his paper previously cited on
“Elementary Species and Hybrids of Bursa’’ that reces-
sive mutants may have an advantage over dominant mu-
tants if fluctuating conditions tend to eliminate now one
form, now the other. The killing off of dominant mutants
may be easily accomplished, but this is not the case with
recessives.'' This is perhaps only another way of sta-
ting the well-known fact that it is hard in plant or animal

° The claim that light-skinned and light-eyed people are not adapted to
d

* Cf. the observations by Professor Thomas Hunt Morgan upon .mutants
of the fruit Pe Drosophila, in the Journal of Experimental Zoology, Vol.
XI, p. 408,

“To a own words: ‘‘The recessive mutant may be preserved
inde: gautis under the protection of the dominant characteristics of its more
successful parent. Such prolongation of the life of a recessive may serve
to tide it over times of special stress or may continue its existence until the
various distributing agents have carried it beyond the limits of the habitat
in which it is a failure into others in which it may become a success.’’
This same idea has been suggested by my colleague, Professor T. D. A.
Cockerell, in a conversation regarding the general question of increase of
recessives.

No. 546] MENDELIAN PROPORTIONS 351

breeding to get rid of recessive characters which are not
wanted.

SuMMARY

In the foregoing digegussion an attempt has been made
to exhibit clearly the facts of stable ratios involving
Mendelian dominants, heterozygotes and _ recessives.
While suggested by a study of lefthandedness, what has
been said will apply to any recessive character which is
not selected against in mating and which does not affect
the success of the organism in other ways. I have tried
to accord full value to the mathematical features of the
case, and have pointed out the various checks which tend
to hold the population in a given ratio. Yet I have been
unable to escape the conclusion that recessive mutants,
unless inherently weak in some respect, must tend to
increase in numbers at the expense of original dominant
types. These conclusions are reached from a considera-
tion of the following points: (1) The greater ease with
which characters may be lost than gained. (2) The great
number of combined dominants and heterozygotes which
through mutation may reach a simpler condition as com-
pared with the small number of recessives and hetero-
zygotes which may be imagined as affording opportunity
for mutation to dominance.!2 (3) The more likely sur-
vival of recessives in an environment of changing con-
ditions in which now the dominant and now the recessive
is hard pressed to maintain its existence.

12 Unless the ratio of the three types is 1:2: 1, when the opportunities
are the same in each direction.

THE INCONSTANCY OF UNIT-CHARACTERS!

PROFESSOR W. E. GASTLE

HARVARD UNIVERSITY

THERE can be no reasonable doubt that Mendel’s law
is of fundamental importance in genetics. It explains so
many of the anomalous facts and seeming contradictions
encountered in practical breeding. The basic fact under-
lying this law is the existence of unit-characters, inde-
pendently inherited. Their independence makes it pos-
sible to combine them in any desired way through the
agency of cross-breeding.

In the first flush of enthusiasm over the rediscovery of
Mendel’s law it was thought by some that recombination
of unit-characters through crossing was to solve all the
problems of breeding relating to the production of new
and improved varieties. But experienced animal breed-
ers have, as a rule, been very conservative in their ex-
pectations, a conservatism justified by the knowledge of
how painfully slow and tedious is the process of improv-
ing a breed in any essential regard. For though it is
easy enough in two generations to get new color varieties
by crossing breeds of different color, the new creations
will, in respects other than color, not be the same as
either of the breeds crossed; they may be inferior to both
in every respect but color, and it will be a difficult, if not
impossible, task to restore the desirable qualities lost.
The reason is that our improved breeds differ from each
other in so many minor characteristics that it is quite
impossible to give attention to all of them simultaneously.
For as the number of variable characters resulting from
a cross increases, a particular combination of characters
will become more rare in occurrence and harder to fix.

Soon after it was discovered that unit-characters exist,

* An address delivered at the University of Illinois, April 19, 1912.

352

No. 546] INCONSTANCY OF UNIT-CHARACTERS 353

the question was raised whether they are or are not con-
stant.

In our descriptions we call these characters A, B, O,
etc., and the recombinations are AB, BC, AC, ete. In
our formule A is always A, and B is always B, but it is
an open question whether in our living animals the char-
acteristics or qualities designated by these symbols are
from generation to generation as constant and change-
less as the symbols. Bateson and Johannsen and Jen-
nings have assumed that they are, that a horn is always
a horn, and a toe a toe. When it is pointed out that
horns are not all alike, that they differ in size, shape and
color, the reply is made that these differences are due to
other things, that is, that these are independent qualities
not inherent in the horn itself. Now there is force in this
argument because we know that a particular color can be
dissociated from the horn, why not also size and shape?
Nevertheless if we dissociate from the horn all color,
size and shape we shall have no horn left. The real unit-
characters, therefore, which we can think of in a con-
crete way and deal with in actual breeding operations are
differences in degree of horn-development, in length,
thickness, curvature or coloration. Who shall say
whether these differences are few or many? We can con-
ceive of an infinite number of gradations in size, shape
and color between known extremes and it is difficult to
believe that any one of these is impossible of realization.
Nevertheless an important body of present-day natural-
ists, those who with De Vries believe in mutation, would
have us think that these minor gradations are not herit-
able. Their reasoning is as follows. Suppose we cross
horned with hornless cattle. All the immediate offspring
are hornless, and the grandchildren 3 hornless to 1
horned. The horned grandchildren breed true. No
intermediates occur. Clearly one unit-character differ-
ence exists between horns and no horns. Therefore no
stable intermediate class can exist unless this unit-char-
acter changes. This they consider to be impossible. If
we call attention to a short-horned race as evidence

354 THE AMERICAN NATURALIST [Vou. XLVI

that the horn character may vary, they assure us that
this condition is due to a different unit and is not deriv-
able from the other, and they challenge us to produce it
from the other. If we begin measuring the horns of our
cattle and picking out those a little shorter than the
average, we find that offspring are obtained with horns
of practically average length. Perhaps we repeat the
selection half-a-dozen times and begin to get a barely
appreciable result. They interrupt, ‘‘See here,” they
say, ‘‘you’re not getting anywhere; give up and acknowl-
edge yourself beaten. If you stop your selecting for a
single generation, the little you have accomplished will
disappear. See meantime what we mutationists have ac-
complished; we have dehorned half-a-dozen breeds by
simple crossing. This is more than you could do in a
thousand years.” Such comment on our work is ex-
tremely disquieting, for our progress is slow, and we can
only reply, ‘‘Your method is the quicker one to get rid
of a character altogether, but you admit yourself power-
less to create a condition which you do not possess fully
realized at the outset. We do not admit ourselves so
helpless; we hope to get something which we do not now
have, and we are willing to wait a while for it. We be-
lieve that we can create what does not now exist. This
you confess yourself powerless to do.”

The foregoing states fairly, I think, the present views
regarding selection as a tool of the breeder held on one
hand by the mutationists and pure-line advocates, and
on the other hand by a minority of Mendelians who like
myself consider selection an important creative agency
in breeding.

The fundamental point of difference between these two
views lies in their different conception of unit-characters.
To the mutationist unit-characters are as changeless as
atoms and as uniform as the capacity of a quart meas-
ure. Theoretically an atom is an atom under all circum-
stances, and a quart holds the same anywhere and every-
where. But the worldly-wise know that the actual quart
is not the same in all places; it is apt to be smaller at the ©

No. 546] INCONSTANCY OF UNIT-CHARACTERS 355

corner grocery than in the U. S. Bureau of Standards,
and the dishonest tradesman will select effectively for
diminished size among the various quart measures
offered on the market, unless his selection is carefully
restrained by legislation. Similarly actual unit char-
acters are modifiable under selection; only one blindly
devoted to a contrary theory will be able long to shut his
eyes to this fact. For several years I have been engaged
in attempts to modify unit-characters of various sorts
by selection and in every case I have met with success.

I shall speak first of the case least open to objection
from the genotype point of view which requires:

1. That no cross breeding shall attend or shortly pre-
cede the selection experiment, lest modifying units may
unconsciously have been introduced, an

2. That only a single unit-character shall be involved
in the experiment.

These requirements are met by a variety of hooded rat
which shows a particular black and white coat-pattern.
This pattern has been found to behave as a simple Men-
delian unit-character alternative to the self condition of
all black or of wild gray rats, by the independent investi-
gations of Doncaster and MacCurdy and myself. The
pigmentation however in the most carefully selected race
fluctuates in extent precisely as it does in Holstein or in
Dutch Belted cattle. Selection has now been made by
Dr. John C. Phillips and myself through 12 successive
generations without a single out-cross. In one series se-
lection has been made for an increase in the extent of the
pigmented areas; in the other series the attempt has been
made to decrease the pigmented areas. The result is that
the average pigmentation in one series has steadily in-
creased, in the other it has steadily decreased. The de-
tails of the experiment can not be here presented, but it
may be pointed out (1) that with each selection the
amount of regression has grown less, i. e., the effects of
selection have become more permanent; (2) that advance
in the upper limit of variation has been attended by a
like recession of the lower limit; the total range of varia-

356 THE AMERICAN NATURALIST [Vou. XLVI

tion has therefore not been materially affected, but a
progressive change has been made in the mode about
which variation takes place.

3. The plus and minus series have from time to time
been crossed with the same wild race. Each behaves as
a simple recessive unit giving a 3:1 ratio among the
grandchildren. But the extracted plus and the extracted
minus individuals are different; the former are the more
extensively pigmented.-

4. The series of animals studied is large enough to
have significance. It includes more than 10,000 individ-
uals.

The conclusion seems to me unavoidable that in this -
case selection has modified steadily and permanently a
character unmistakably behaving as a simple Mendelian
unit.

In my experience every unit-character is subject to
quantitative variation, that is, its expression in the body
varies, and it is clear that these variations have a ger-
minal basis because they are inherited. By selection
plus or minus through a series of generations we can in-
tensify or diminish the expression of a character, that is,
we can modify the character.

In an earlier lecture I showed that long hair and rough
coat in guinea-pigs each differ from the normal condi-
tion by a single unit-character. In 1906 I showed that
both these characters are subject to quantitative varia-
tion, and that such variations are heritable. The same
is true of polydactylism in guinea-pigs, a condition in
which a fourth toe is present on the hind-foot. A poly-
dactylous race of guinea-pigs was unknown until I cre-
- ated one by selection from among the descendants of a
single abnormal individual which had a rudimentary
fourth-toe on one hind-foot. For several generations in
succession those individuals were selected for parents
which had the best-developed extra-toe, and thus was
obtained a good 4-toed race.

Another character built up slowly from small begin-
nings is the silvered variety of guinea-pig. It originated

No. 546] MENDELIAN PROPORTIONS 357

from a tricolor race in which was observed an individual
having white hairs interspersed with red on the lower
side of the body. Selection has been made to increase
the amount of the silvering and to get it on a black back-
ground. This involved increase of the black areas in the
coat as also of the silvered areas. In this task, difficult
because it involved simultaneous modification of two un-
related characters, steady progress has been made. The
best animals are now silvered all over except at the ex-
tremities.

Even albinism, the first-discovered of all Mendelian
characters in animals and by every one acknowledged to

be a single and simple unit-character, even this is vari-
able. In rabbits, for example, some albinos are snow-
white without a trace of pigment in the fur or skin, while
others (the so-called Himalayan type) are heavily pig-
mented at the extremities (nose, ears, feet and tail).
And yet we can not discover that these two kinds of al-
binism differ by any second unit-character which might
account for the difference. Their albinism is different.
Between the extreme types of the snow-white and the
Himalayan albino, various intermediates exist, but all
are clearly albinos, producing only albinos when bred
inter se. They represent quantitative variation within
the albino type.

Similar quantitative variation within colored classes
of animals is well known. Thus in mice an extreme
quantitative reduction of the pigmentation has produced
an animal with pink eyes and faintly colored coat. Such
an animal, however, is not an albino, though less heavily
pigmented than many Himalayan rabbits, for if the pink-
eyed mouse is crossed with an albino, fully pigmented
young result.

In guinea-pigs I several years ago set myself the task
of reducing as much as possible the pigmentation of a
black race, in hopes thus of obtaining blues. I first
crossed the blacks with a light yellow (cream-colored
race). In the heterozygotes the black was somewhat re-
duced in amount. The lightest of these were selected

358 THE AMERICAN NATURALIST [ Vou. XLVI

and again crossed with yellow. By this means the black
was after several generations much reduced. The hairs
were distinctly yellowish at base and the part above sooty
black in appearance. Recently a pink-eyed animal has
appeared in this race with hair light sooty-black in spots.
This evidently is an extreme variation in the direction
which selection has taken throughout the experiment and
probably similar in nature to the pink-eyed variation in
mice. There can be no question of recombination of
independent Mendelian factors in this experiment, be-
cause (aside from albinism) only a single Mendelian fac-
tor is involved. The heterozygotes, as regards black,
have consistently behaved as simple heterozygotes, and
the experience of all observers agrees with my own that
black in guinea-pigs is a simple Mendelian unit. If so it
is clearly a unit modifiable under selection.

In yellow animals, as in blacks, individuals of varying
intensity 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 sep-
arated stages in this series fairly stable in their breeding
capacity and cross these, they Mendelize, 7. e. they behave
as if there were a single unit-character difference be-
tween them. Now this fact is instructive, for it throws
light on the nature of unit-characters in such cases.
They are not things in themselves distinct and separate
from the organism concerned; each is a quantitative
variation plus or minus in some one character possessed
by the organism. Each quantitative condition of a char-
acter tends to persist from cell-generation to cell-genera-
tion. When two quantitatively unlike conditions of a
character are brought together in a fertilized egg, they
- tend to keep their distinctness in subsequent cell-gen-
erations and to segregate into different gametes at re-
production, i. e., they Mendelize. Only by a figure of
speech are we justified in recognizing a unit difference
between them. That difference might equally well be
half as great as it is, or a quarter as great, or a thou-
sandth part as great. _A mono-hybrid ratio would result

No. 546] INCONSTANCY OF UNIT-CHARACTERS 359

equally in each case, upon crossing the two quantitatively
different stages. It is the substantial integrity of a
quantitative variation from cell-generation to cell-gen-
eration that constitutes the basis of Mendelism. All else
is imaginary.

We can distinguish and trace the history of these
quantitative variations from generation to generation
only when the differences between them are of some size.
This has led many to think that only variations of some
size are inherited (the mutation theory) and others to
deny that such variations can be increased in size by se-
lection (the genotype theory). Others still observing un-
mistakable evidence that small variations are heritable
no less than large ones conceive that the large variations
which can be increased or decreased by selection are com-
posed of a certain number of smaller ones cumulative in
their effects (the multiple factor hypothesis). A fatal
objection to this idea is the fact that these quantitative
variations behave as simple units, not as multiple ones,
and so give mono-hybrid ratios, not polyhybrid ones.
The only logical escape from this dilemma for one de-
voted beyond recall to a pure-line hypothesis will be to
assume further that the assumed multiple units are all
coupled, 7. e., all united in a single material body so that
in cell-division they behave as one unit, for practical
purposes are one unit. This position will be logically
unassailable, for we shall never know whether the body
which in practise behaves as one is in the last analysis
composite. Chemists tell us (or used to) that water is
composed ultimately of atoms of hydrogen and atoms of
oxygen not further dissociable, uniform in size and
weight and hard and indestructible as rocks; neverthe-
less for practical purposes we drink our cup of water and
do not chew it. I for one will be content with the ad-
mission that variation is as continuous as water and will
not press the argument against discontinuity into realms
of the ultimate.

The majority of the characters dealt with by the ani-
mal breeder are less simple in behavior than color char-

360 THE AMERICAN NATURALIST [Vou. XLVI

acters. They are also from the economic standpoint
more important. Their careful study is therefore desir-
able. Several years ago I undertook the study of size
inheritance in rabbits. I found that when rabbits of un-
equal size are mated, the young are of intermediate size,
i. e., neither large nor small size dominates in the cross.
Further, segregation does not apparently occur among
the grandchildren, for these vary about the same inter-
mediate mode as the children, though somewhat more ex-
tensively. My conclusion was that the inheritance in
such cases is non-Mendelian, since neither dominance nor
segregation occurs. I called it blending. The experi-
ment with rabbits has been repeated on a much larger
scale by my pupil, Mr. E. C. MacDowell. He finds, how-
ever, that the variability of the grandchildren is consid-
erably greater than that of the children, though it
seldom extends far enough to include the extreme condi-
tions found in the grandparents. This result is’ con-
firmed by observations upon ducks made by Dr. Phillips.
It is evident therefore that size is not a simple unit-char-
acter, for there is no dominance and no evidence of seg-
regation other than the increased variability of the sec-
ond hybrid generation. But cases of this kind have
recently been interpreted as involving multiple unit char-
acters and so as possible Mendelian. This interpretation
has been suggested by interesting observations made by
Nilsson-Ehle on color-inheritance in oats and wheat.

In crossing colored with uncolored varieties he ob-
tained inheritance ratios of 15:1 or 63:1, instead of the
usual 3:1 of colored to uncolored progeny in the second
generation from the cross. The ratios obtained in these
exceptional cases were such as should occur when two or
three independent unit-characters are involved in a cross.
But Nilsson-Ehle could discover only a single kind of color-
production. The conclusion which one naturally draws
from these facts is that the color factor in these cases
was localized in two or three distinct bodies independent
of each other in their splittings and migrations during
cell division. Now Nilsson-Ehle argues with much plaus-

No. 546] INCONSTANCY OF UNIT-CHARACTERS 361

ibility that if in a case such as this dominance were
wanting, i. e., if the cross always produced intermediates,
the bulk of the second-generation offspring would also be
intermediate, with only an occasional complete segrega-
tion. He suggests that size differences may involve units
of this sort, without dominance though fully segregating.
This attractive hypothesis would account for the known
facts of size inheritance fairly well, involving only the
existence of multiple units which may be perfectly stable
and changeless in character. Nevertheless this hypothe-
sis has not been established beyond question. It is quite
possible that we are stretching Mendelism too far in
making it cover such cases. Dominance is clearly absent
and the only fact suggesting segregation is the increased
variability of the second as compared with the first hy-
brid generation. This fact however may be accounted
for on other grounds than the existence of multiple units
of unvarying power.

If size differences are due to quantitative variations
in special materials within the cell, it is not necessary to
suppose that these materials are localized in chunks of
uniform and unvarying size, or that they occur in any
particular number of chunks, yet the genotype hypothe-
Sis involves one or both of these assumptions. Both are
unnecessary. Variability would result whether the
growth-inducing substances were localized or not, pro-
vided only they were not homogeneous in distribution
throughout the cell. Crossing would increase variability
beyond the first generation of offspring because it would
increase the heterogeneity of the zygote in special sub-
stances (though not its total content of such substances)
and this heterogeneity of structure would lead to greater
quantitative variation in such materials among the
gametes arising from the heterozygote. Thus greater
variability would appear in the second hybrid genera-
tion.

As a matter of fact we know that protoplasm is not
homogeneous, and that there are substances widely distri-

362 THE AMERICAN NATURALIST [Vou. XLVI

buted in the cell, not localized in chromosomes, which
may well have an influence on size.

But whatever our conclusion may be concerning the
theoretical explanation of size inheritance, the practical
manipulation of it must clearly be different from that of
color inheritance. All possible combinations of color
factors existing in two distinct races we can secure
within two generations by crossing. New conditions of
color we can often obtain by the slower process of selec-
tion, thus modifying existing color factors. Modification
is, I believe, often accelerated by crossing, quite apart
from the effect it has in bringing about recombination,
because it has a tendency to increase quantitative varia-
tion.

Size is an unstable character, ever varying. Slow
changes in size can be effected by selection without any
crossing whatever. Change in size is made more rapid
by crossing, because variability is increased thereby.
If further increase in size is desired regardless of other
qualities two large races should be crossed and the larg-
est second-generation offspring should be selected. Pro-
gressive diminution in size should be sought in a similar
way, crossing the smallest breeds.

If a medium-sized race is desired, it may be obtained
by crossing a large with a small race and inbreeding the
offspring. Physiological limitations undoubtedly would
prevent unlimited size variations either plus or minus,
yet when we consider what extreme differences exist
among dogs, as for example between ‘‘toy terriers” and —
‘‘oreat Danes,’’ we can scarcely doubt that the limits of
possible size variation have not been approached in most
of our domesticated animals.

SHORTER ARTICLES AND DISCUSSION

INHERITANCE OF COLOR IN THE ALEURONE CELLS
OF MAIZE

In those plants of which there is a considerable knowledge of
the heredity of flower sap color, namely, Antirrhinum, Lathyrus,
Matthiola and Primula, it has been found that an hypostatie
color factor is often necessary for the production of an epistatic
color. For example, a basic factor generally designated as C
being present, a flower becomes red by the addition of a factor
FR, and becomes magenta or purple by the addition of still
another factor P. The zygotic formula of a pure red flower is
RRCC and of a pure purple flower is PPRRCC; ‘but a flower
with the zygotic formula PPCC is colorless.

On the other hand, certain seed coat and other colors of wheat,
of beans and of other plants do not need the presence of the hypo-
static factor for the formation of the epistatic color. For ex-
ample, Nillson-Ehle crossed a black glumed oat BBGG with a
white glumed oat bbgg. In the F, he obtained 12 black: 3 gray:
1 white. The actual ratio was 9BG: 3Bg:3bG@:1bg, but as the
black was produced whether the gray factor was present or not,
the visible ratio was as given above.

The natural conclusion is that in the first category the epis-
tatic factor is more specific in its action than it is in the second
category. If one accepts the interpretation that color is formed
by the action of an enzyme on a colorless chromogen, he must
conclude that the epistatic enzyme of the first kind can only
produce its action, if, by the presence of the hypostatic enzyme,
the chromogen has already been carried through a necessary
preliminary reaction. -An epistatic enzyme of the second kind,
however, is sufficient unto itself and is absolutely independent
of the action of the hypostatic enzyme. One may even assume
that the chemical bases upon which the two enzymes of the
second category act are independent of one another.

Perhaps a concrete illustration will show the difference of
action in these two cases better than description. In the black
glumed oat BBGG, one can imagine the black color or the gray
color wiped out mechanically. Thé other color remains. In the

863

364 THE AMERICAN NATURALIST [Vou XLVI

purple flower PPRRCC, if the red factor is wiped out no color
is left.

In an earlier paper East and Hayes' found four independent
gametic factors in maize, each of which affects the production
of color in the aleurone cells of maize. These four factors are a
basic color factor C, a reddening factor R, a purpling factor P,
and an inhibiting factor J which prevents the development either
of the red or of the purple color. Of the various points of
interest in the interpetation of the inheritance of these factors,
two have been investigated further. The first is the cause of
modified colors. This will be discussed at length in another
paper. The second is the action of the reddening factor R and
the purpling factor P. It was then thought that the presence
of the factor P together with C was all that was necessary for
the production of the purple color. It can now be shown that
the purple color develops only when the three factors P, R and
C are present. The production of color in the aleurone cells
of maize is therefore analogous to that in the flowers of the
genera described above, which was designated as category one.
This interpretation of the facts makes little difference in the
general behavior of these colors in inheritance and is only in-
teresting in so far as it unifies the interpretation of the aleurone
colors in maize with the sap colors of certain flowers.

The following scheme will show the differences in behavior in
the two schemes.

1. A purple crossed with a non-purple gives 3 purple : 1 non-
purple in F,. Here there is no difference in the two schemes.
The proper interpretation gives this result from crosses

PPRRCC X PPERce or
PPRRCC X PPrr€C.

2. A purple crossed with a non-purple gives 9 purple : 7 non-
purple in F,. The old interpretation was that this occurs when
the F, has the formula PpCc. The present interpretation is
that it occurs when the formula of the F, is PPRrCc.

3. A purple crossed with a non-purple gives the formula
PpRrCc in F,. If the R factor is unnecessary for the pro-
duction of purple, the ratio in F, will be (a) 36 purple : 9 red :
19 white. If R is necessary for the production of purple the
ratio in F, will be (b) 27 purple : 9 red : 28 white. A sample

1<<Tnheritance in Maize,’’ Conn. Agr. Exp. Sta. Bull., 167: 1-141, 1911 -

No. 546] SHORTER ARTICLES AND DISCUSSION 365

family of F, segregates gave the following ratio which may be
compared with the closest possible expectancy under each theory.

AGtHAL ORUIE sosro. eek es 191 purple : 56 red : 180 white.
EROF (A)... 240 purple : 60 red : 127 white.
Theory (D) rr ee 180 purple : 60 red : 187 white.

This suggests theory B, but is not conclusive. Conclusive
evidence comes from the F, generation. On theory A, every
36 purple F, seeds should give on the average the following
results in F,:

4 ears all purple. f

10 ears segregating 3 purple : 1 white.

4 ears segregating 9 purple : 7 white.

2 ears segregating 3 purple : 1 red.

4 ears segregating i purple : 3 red : 1 white.
4 ears segregating 9 purple : 3 red : 4 white.
8 ears segregating 36 purple : 9 red : 19 white.

wo

On theory B, every 27 purple F, seeds should give on the
average these results in F,:

1 ear all purple.

4 ears segregating 3 purple : 1 white.

2 ears segregating 3 purple : 1 red.

4 ears se ing 9 purple : 7 white.

8 ears segregating purple : 3 red : 4 white.
8 ears segregating 27 purple : 9 red : 28 white.

is7]
ae
S
w

The crucial test is the appearance of families showing the ratio
12 purple : 3 red : 1 white. No such family has ever appeared.
On the other hand they did divide nicely into families with ratios:
of 9:3:4 and 27:9:28. Of the first type the total progeny
of nine families was 935 purple: 318 red: 436 white. The closest
theoretical ratio on the basis of 9:3:4 would be 950 purple : 317
red : 422 white. Of the second type the total progeny of four
families was 423 purple : 127 red : 396 white. The closest pos-
sible ratio on the basis of 27: 9:28 would be 400 purple : 133 red
: 414 white.

All other tests made corroborated the interpretation that the
P factor can produce purple only when R and C are present.

E. M. East
TORY OF GENETICS,
Bussey INSTITUTION OF HARVARD UNIVERSITY

366 THE AMERICAN NATURALIST [Vou. XLVI
NUCLEAR GROWTH DURING EARLY DEVELOPMENT

In reading Conklin’s recent paper on the relative growth of
nucleus and cytoplasm in developing eggs,! I was at first some-
what puzzled by certain of the relations brought out. The mat-
ter is one that bears directly upon so many important problems,
and Conklin’s paper is one of such fundamental importance,
that possibly a statement of the difficulty and its apparent solu-
tion may be worth while. Work done with the thoroughness
that characterizes all that Conklin puts forth partakes to a cer-
tain degree of the inexhaustibleness of nature, in that it is pos-
sible to discover in it relations not explicitly set forth by the
author.

Conklin’s most notable result is that the relative proportions
of nuclear and cytoplasmic materials do not appreciably change
during early development, as they have been supposed to do.
The bearing of this upon the theory that cleavage is a process of
rejuvenescence, owing to the enormous increase of nuclear ma-
terial relative to the cytoplasm, is evident. The point which I
wish to discuss has no bearing upon this fundamental result, but
relates to the rate of nuclear growth.

On this point Conklin sums up his results for Crepidula as
follows:

“The rate and amount of nuclear growth during cleavage is much
less than is generally believed. Whether the nuclear volume is taken
when the nuclei are at their maximum, mean or minimum size, the
nuclear growth is far from 100 per cent. or a doubling, in each division.
In Crepidula the nuclear growth is not more than 5 per cent. to 9 per
cent. for each division from the 2-cell to the 32-cell stage, and less than
1 per cent. for each division after the 32-cell stage” (p. 40).

Similar figures are given for the other animals studied.

Now, if I have gotten clear on the matter, what Conklin here
means is that when any cell divides, the increase of nuclear ma-
terial thus produced is on the average but 5 to 9 per cent. of the
amount that was present in an early stage of the egg, and spe-
cifically in the two-cell stage. This is a different method of ex-
pressing the rate of growth from that often employed. The
question perhaps most often answered when the rate of nuclear
growth at cell division is given is the following: How much is

1 Conklin pes ‘‘Cell Size and Nuclear Size,’’? Journal of Experimental
Zoology, 12, 1-9

No. 546] SHORTER ARTICLES AND DISCUSSION 367

the nuclear material of a given cell increased when that cell
divides? Or, what is essentially the same, when several cells
are present, as usually in a developing egg: In what proportion
is the total nuclear material increased when all of the cells
divide once? It appears to me that these questions are the
ones that have been in mind when it has been held that the nu-
clear material increases nearly 100 per cent. at each cleavage,
so that the relation of the ratio used by Conklin to the ratio
implied by them is of interest.

The ratio implied in the questions just set forth—the ratio
of increase in the nuclear material of a given cell after that cell
has divided—is of course obtained by dividing the nuclear vol-
ume of the two resulting cells by this volume in the mother cell ;
or by dividing the total nuclear volume of the egg after all its
cells have divided once by the total volume before the cells
divided. (Inclusion of several cells, each of which divides once,
of course does not of itself alter the ratio.) Performing these
operations for the mean nuclear volùmes in Crepidula, as given
in Conklin’s Table 9, one finds the ratios in question to be for
the second cleavage 1.40, for the third 1.25, for the fourth 1.19,
for the fifth 1.89. That is, in passing from the 2-cell to the 4-cell —
stage, the nuclear volume of each mother cell increases 40 per
cent.; in passing from the 4-cell to the 8-cell stage, the increase
is 25 per cent.; from the 8- to the 16-cell stage it is 19 per cent. ;
from the 16- to the 32-cell stage it is 89 per cent. These ratios
are not 100 per cent. at each division, but they approach it more
nearly than the ratio Conklin employs. If we average the in-
crease for these four cleavages, we find the mean to be 43 per
cent. That is, at each cleavage, the nuclear volume of the cell
increases on the average by 43 per cent.

It may be noted that even with an absolutely constant ratio
between the nuclear volume in a given cell before cleavage and
that in its products after cleavage, the ratio employed by Conk-
lin would, as a rule, if I have correctly interpreted it, decrease
rapidly as we pass to later cleavage stages. The relation be-
tween the two ratios would be that shown by the following
formula, in which «== Conklin’s ratio; r= the (constant) ratio
of the nuclear volume after division of a given cell to the nuclear
volume before that war n= the number of the cleavage
(1, 2, 3, ete.) :

368 THE AMERICAN NATURALIST [Vou. XLVI

For example, if r—1.5 (so that the nuclear volume of any
cell increases 50 per cent. when that cell divides), then for the
result of four cleavages (producing 16 cells) the formula gives

3
z= soc Be = se a 16.96 per cent.

If Conklin employed the 1-cell stage as his standard of com-

parison, the above formula would be

t= 4° (2)

It will be found that for any increase less than 100 per cent.
of what was present before division (that is, r— 2), Conklin’s
ratio (from formula 1 or 2) decreases in the later stages of
cleavage, even though the law of increase, so far as each cell by
itself is concerned, remains absolutely the same. Thus, if at
the division of every cell its nuclear volume increases 50 per
cent., Conklin’s ratio (formula 1) will give 25 per cent. for the
result of the second cleavage, 20.83 for the third, 16.96 for the
fourth, 13.54 for the fifth, 10.64 for the sixth, 8.25 per cent. for
the seventh, ete. This appears to-be the reason why Conklin
finds the rate of nuclear increase, as shown by his ratio, to bè
less in later stages; it is not due to any change in the relations
so far as what happens in each cell is concerned.

H. S. JENNINGS

IS THERE ASSOCIATION BETWEEN THE YELLOW
AND AGOUTI FACTORS IN MICE?

IN the generally accepted formulæ for the colors of mice, as
worked out by Cuénot, Bateson, Durham and others, there is
assumed to be a factor, Y, for self yellow color, which is epi-
static to T, the ticking or agouti factor (also known as G). The
various types have the following constitution :

YoUowW 5 pa Sak e ae es eee Sac Yt yt, or Yt yT?
Agouti (including cinnamon) ................. yT yT, or yT yt.
Black and chocolate (including dilute forms) .. yt yt.*

*On the formule adopted here this factor Y is probably to be considered

an inhibitor.

been shown by Cuénot and others (see especially Castle and ~
Little, Science, N. S., 32, 868, 1910) that mice homozygous for Y do not
exist, the reason probably being that the YY zygotes, though formed in the
expected proportions, do not develop.

* Blacks differ from chocolates in having a black factor, B. Agoutis
also carry this factor, while cinnamons lack it. Yellows may or may not
bear it, but the two types are distinguishable by their eye color.

No. 546] SHORTER ARTICLES AND DISCUSSION 369

If yellow mice of the sort one usually obtains be bred together
they produce yellows, blacks and chocolates—almost never any
agoutis. But if such yellows be bred to agoutis and their yellow
offspring be mated together the result is only yellows and
agoutis. It has been pointed out by Morgan‘ that this last result
is inconsistent on the current formule, since blacks or chocolates
should also be expected. For example, if we assume, as I think
we must, that ordinary yellow mice usually have the constitution
Yt yt, then the cross under discussion would be as follows:

Yellow—Yt yt
Agouti—yT yT

Yt yT—yellow
Y —agouti
F, yellow—gametes—YT Yt yT yt

YI ILT IY a gs
x24 38 y7 | yt

Yt re Yt Yt
yr | Yt | yP | yt
ye £2 ch R
YF i Yi yT | yt

yt | yt | yt | yt
YT | Ye | yT | yt

Total, omitting all YY’s: 8 yellows, 3 agoutis, 1 black or chocolate.

It seems to me that the easiest way of explaining why this
mating actually does not produce blacks and chocolates is by the
assumption of linkage or association (‘‘gametic coupling’’)
between the agouti and yellow factors. The fanciers, from whom
most yellows come, ordinarily keep few agouti mice. It is there-
fore probable that most yellows carry no T, and for this reason
Y and T really show ‘‘spurious allelomorphism’’ or repulsion
instead of ‘‘gametic coupling.” There seems to be no evidence
that Y and T ever occur in the same gamete. The evidence
which has led to this conclusion is as follows:

Miss Durham has found that if ordinary yellows be mated

‘Ann. N. Y. Acad. Sci., 21, 87, 1911. It was at Professor Morgan’s
Sas ae that I took up this problem.

šI shall use the term association to cover both coupling and repulsion.
“Journ. Genet., 1, 159, 1911.

370 THE AMERICAN NATURALIST [Vou. XLVI

together they produce two yellows to one black or chocolate,
no T being present. Her total figures are 451 yellows to 241
blacks and chocolates, to which should be added Little’s’ record
of 31 yellows to 24 blacks and chocolates. If ordinary yellows
be mated to chocolates or blacks the result is an equal number
of yellows and of blacks or chocolates. Miss Durham’s figures
are 282 yellows and 319 blacks and chocolates. Both these
crosses are cases of monohybridism, T never being present, and
Y alone being involved.

If ordinary yellows be bred to pure agoutis, the result should
be equal numbers of yellows and agoutis, as shown above. The
actual results obtained are as follows:

Yellow Agouti Authority
53 39 Durham.
16 14 Morgan.
69 53

If the P, agouti used in this cross were heterozygous in T, then
the expectation would be two yellows, one agouti, and one black
(or chocolate). Apparently Morgan is the only one who has
reported such a cross. He obtained 4 yellows, 5 agoutis, and.
1 black.

' The results of the four crosses above are explicable without
the assumption of association between T and Y, since in no case
was an animal bred from which was heterozygous in both, and
only in such cases does association ever become apparent. But
such heterozygous mice should be obtained in the cross between
agouti and yellow. If these be mated together the ordinary
expectation, as shown by the diagram above, would be 8 yel-
lows, 3 agoutis, and 1 black or chocolate. The actual offspring
recorded from yellow by yellow giving agouti are:

Yellow Agouti Black and Chocolate Authority

108 62 Durham.

15 9 0 Morgan.
123 71 0

This is approximately two yellows to one agouti. It is the result
which would be expected if one of the parent yellows were pure
for T (TY Ty). But, with the possible exception of 20 of Miss
Durham’s,’ all the above 194 mice were from yellows out of yel-
* Science, N. S., 33, 896, 1911.
*It would appear from the context that these 20 also belong to the
category under discussion, but a definite statement to that effect is not given.

No. 546] SHORTER ARTICLES AND DISCUSSION 371

low by agouti. No other crosses of yellow by yellow gave agouti,
so that it seems in the highest degree probable that the original
(P,) yellows were pure for t. That being the case the F,
yellows must all have had the formula tY Ty. But since they
produced only 24 yellows to 14 agoutis it follows that association
occurs between T and Y, thus:

Yellow—t¥ Ty
Yellow—tY Ty
Ty tY
tY Ty
(tY tY)
Ty Ty—1 agouti.

—2 yellows.

These 7-bearing yellows have also been bred to chocolates and
blacks. If there were no association this mating should produce
two yellows, one agouti, and one chocolate or black. But Miss
Durham has obtained only 30 yellows and 37 agoutis—practi-
eally equal numbers, with no blacks or chocolates. On the asso-
ciation hypothesis this cross should produce the following result:

Yellow—tY Ty
Chocolate or black—ty ty

tY ty—1 yellow.
Ty ty—1 agouti.

Apparently the ticking factor and the factor which produces
yellow mice are associated very closely. There appears to be no
evidence that ‘‘crossing over’’ ever occurs. It can not be sup-
posed that Y is the same as ¢, with whiich it always occurs, since
in that case all mice not agouti, or even heterozygous for T,
would be yellow, and black and chocolate would not exist. It
should be noted that if one adopts Castle’s mouse formule (see
Castle and Little,” ete.), it is still necessary to suppose that
association occurs but now between the restriction factor, R,
and the agouti factor, A.

A. H. STURTEVANT

COLUMBIA UNIVERSITY,

February, 1912

? Science, N. S., 30, 313, 1909.

372 THE AMERICAN NATURALIST [Vou. XLVI

THE MALTHUSIAN PRINCIPLE AND NATURAL
SELECTION

In the last edition of his essay on population, page 2, Malthus
laid down the following biological proposition from which he de-
rived his well-known sociological conclusion :

The cause to which I allude is the constant tendency in all animated
life to increase beyond the nourishment prepared for it.

It is incontrovertibly true that there is no bound to the prolific plants
and animals, but what is made by their crowding and interfering with
each others’ means of subsistence. (Italics mine.)

In plants and irrational animals, the view of the subject is simple.
They are all impelled by a powerful instinct to the increase of their
species; and this instinct is interrupted by no doubts about providing
for their offspring. Wherever, therefore, there is liberty, the power
of increase is exerted; and the superabundant effects are repressed
afterwards by want of room and nourishment.

The great influence of this book upon Darwin is well known
and so it is not surprising to find him writing in the ‘‘ Origin of
Species,’’ page 60,

A struggle for existence inevitably follows from the high rate at
which all organic beings tend to increase.

And again on page 72,

Each organic being is striving to increase in geometrical ratio; each,
at some period of its life, during some season of the year, during each
generation or at intervals, has to struggle for life and to suffer great
destruction.

It is quite natural, therefore, to find that in the current texts
and in the class room, that natural selection is taught as starting
from the contrast of a limited subsistence and a very large birth-
rate (and I confess that I have been guilty). So strong an im-
pression is thus made, that to most people, and unfortunately
many sociologists, natural selection has come to mean that factor
of evolution which is caused by an excessive birth rate.

The fact is frequently lost sight of that natural selection effects
its results by differential success in mating (sexual selection),
and differential fecundity (fecundal selection), as well as by a
differential age at death (lethal selection). Even when we con-
fine our attention to lethal selection, we shall see that a very
large share of its action is in no way dependent upon the ade-
quacy of the food supply. Such selection may well be called non-

No. 546] SHORTER ARTICLES AND DISCUSSION 373

sustentative selection to distinguish it from that which is so de-
pendent.

Sustentative selection is generally thought to be nearly always
in operation. In every group of animals in which I have made
any special field observations, namely, bryozoa, birds and
beetles, the falsity of this belief has been impressed upon me.
With fresh-water bryozoa the food supply can scarcely ever be
taxed. The limiting conditions seem to be largely inconstancy
of the bodies of water, the danger of being eaten, and the limited
extent of suitable substrata.

With birds, when one really sees an emaciated individual, the
result of some accident which has made it difficult for it to ob-
tain food, one is struck by the very great contrast with other
birds. In my experience, in skinning birds in the state of Wash-
ington, summer and winter, I never opened one not well equipped
with abdominal fat. On the other hand, the great loss of the
young birds by adverse weather and the large variety of pre-
daceous enemies is common knowledge.

With lady beetles there is a more direct relation of the num-
bers to food supply, but even here it is a question of finding food,
rather than any real lack of it. The food supply of the adult
beetles, embracing aphids, pollen and spores, is superabundant.
The principal causes of death seem to me to be due to inability
of the females to distribute the eggs proportionately to the dis-
tribution of aphids, and, secondly, the unreliableness of aphid
stocks, owing to their rapid annihilation when one of the numer-
ous aphid diseases or parasites becomes rampant. One may
often hunt over many colonies of aphids without finding any
coccinellid larve, and then at last find one of the same species
with several large egg masses. So many larve will hatch in this
case that they will consume all the aphids before they themselves
have all become mature. As a result, they will wander, only a
few surviving who may have the efficient instincts and good for-
tune necessary to discover another stock of aphids.

Of the half dozen or so species of the large coccinellids most
common in the United States, it is common to find here one
species abundant, and there another, though some others are also
found. Attempts to account for these contrasts in numbers on
grounds of temperature, humidity, altitude and the like, have
proved unsuccessful. I think it is because the instincts of egg-
laying and of migration of the larve of one species or another is
better adapted to the species of spar ye TENE pane

374 THE AMERICAN NATURALIST [ Vou. XLVI

in that vicinity. Evidence to this effect is seen in the fact that
certain shrubs that are aphis-infected for a short time only in the
early spring sustain only Adalia bipunctata. The wild parsnip,
which becomes infected with other aphids later, sustains Adalia
bipunctata in numbers smaller than three of its competitors. If
we take the herbivorous coccinellids, we will find similarly in
some places an abundance of food that they have not touched,
while some other patches of suitable plants may be stripped.

In conclusion, then, we see that in the bryozoa and birds,
sustentative selection does not play the dominant rôle imputed to

it. In the lady beetles, where the supply of food is seen to limit
their numbers, it is not because there is not food enough, but
because the individuals are not properly distributed with refer-
ence to that food. The sustentative selection in this case must be
differentiated as indirect. The evolutionary significance lies in
the fact that where the sustentative selection is indirect, the
species may become more abundant through variations which ad-
just the individuals better to the food supply.

I believe there is a fundamental reason for this subordination
of sustentative selection. The reproductive rate is not merely
an arbitrarily large number, which necessarily causes a severe
struggle, but is just such a number as is best adapted, in general,
to the needs of the species. The extreme members of that school
which emphasizes the importance of the variations at the expense
of selection can scarcely object to this, for fecundity is always
variable and these differences are known to be inheritable in
many cases.

Now the number of progeny which is best adapted to the needs
of the species is that number which is large enough to sustain
the losses from all non-sustentative causes of death, but not large
enough to invoke death by starvation. Such a species is obvi-
ously less liable to extermination than one in which the hostile |
influence of underfeeding always handicaps. If grasshoppers
conformed to the Malthusian conceptions of many evolutionists,
there should be no alfalfa, for that favorite food would all be
eaten up before it could be harvested. The world teems with
herbivorous animals of one kind or another, and yet also teems
with plants, most of which are eaten by many species of animals.
I can see no other explanation than that the species are not ordi-
narily subject to sustentative selection, and that when it is, it is.
generally the indirect selection rather than the direct. oe

Indirect sustentative selection is less injurious to a species, be-

No. 546] SHORTER ARTICLES AND DISCUSSION 375

cause those individuals which do become well placed thrive. The
burden of starvation passes by a number of the individuals to
fall upon the others. In this way, the evil effects of a general
underfeeding, which is the necessary result of direct sustentative
selection, is avoided. In cases, then, where indirect sustentative
selection is operative, the reproductive rate is that which will
produce enough individuals to find many of the favored places
and withstand the non-sustentative causes of death. Where
direct sustentative selection might be theoretically expected, as
in the case of the large birds of prey, in regions where they are not
persecuted, the rigors of a mers struggle are avoided by low
reproductive rates.

While the reproductive rate must be looked upon as a char-
acteristic which has its adaptations like other characteristics, it
has one peculiarity—its increase is always opposed by lethal se-
lection. The’ chances of life are reduced by reproducing inas-
much as more danger is entailed by the extra activities of court-
ship, and later, of the care of the young, since they reduce the
normal wariness of individual life. The species, therefore, al-
ways keeps the reproductive rate at the lowest point which will
suffice for the reproductive needs of the species. For this reason
alone we should expect the non-sustentative selection to be the
predominant kind.

Gulick and Pearson have shown that there is a normal conflict
between natural selection and fecundal selection. Fecundal se-
lection is said by them to be constantly tending to increase the
reproductive ratio, while lethal selection asserts its power to re-
duce it, because the reproductive demands on the parents reduce
their chances of life by interference with their natural ability of
self-protection. This is quite true, but the analysis is incomplete,
for an increased number of progeny not only decreases the life
chances of the parents, but also of the young, by reducing their
endowment and care.

A further reason for believing in the predominance of non-
sustentative selection is the fact that the species that have
evolved furthest are well known to be of low fecundity.

imself, even where there is no artificial restraint, has one of the
smallest reproductive rates known. If sustentative selection had
been predominant, we should expect higher fecundities in these
highly evolved species than in the lower ones.

The fundamental formula of Malthusianism, that the number
of individuals i in a species tends to increase in geometrical ratio,

376 THE AMERICAN NATURALIST [Vou. XLVI

is misleading, and a great mass of biological and sociological
writing has been led into error by it. Of course, there can be no
objection to the position that since the number of progeny ex-
ceeds the number of parents, there must be many premature
deaths, or the species will increase in numbers. But this is a
truism. The real essence of Malthusianism lies in the conclusion
that a sustentative struggle must arise, and there lies the error.
The Malthusian conception of the rate of reproduction is that the
rate is such that the food supply must be overtaxed and a struggle
for existence will take place. The conception here urged is that
each species has such a reproduction rate as will suffice to with-
stand the premature deaths and sterility of some individuals, and
yet not so large as to press normally upon the limits of the food
supply.

I believe the common over-estimation of the rôle of sustenta-
tive selection, and the neglect of the non-sustentative, is largely
historical in origin, and that it is maintained by improper teach-
ing. : :

In teaching natural selection, the fault is generally made of
starting with the Malthusian contrast between the limitation of
the food supply and the rate of reproduction. The current con-
ception will not be righted until we learn to teach natural selec-
tion more correctly. While the rate of reproduction is the
proper place from which to start, this should be treated, not as a
fixed quantity to which nature must accommodate itself, but as
that number which just exceeds the great number of premature
deaths and suffices to replace the parents. Then the premature
death of the 999 must be explained. After an examination of
these causes the student can not but grasp the master rôle of the
non-sustentative form.

Finally, then, we see from these considerations, that the com-
mon assumption that every species is as common as it can be,
because it is living up to the limits of subsistence is erroneous.
A relaxation of any of the causes of death in most cases will in-
crease the numbers.

[This article was written while the author was on the staff of the Station
for Experimental Evolution at Cold Spring Harbor, N. Y., of the Carnegie
Institution of Washington. ]

RoswELL H. JOHNSON.
BARTLESVILLE, OKLA. .

VOL. XLVI, NO. 547

THE
AMERICAN NATURALIST

VoL. XLVI July, 1912 No. 547

GENETICAL STUDIES ON CENOTHERA. III

FurtHeR Hysrips or (nothera biennis ann O. grandi-
flora THAT RESEMBLE O. Lamarckiana‘

DR. BRADLEY MOORE DAVIS

UNIVERSITY OF PENNSYLVANIA

Tue following paper will describe my cultures of
(Enothera, grown during the season of 1911, which bear
upon the working hypothesis announced in a previous
contribution (Davis, ’11) to the effect that Gnothera
Lamarckiana arose as a hybrid between types of O.
biennis and O. grandiflora. The cultures of 1911 gave
results much more striking than those of 1910 and 1909
and marked progress has been made towards the synthe-
sis of a hybrid between these two species which will be
so similar to Lamarckiana as to be practically indistin-
guishable by the usual taxonomic tests.

The problem before me seems to be chiefly that of find-
ing and selecting among the various strains or races of
biennis and grandiflora the most favorable types with
which to work. As the study has progressed I have been
surprised at the wealth and variety of forms which
taxonomically are included in the species O. biennis, but
which can be readily differentiated in ‘‘pure line” cul-

*An abstract of this paper was presented before the American Society
of Naturalists on December 28, 1911.

377

378 THE AMERICAN NATURALIST [Vou. XLVI

tures as biotypes and which so far have held their char-
acters through two and three generations. (Mxnothera
biennis is much more rich in biotypes than O. grandiflora
probably because it is more hardy and adaptive in its life
habits, growing over an immense geographical range and
under a great variety of climatic and soil conditions. It
is clear that any one who cared to give an extended
period to the study of O. biennis could differentiate
scores of elementary species in the assemblage of forms
included under this name.

Further acquaintance with O. Lamarckiana has led me
to believe that under this name must be included a num-
ber of races. Excluding the more striking of the‘‘mutants’’
of De Vries, there still remain strains that differ from
one another in such characters as the size of buds and
flowers, relative height of the stigma, forms of the cap-
sules, tint of the foliage, ete., and these strains, when
selected and carried on in ‘‘pure lines’’ hold their pecul-
iarities and are true biotypes, although the differences
between them may be so small as to have little or no taxo-
nomic value. We have then under O. Lamarckiana a
diverse assemblage and no one will ever be able to prove
that any one type is identical with the original since the
account of the original and the herbarium material
available do not give the information necessary for a full
description.

The types called Lamarckiana, as now cultivated, have
come down to us perhaps entirely from cultures of
Messrs. Carter and Company of London at about 1860,
but they have come down through diverse channels with
abundant opportunities for hybridization and that differ-
entiation that results from selection, conscious or other-
wise, on the part of gardeners. In the experimental
gardens we are working chiefly with the strains that have
passed through the hands of Professor De Vries, who,
however, began his work (about 1886) many years after
the plant had been in cultivation. This material of the
experimental gardens has then been subjected to a large

No. 547] GENETICAL STUDIES ON ŒNOTHERA 379

amount of selection and every worker who is following
genetical methods is continuing to differentiate, more or
less perfectly, strains or biotypes. I have in my own cul-
tures separated in ‘‘pure lines’ several strains which
differ in the size of the flower, the height of the stigma
relative to the tips of the anthers, and the depth of color-
ation in the foliage, and these strains have held true in
my limited cultures.

These points must be borne in mind in judging the
results of my studies, for it is one thing to attempt the
synthesis of a hybrid taxonomically similar to Lamarcki-
ana, and it would be quite another to attempt to match
exactly a particular biotype of this plant. The prob-
abilities of obtaining a hybrid the exact counterpart of
a specific biotype are small; for the reason that very many
characters or groups of characters give to these strains
their peculiarities. The probabilities of obtaining a
hybrid the characters of which will be matched largely or
wholly by forms of Lamarckiana are, however, in the
writer’s opinion, excellent and such a type is meant when
we speak of a hybrid taxonomically similar to Lamarck-
iana or indistinguishable from it.

e announcement of Gates (’10) that certain marginal
notes in a copy of Bauhin’s ‘‘ Pinax,’’ 1623, give an accu-
rate description of Lamarckiana and establish its pres-
ence in Europe at this early date has proved to be a false
alarm. These notes consist of matter, copied on the
margin of a page, from the longer description of Lysi-
machia lutea corniculata found in the appendix (pp. 520,
521) of the ‘‘Pinax.’’ The readiness with which Gates
and Hill (’11) have accepted as reliable the statement of
Bauhin that the flowers of this form are 3 inches long
above the ovary, in the face of the statement a few lines
below that the capsules become 2-3 inches long (a mani-
fest absurdity), is inexplicable to the writer. It shows a
naive confidence in the descriptions of the early botanists
which is searcely to be expected in the consideration of so
difficult a problem as the origin of O. Lamarckiana.

380 THE AMERICAN NATURALIST [Vow. XLVI

Gates (’lla, p. 101) also believes that the description of
Lysimachia Americana by Hernandez (‘‘Nova Plant.
Anim. et Miner. Mex.,’’ p. 882, 1651) is that of Lamarck-
iana ‘‘in the strict sense.’’ This account is quite as
vague in character as others of the period and the figure
is very crude. Gates emphasizes the statement concern-
ing the leaves ‘‘sinibus levibus excavata’’ and regards
this as descriptive of the characteristic crinkling of the
leaves of Lamarckiana. He gives no consideration to the
petals clearly drawn as mucronate or to the stigma
figured on about the level of the anthers, and is not im-
pressed with the description and figures of the leaves as
like the willow. The writer must express his astonish-
ment that an identification of this plant with Lamarckiana
should be claimed chiefly on a single character loosely
described, ignoring important points that radically dis-
agree, and giving no weight to the evident inaccuracy of
the description and figure. Gates has changed his posi-
tion respecting the account of Bauhin which in this later
paper (Gates, ’1la) is referred to biennis in agreement
with the general opinion of botanists, but I have found
no further reference by him to the description of Her-
nandez. ;

There have been then no new developments to modify
my view that O. Lamarckiana was unknown previous to
the description of Lamarck’s plant at Paris in 1797,
about eighteen years after the introduction of O. grandi-
flora at Kew in 1778. There is less reason to lay stress
upon the appearance of this plant in the gardens at Paris
since the evidence seems clear that the Lamarckiana of
to-day has genetic relation to the cultures of Carter and
Company of London about 1860. I shall have more to
say on this point in connection with the valuable sheet in
the Gray Herbarium and the interesting history of its
probable relation to these same cultures of Carter and
Company. This seems to the writer perhaps the most im-
portant herbarium sheet known bearing on the problem of
the origin of nothera Lamarckiana. It will be con-
sidered in the latter part of this paper.

No. 547] GENETICAL STUDIES ON ŒNOTHERA 381

I have recently had the opportunity of examining two
herbarium sheets of American wild plants, referred to
in my earlier paper (Davis, ’11, p. 227), which were
thought by De Vries (’05, p. 386) to be O. Lamarckiana.
The first of these, at the New York Botanical Garden, is
a specimen collected by A. W. Chapman in Florida (1860
or earlier). Duplicates of this material are said to be at
the Biltmore Herbarium and at the Missouri Botanical
Garden (MacDougal, ’05, p.6). The second sheet, at the
Philadelphia Academy of Natural Sciences with an appar-
ent duplicate at the New York Botanical Garden, is of a
specimen collected by C. W. Short near Lexington, Ky.
This specimen was later considered by Miss Vail to be
O. grandiflora and a possible escape from cultivation.
There is nothing on these sheets from Florida and Ken-
tucky that is not represented in a fair range of herbarium
material of grandiflora such as may now be found in my
own collections and at the New York Botanical Garden.
In no point do the sheets closely approach Lamarckiana
except that they have large flowers. In justice to Pro-
fessor De Vries it should be stated that he expressed his
opinion before the rediscovery of the habitat of Enothera
grandiflora by Tracy in August, 1904, and consequently
before there was available the extensive material of this
species now assembled. Could he have made the com-
parisons at present possible he would not, I am sure,
have given the opinion quoted above.

There has then so far been found in the American
herbaria and records, and these have been very thor-
oughly examined by various workers, no evidence that
(nothera Lamarckiana is at present or ever has been a
component of the American flora as a wild native species.
There are in the south and west certain large-flowered
species of @nothera of considerable interest because of
their possible affinities with grandiflora, and these should
be studied by those in a position to do so. A number of
these are represented in American herbaria; others in
the British Museum are referred to by Gates (’1la, p.

382 THE AMERICAN NATURALIST [Vou. XLVI

591). That an identification with Lamarckiana can ever
be made from the average herbarium sheet seems to the
writer almost impossible, for the specimens of Œnothera
formerly collected rarely give a fourth of the informa-
tion necessary to make a critical comparison with
Lamarckiana. With the clear evidence that the present-
day Lamarckiana holds a genetic relation to the cultures
of Carter and Company, about 1860, the problem of its
origin has become much more tangible than formerly.
and this matter will be taken up in our discussion of these
cultures in relation to the sheet in the Gray Herbarium
at Harvard University.

American botanists are not likely to believe that
Lamarckiana, if present in America in 1860, has become
so quickly extinct, knowing as they do the vitality of our
rich @nothera flora. For example, O. grandiflora has
actually persisted in the same locality since its first dis-
covery by William Bartram in 1776. Let those interested
in the problem of the status of Lamarckiana use their
best endeavors to discover this plant in the field, but let
them give us their results not only by herbarium material
covering the entire life history, but above all through seed
that can be sent to the workers in the experimental
gardens.

The material of this paper will be arranged under the
following headings: (1) Methods, (2) Large- and Small-
flowered Biotypes of @nothera Lamarckiana, (3) Further
Races of Œnothera biennis L., (4) Further Races of
nothera grandiflora Ait., (5) Hybrids in the F, Gener-
ation from the Cultures of 1911, (6) Hybrids in the F,
Generation from the F, Hybrid Plants 10.30La, and
10.30 Lb, (7) The Probable Composition of the Cultures
of Carter and Company from Evidence Furnished by the
Sheet in the Gray Herbarium, (8) Further Considera-
tions on the Possible Origin of nothera Lamarckiana
as a Hybrid of O. biennis and O. grandiflora.

As in previous seasons, I am greatly indebted to the
Bussey Institution and to the Botanic Garden of Harvard

No. 547] GENETICAL STUDIES ON ŒNOTHERA 383

University for the facilities that have made possible
these studies.
1. MzTHODS

The methods of culture were the same as those of the
previous season (Davis, 711, p. 196). I have, however,
adopted the system of collecting and sowing seed capsule
by capsule as being the safest way of regulating the
size of the cultures and obtaining a fair average of
results both qualitative and quantitative. Furthermore
the seeds which go into a seed pan are counted so there
is obtained some data on the percentage of germinations.
The count can not be made strictly accurate, for there are
in @nothera varying proportions of obviously abortive
seeds which so grade into seed of questionable fertility
that good and bad could not be separated unless dis-
sected. Nevertheless, these counts are important, espe-
cially in cases where it is fundamental that all fertile
seed be germinated, as in the comparison of reciprocal
crosses.

2. LARGE- AND SMALL-FLOWERED BIOTYPES OF
(Enothera Lamarckiana

The experience of the writer during the past six years
has forced upon his attention the fact that there is a wide
range in the bud and flower measurements of Gnothera
Lamarckiana in cultures that are practically indistin-
guishable as to their vegetative characters.

De Vries in his ‘‘analytical table of flowers, fruits and -
seeds’’ (‘‘The Mutation Theory,’’ Vol. I, p. 452, 1909)
gives the measurements of the petals of Lamarckiana, on
the average, as 3-4 em. long. I am growing strains or
biotypes of Lamarckiana derived from seeds of De Vries
in which the petals measure from 4-4.5 em. in length and
similar strains have been sent tome from England, where
essentially the same form is cultivated under the name
biennis var. grandiflora. These very beautiful plants
constitute a sort of élite race and apparently represent

384 THE AMERICAN NATURALIST [Vou. XLVI

the best that the gardener’s art (probably in large part
selection more or less in ‘‘pure lines’’) has been able to
accomplish.

There have, however, twice come to me seed of La-
marckiana through different sources (all originally from
De Vries) that has given numbers of plants with much
smaller flowers, but otherwise presenting essentially the
same characters as the large-flowered types. From
such plants I have had no difficulty in establishing strains
(B, D, Y, and Z) in which the petals measure about
2.5 em. (for figure, see Davis, 711, p. 216). The strains
have been perfectly true through two generations,
although the cultures have been small. The stigmas
in these plants are about on the level of the tips of
the anthers, sometimes a little above, in one strain
(B) somewhat below. In this respect the flowers resem-
ble those of biennis in sharp contrast to some of the large-
flowered Lamarckiana in which the stigma is 6-7 mm.
above the tips of the anthers, even higher than is typical
of grandiflora. The point should be emphasized that
when these small-flowered plants are grown side by side
with the large-flowered forms there is no hint of impor- —
tant differences between the plants until the time of
flowering, when the large and small buds first clearly
define the two types.

Some authors will refuse to admit that the small-
flowered plants are Lamarckiana. They will insist that
the true Lamarckiana is always large-flowered and that
these variants are ‘‘mutants’’ or perhaps aberrant types. _
Yet the fact remains that the large- and small-flowered
types are indistinguishable in taxonomic practise except
for the bud and flower characters, and the writer can
but believe that the large-flowered forms have been
steadily selected by those who have for so many years
carried Lamarckiana along to its present state.

Whether the small-flowered forms illustrate reversion
towards a biennis type of flower is a matter worthy of
critical attention. The behavior of my hybrids between

No. 547] GENETICAL STUDIES ON ŒNOTHERA 385

biennis and grandiflora so far tested in the F, genera-
tion (briefly described later in the paper) showed clearly
that there is a segregation of flower size. Some types of
flowers appeared in the F, that were even larger than the
grandiflora parent species and as large as the largest
Lamarckiana; others were smaller than the flowers of the
F, hybrid parent, although none were so small as the
biennis parent species. Between the large and the small-
flowered F, hybrids was an apparently perfect range of
intermediates.

Gates (’11b) has criticized the comparison of the
flowers of my hybrid plants 10.30La and 10.30Lb to those
of Lamarckiana (Davis, ’11) on the ground that their
measurements were too small (petals 2.2 em. long), being
unwilling to recognize the existence of small-flowered
types of Lamarckiana. He apparently, however, fails to
appreciate that the problem is only in part what may
appear in the F, generation. It is the behavior of the
hybrids in the second and later generations that will
demonstrate the possibilities of the double organization
of the F, hybrid, and the results of my cultures of the F,
generation indicate that forms with flowers fully as large
as those of the largest-flowered Lamarckiana may be
readily obtained in abundance. It may prove more diffi-
cult to establish in the hybrids certain points of stem
coloration and leaf form, but my later studies indicate
that these results will depend chiefly upon proper
discrimination in the choice of parents for the cross,
especially among the large variety of biotypes included
in biennis. 3

3. Furruer Races or @nothera biennis L.

I have discarded for experimental purposes the races
biennis A and biennis B which were used in the first
crosses with O. grandiflora (Davis, 11). These have
been supplanted by biotypes much more favorable for
the purposes of the investigation. Among the American
wild forms of biennis the best so far obtained is biennis D

386 THE AMERICAN NATURALIST [Vow XLVI

of my cultures—a fairly large-flowered type (petals
about 2 em. long) with relatively broad leaves and a
green stem the papillate glands? of which are colored red
by anthocyan.

This form, biennis D, is widespread. It is common in
the suburbs of Boston, and I have seen it at Woods Hole,
Plymouth, and in the neighborhood of Philadelphia.
There is considerable variation in the breadth of the
leaves. Similar plants may also be found with clear
green stems indicating that the red coloration of the
glands may not always be a firmly established character
in the strains that show it.

The plant which was the starting point of the strain
biennis D grew wild in the grounds of the Bussey Insti-
tution in company with a number of similar types. From
self-pollinated seed a culture of 51 plants was brought to
maturity in the summer of 1911, all the plants being alike,
even to the red coloration of the glands on the stems.

7 Apparently there has been but little study of the surface tissues of

he tions.

ever, former statements of the absence of external glands 7 the group

younger portions of the plants more or less of a somewhat sticky moisture,
and the problem is from what cells do these secretions come. The hairs
on these plants are of two types, both unicellular, (1) short hairs attached
directly to the surface, (2) much longer and stouter hairs each arising
from the top of a papilla. The papilla in section is seen to consist of a
ages of the epidermis into which extends a number of hypodermal
cells. These hypodermal cells in younger portions of the plant are filled
eek a dense viscous-like subst ance, as are also some of the epidermal cells.
old portions of the plants these cells, like those of the hypodermal tissue
in general, are found to be quiteempty. Thus the appearance of the contents
of the cells composing the papilla indicates that it is secretory in function
and I have consequently termed it a gland. The structure is important in
experimental studies since its coloration in some forms follows that of the
stem on which it lies (green or reddish) while in other pay the papilla
may be colored red upon green stems and ovaries. We trust that the evi-
dence presented above will justify the term gland ati if correct,
preferable to the designations papille, pustules, tubercles, red tubercle- ie ke
bases, red prickles, papillose based, red tuberculate, ete., that have been
applied by my correspondents to this structure, or to ‘the hair.

No. 547] GENETICAL STUDIES ON GENOTHERA 387

The original plant was crossed in 1910 with the strains
grandiflora B and D (for descriptions see Davis, ’11,
pp. 205-207) and the interesting F, hybrids described in
this paper were from this cross.

The chief characteristics of the strain biennis D, when
under good cultivation, are as follows:

1. Rosettes-——Mature rosette (Fig. 1) about 4 dm.

Fic. 1. Mature rosette of @nothera biennis, D (11.13a).

broad. Leaves broadly elliptical, 2-2.5 dm. long, some-
what crinkled, margin sinuate, irregularly toothed and
cut below, green with occasional red spots.

9, Mature Plants.—The mature plants ( Fig. 2), 1-1.5
m. high, have long irregularly spreading side branches
from the rosette and main stem; collar? at the base of the
branches inconspicuous. Stems green above, punctate

collar is suggested for the swollen ring at the base of the
larger branches, which in some species of Œnothera (e. g., grandiflora) is
very conspicuous.

388 THE AMERICAN NATURALIST [Vou. XLVI

Fig. 2. Mature plant of @nothera biennis, D (11.13a).

with red papillate glands at the base of long hairs. Basal
leaves on the main stem elliptical (Fig. 3), about 17 cm.
long, without ‚marked crinkles, irregularly toothed ;
leaves above lanceolate.

3. Inflorescence.—Bracts lanceolate, 4—4 length of buds
(Fig. 3), frequently deciduous, leaving the fruiting
branches destitute of leaves.

4. Buds.—About 5 em. long, the cone 4-angled. Sepals
green, pubescent with numerous long. hairs arising from
papillate glands among which are short hairs; sepal tips
not markedly attenuate.

5. Flowers.—Fairly large (Fig. 3). Petals about 2

No. 547] GENETICAL STUDIES ON ŒNOTHERA 389

em. long. Stigma lobes slightly below tips of anthers,
3 mm. long, pollinated in the bud. Papillate glands on
ovaries red.

6. Capsules —Gradually narrowing from the base,
2-2.5 em. long.

7. Seeds.—Light brown.

Fic. 3. Flowering side branch of @nothera biennis, D (11.13a), with a leaf
from the lower portion of the main stem.

A comparison of biennis D with the strains biennis A
and B (Davis, ’11, pp. 198-200) will show that it has
certain important characters, present in Lamarckiana,
which were not exhibited by those types of biennis first-

390 THE AMERICAN NATURALIST [Vou. XLVI

employed in my crosses. The most important of these
characters are (1) the stem coloration, green punctate
with red papillate glands, (2) a broader rosette leaf
somewhat crinkled, and (3) broader and larger foliage
leaves. The flowers, half again larger than those of
biennis A and B, are of a size more favorable to give
large-flowered hybrids approaching Lamarckiana when
grandiflora is employed as the other parent of the cross.

Although the differentiation of biennis D as a biotype
has marked a great advance in the possibilities of my
experimentation, there are undoubtedly other races of
biennis which will prove better for my purposes. Thanks
to correspondents, I have received seed of biennis from
England and the continent that is likely to give types of
great interest, and more favorable strains among the
American forms are likely to come to hand. There are
some beautiful large-flowered English and Dutch forms
of biennis which in their broad and crinkled leaves and in
their habit are very similar to Lamarckiana, but I have
not as yet found among them the stem coloration desired.
However, it is to be expected that broad and crinkled-
leaved types of biennis will be discovered with red-colored
glands upon the stems and ovaries and such forms when
crossed with grandiflora are likely to give the hybrids
most like Lamarckiana. It will take some time to differ-
entiate such strains, but they are certain to exist since
these characters are presented in part by various types of
biennis. Indeed they seem likely to prove not uncommon
judging from the collections that have come to me during
the past year from botanists who have kindly interested
themselves in the problem.

In my previous paper (Davis, ’11, p. 201) mention was
made of a southern Œnothera (strain S) which appeared
in cultures from the wild seed collected by Tracy as
(nothera grandiflora. Further studies have shown that
the plant is annual, its rosette being small and transitory
as in grandiflora. The form has been described and
named nothera Tracyi by H. H. Bartlett (711). I have

No. 547] GENETICAL STUDIES ON ŒNOTHERA 391

crossed this species with grandiflora, but the hybrids were
very far from Lamarckiana chiefly since the hybrids had
narrow leaves and lacked the persistent rosettes asso-
ciated with the usual biennial habit of the latter plant.
The results show clearly that any form crossed with
grandiflora to produce Lamarckiana-like hybrids must be
one with large persistent rosettes such as are presented
by the northern types of biennis.

4. Furrner Races or Gnothera grandiflora Arr.

A type of Gnothera grandiflora appeared in my cul-
tures of 1910 (Davis, ’11, p. 204), first noted because of
its rosette of green, much crinkled leaves. This strain,
grandiflora I, was further cultivated in 1911 but has
proved less favorable for my purposes than the strains
A, B, and D, chiefly for the reason that its leaves have a
pronounced petiole and the stems bear towards their tips
dense clusters of flowering side shoots. The strain
grandiflora I is a well-defined type very different from
the much more common forms of grandiflora and as
such is of interest.

To add to the data on the composition of Œnothera
grandiflora as it grows wild (Davis, ’11, pp. 202-205) a
culture of 169 plants from wild seed, collected by Tracy
at Dixie Landing, Alabama, in 1907, was brought to
maturity during the season of 1911. In this culture 42
plants proved to be @nothera Tracyi referred to above,
9 plants were unmistakably of the strain grandiflora I,
and the remaining 118 plants were close to the strains
grandiflora A, B, and D, which it will be remembered are
so similar as to be essentially of one type.

It seems clear from my cultures of wild seeds, a total
during the past four years of about 300 plants, that the
prevailing form of grandiflora is that represented by the
strains A, B, and D, previously described (Davis, ’11,
pp. 205-207). There is a range of variation, chiefly in the
breadth of the leaves, and these strains (A, B, and D) are

392 THE AMERICAN NATURALIST [Vou. XLVI

the broader-leaved, more luxuriant forms such as the
gardener would be likely to select for cultivation.

5. HYBRIDS IN THE F, GENERATION FROM THE
CULTURES oF 1911

The most important cultures of 1911 in the F, genera-
tion were those of the following combinations of parent
species.

1. grandiflora B X biennis D (11.35).

2. grandiflora D X biennis D (11.32).

3. grandiflora I X biennis D (11.37).

4. grandiflora D X Tracyi (11.33).

Of these the most interesting, with respect to the
resemblance of some of its plants to Lamarckiana, was
the first culture in the list—grandiflora BXbiennis D
(11.35). The greater part of this account will conse-
quently be devoted to this culture, but the others will be

Fig. 4. E rosette of an F, hybrid grandiflora B x biennis D (11. 35),
compared with that of Lamarckiana Z (11.7).

briefly described, chiefly with reference to the coloration
of the papillate glands in which all four cultures agree in
exhibiting a behavior that was not tò be expected.

l. grandiflora B X biennis D (11.35).—This culture

No. 547] GENETICAL STUDIES ON ŒNOTHERA 393

was derived from the contents of two capsules containing
about 300 seeds. From these 247 seedlings appeared in
the pans within six weeks ; 198 young rosettes were potted
and finally 180 large rosettes were set out. Thus, in re-
ducing the culture, 67 smaller rosettes which gave less
promise of developing into vigorous plants were dis-
carded.

The similarity of the young rosettes of the hybrid to
those of Lamarckiana is shown in Fig. 4. The mature

Fic. 5. Mature rosette of the F, hybrid 11.35a, preven B x biennis D;
representative of the mass of the ¢

rosettes (Fig. 5) presented characters that were clearly
blends in various degrees between the parent types,
blends very hard to define because of a range of varia-
tion in the leaves. The leaves were conspicuously
crinkled as in Lamarckiana, differing from the latter
types of my cultures chiefly in being more deeply toothed
and cut at their bases and in being colored a slightly
darker shade of green with more’numerous reddish spots

394 THE AMERICAN NATURALIST [Vou. XLVI

of anthocyan. The rosettes resembled those of Lamarck-
iana very closely in their morphology, much more closely
than those of my earlier hybrids (see Davis, ’11, Figs. 9
and 12), as would be expected from a cross involving the
broader-leaved biotype biennis D.

Continued experience with the rosettes of Gnothera
has shown that there is a greater range of variation
among the F, hybrids of the species crosses than might
be expected by those whose studies have been upon
hybrids of closely related races. It has always been pos-
sible to arrange the mature rosettes of a large culture in
a series with a small group at each end that differ and
incline towards the two parents. It is my method to set
out the rosettes in this order and generally the rosettes
at the ends will develop into mature plants exhibiting a
similar degree of divergence from the mass of the cul-
ture. But this is by no means a fixed rule and rosettes
giving promise of a certain line of development may
grow into mature plants of an unexpected type. I have
emphasized the divergence between the rosettes at the
extremes of an F, generation, but it must not be supposed
that these plants constitute classes. On the contrary, in
large cultures they apparently range insensibly into the
mass.

A comparison of the parents of this cross, grandiflora
B (see Davis, ’11, pp. 205-207) and biennis D described
above, will show that they differ chiefly in the relative
form, size and proportions of their organs, and their
characteristics would be expected to be present in the F,
hybrids in blended relations, as is the fact. The only
character observed that might be sharply contrasted as
present or absent in these parents is the coloration of the
papillate glands which in biennis D are red on green
stems and ovaries, and in grandiflora D are uncolored,
i. e. lack the red. The hybrids of the F, generation
naturally might be expected to exhibit either the presence
or absence of the red coloration in these glands as one oF
the other condition might be dominant.

No. 547] GENETICAL STUDIES ON ŒNOTHERA 395

To my surprise, with the appearance of the main stem
from the rosettes two classes of plants became at once
defined in the culture of these F, hybrids: Class I, repre-
sented by 12 plants with red glands on the ovaries and
the green portions of the stems, and Class II, repre-
sented by 168 plants in which the papillate glands simi-
larly situated lacked the red coloration. There were
apparently no intermediates, with respect to this char-
acter of gland coloration, between these two well-defined
types which usually differed in other respects, as noted

elow.

The mature plants of the two classes could generally
be sharply contrasted with one another in the following
respects.

Crass I (12 PLANTS) Crass II (168 PLANTS)
Papillate glands colored red on the Red coloration absent from the
ovaries and green portions of the papillate glands on the ovaries

Mature SPR (Fig. 6) symmetrical Mature plants (Fig. 8) somewhat
ollar at base of the long straggling in habit, collar at ba:
hai inconspicuous. of s branchés more ĉon-

spicu
Lower leaves (Fig. 7) narrowly el- Lower Tivi (Fig. 9) broadly el-

TEEM, lanceolate on upper por- a ovate on upper portion o:
tions lant.
Bud cones 4-angled. me cones round in section,
Stigma lobes 6-7 mm. long. Stigma lobes 3-4 mm. long.
Capsules about 3.3 em. long. Capsules about 2.3 em. long.
racts persistent, developing on Bracts deciduous, the fruiting
fruiting branches into lanceolate branches becoming nearly or
leaves, wholly destitute of leaves.

We have said above that the two classes of plants
could usually be sharply contrasted with respect to the
characters listed. There were found no exceptions as
regards the coloration of the glands, but with respect to
the other characters some variation was exhibited, espe-
cially among the large number of plants in Class II. For
example, the plant of the culture most resembling
(Enothera Lamarckiana (11.35La, to be described later)
was representative of Class II in all respects except that

396 THE AMERICAN NATURALIST [Vow XLVI

it had 4-angled buds. Class II was more diversified than
Class I, but this was probably because of its being repre-
_ sented in the culture by fourteen times as many plants.
It is of interest to compare the hybrids of these two
classes with their parents. The plants of Class I resemble
the biennis parent in the red glands on ovaries and green
stems, inconspicuous collars, narrower leaves, and 4-
angled buds; they resemble the grandiflora parent in
having longer stigma lobes, longer capsules and persist-
ent bracts. The plants of Class II resemble the grandi-
flora parent in the absence of the red coloration in the
glands on the ovaries and green portions of the stems,
and in having more conspicuous collars, broader leaves,
and round buds; they resemble the biennis parent in
having shorter stigma lobes, shorter capsules, and decid-
uous bracts. It will be noted that the contrasted char-
acters are mixed for both classes of plants, some of them
being biennis-like and some of them grandiflora-like.
Thus neither class could be claimed as patroclinous or
matroclinous except it were established that gland colora-
tion, size of collar, form of leaves and form of buds are
more important specific characters than length of stigma
lobes, length of capsule and the persistent or deciduous
nature of the bracts, or vice versa. Of these diverse
characters who would be willing to express the opinion
that one set or the other is not of consequence in a de-
scription of the forms? Yet it would be necessary to
disregard one or the other set of characters if either class
of hybrids were defined as patroclinous or matroclinous.
We have stated before that the plants of the culture
were set in the ground with the few more biennis-like
rosettes at one end of the bed, and the few more grandi-
flora-like at the other end, and that these extreme forms
graded insensibly into the mass of the culture. It is
important tọ note that not one of the 12 plants of Class I
was at either end of the series, but they were scattered
irregularly through the culture, in some cases 2 or 3 close
together, but usually wide apart. I know as yet no way in

No. 547] GENETICAL STUDIES ON G@NOTHERA 397

which the representatives of Class I can be distinguished
in the rosette condition; they certainly do not constitute
either of the small groups of rosettes that are more
like one or the other of the two parents.

The differentiation of these two classes of hybrids in
an F, generation appears to be similar to the phenomenon
of ‘‘twin hybrids’’ reported by De Vries (’07) although
the contrasted characters are somewhat different. The
behavior is most interesting, especially in its relation to
the results expected in F, generations according to gen-

Fic. 6. Mature plant gea i sah hybrid 11. “sang — B x biennis D;
tative of Class

eral experience and Mendelian laws. We should have
expected the red coloration of the papillate glands of the
biennis parent to be either dominant or recessive to the
uncolored glands of the grandiflora parent; as a matter

aoe THE AMERICAN NATURALIST [Vou. XLVI

of ‘fact, both conditions appeared with apparently no
UN RE E It would, however, be unsafe to regard
this behavior as an exception to Mendelian laws since
there is the possibility that the strain biennis D is hetero-
zygous with respect to the red coloration of its papillate
glands. To be sure, a culture of 51 plants from self-polli-
nated seed of the wild plant (biennis D) were uniform
even to the red coloration of the glands on the green
stems and ovaries, but, nevertheless, there are similar
wild biennis types that lack this coloration, and there has

Fic. T. plas tiles side branch of the F, hybrid 11.35m, grandiflora B x biennis D;
esentative of Class I. At the left is a leaf from the
lower portion of the main stem

No. 547] GENETICAL STUDIES ON CENOTHERA 399

not yet been time to determine whether or not the strain
biennis D is homozygous in all of its characters. The
matter has no especial bearing on the immediate purposes
of my studies, but will become vital to further investiga-

Fig 8. Mature plant me a F, hybrid 11. =f Agger B x biennis D;
presentative of Clas

tions on the behavior of the characters of gland colora-
tion in the Œnotheras.

My plans for the further study of these hybrids in
second generations involve cultures in pure lines from
representative plants of both Class I (11.35m, Figs. 6
and 7) and Class II (11.35a, Figs. 8 and 9), and cultures

400 THE AMERICAN NATURALIST [Vou. XLVI

of reciprocal crosses between these same representatives
(11.35mXa and 11.35axm). The results of the reci-
procal crosses will be awaited with especial interest in
view of De Vries’s (711) recent paper on double reciprocal
crosses. Furthermore, a large second generation will be
grown from self-pollinated seeds of a plant (11.35Za) in
Class II selected as being in more respects similar to

Fig. 9. s side branch of the F; hybrid 11.35a, grandiflora B x biennis D;
representative ~ Class II. At the left is a leaf from the
lower portion of the main stem

Lamarckiana than any hybrid of my crosses so far grown.
Let us consider the hybrids just described with respect
to their resemblance to Lamarckiana. As regards the

No. 547] GENETICAL STUDIES ON @NOTHERA 401

size of the flowers both classes of hybrids presented
essentially the same types (petals 2.5-2.7 em. long),
closely intermediate between those of the parents, but
the flowers of Class I were similar to those of Lamarck-
iana in the length of the stigma lobes, in their 4-angled
buds and in having red glands on their ovaries. In the
form and size of the capsules and in the deciduous habit
of the bracts Class II had the advantage of a closer
resemblance to Lamarckiana. In the broader and some-
what crinkled leaves the advantage was very much in
favor of the plants in Class II, but in the coloration of
the glands on the stem the plants of Class I were
Lamarckiana-like. The situation seems mixed and we
must await the outcome of the cultures as planned above
in the hope of some interesting conclusions. The writer
in selecting a representative of Class II (11.35Za) as
more like Lamarckiana than any other plant has laid the
greater emphasis upon the characters of foliage and
buds (in this plant stout and 4-angled), believing that
these characters of Lamarckiana are more important and
more difficult to obtain in hybrids than many of the
others.

A description of this plant (11.35Za), the F, hybrid, so
far obtained, most closely resembling O. Lamarckiana,
will now follow, arranged to bring out its important char-
acteristics in comparison with those of the parent species
and with Lamarckiana.

Hysrip 11.35La

1. Rosette——The rosette of this plant was unfortu-
nately mutilated by cut worms which destroyed the cen-
tral bud so that the shoots that grew to bear flowers were
all from side buds. The rosette, however, was similar
to that shown in Fig. 5 (11.35a), that is to say, it was

representative of the mass of the culture. My attention
` was first attracted to this plant by the greater length and
breadth of the leaves upon the young side shoots and their
conspicuous crinkling, well illustrated i in Fig. 10. As tiie

402 THE AMERICAN NATURALIST [Vou. XLVI

side shoots elongated (Fig. 11) the foliage retained the
promise of the earlier condition (Fig. 10), the shoots
bearing leaves distinctly larger and more crinkled than
the other plants of Class II, but of the same form.

se

Fic. 10. Rosette of the F, hybrid 11. St grandifiora B x biennis D.
R oe destroyed by cut worms; side shoots developing
h large, cones P a leaves

2. Mature Plant.—Five strong side shoots, 1.1-1.3 m.
high, (Fig. 12) developed from the mutilated rosette.
Their stems. were green above, reddish below, the
papillate glands following the coloration of the stem,
i. e., they were green (uncolored) on the ovaries and
younger portions of the stem as in Class I, and not red
as in my cultures of Lamarckiana. The basal leaves,
about 23 em. long, were broadly elliptical and strongly
crinkled, the margins scarcely toothed except at the base;

No. 547] GENETICAL STUDIES ON CENOTHERA 403

the leaves above were ovate or broadly elliptical; all
leaves had short petioles. The leaves on the plant
throughout the history recorded by Figs. 10, 11 and 12
were so similar in form, size, and texture to those of my
cultures of Lamarckiana that I do not believe the plant
could have been easily separated by its foliage if grown
among them. The absence of red colored glands on the
green stem and ovaries appeared to the writer to be the
only character of importance distinguishing it from

Fic. 11. Young plant of the F, hybrid 11.35La, grandiflora B x biennis D,
showing Lamarckiana-like foliage of large, crinkled leaves.

Lamarckiana during its development up to the time of
flowering.

3. Inflorescence. —The inflorescence (Fig. 13) pre-
sented. shorter internodes than is characteristic of
Lamarckiana as I have observed it, and was in conse-
quence more flattened at the top, in this respect resem-
bling the inflorescence of gigas. The bracts were similar

404 _ {HE AMERICAN NATURALIST [Vou. XLVI

in form to those of Lamarckiana, but were not quite so
closely sessile.

4. Buds.—The buds, about 7 em. long, were strongly
4-angled (in this respect departing from the rule in Class

Fic. 12. Mature e “ Agra F, hybrid 11.35La, gřandiflora B x biennis D,
h Lamarckiana-like foliage.

II). Their pubescence was like that of Lamarckiana,
long hairs arising from papillate glands among muc

shorter hairs which were, however, less numerous than
in Lamarckiana. The sepals were green, their tips of
medium length, not markedly attenuate as in grandiflora;

No. 547] GENETICAL STUDIES ON CENOTHERA 405

the cone of the bud was stout. The strong resemblance
of the buds in form to those of Lamarckiana was an im-
portant point in the choice of this plant as favorable for
further cultivation. They were, however, from 2-3 em.
shorter than those of the large-flowered Lamarckiana,

Fig. 13. Flowering side branch of the F, hybrid 11.35La, grandiflora B x
biennis D. At the left is a leaf from the lower third of a side stem.

but as large as the small-flowered type. There should be
little difficulty in obtaining larger sizes in the P.
generation.

5. Flowers.—The flowers were medium-sized, petals
about 2.5 cm. long. The stigma lobes, 3—4 mm. long, were
slightly below the tips of the anthers. Ovaries were

406 THE AMERICAN NATURALIST [Vou. XLVI

green (without red-colored glands). The flowers were
indistinguishable from those of the small-flowered types
of Lamarckiana except for the absence on the ovaries of
red-colored glands. They differed from the large-flow-
ered types (petals 4-4.5 em. long) not only in size, but
also in the shorter stigma lobes which were not above the
tips of the anthers. My experience has shown that the
flowers in a second generation of these @nothera hybrids
will present a wide range in their measurements, many
flowers surpassing in size those of the grandiflora parent
and equalling the largest forms of Lamarckiana. Conse-
quently there is good reason to be hopeful of obtaining
large-flowered types in the F, generation.

6. Capsules—The capsules, about 2 cm. long, were
stout and similar to those of Lamarckiana. The bracts
exhibited the deciduous tendencies of Class II, a
Lamarckiana-like habit.

7. Seeds.—The seeds were of a shade of color inter-
mediate between the light and dark brown of the parents.

Speculation as to the probabilities of the behavior of
these F, hybrids (11.35La, 11.35a, and 11.35m), and the
reciprocal crosses (11.35 a X m, and 11.35 m x a) in the
second generation, is not called for, since we hope in afew
months to have data on the questions involved. Some of
these problems concern the general behavior of species
crosses and are to the writer as interesting and im-
portant as the attempt to synthesize a Lamarckiana-like
hybrid between biennis and grandiflora. It may, how-
ever, be pointed out that there are no important char-
acters of Lamarckiana which are not represented in some
one of these hybrids or in their parents, except those of
the somewhat larger size of certain organs. Since my
cultures have shown that hybrids of @nothera in the F,
generation frequently present marked advances over
their parents in the measurements of petals, leaves, etc.,
these differences of size in an F, hybrid are likely to
prove more apparent than real.

2, grandiflora DX biennis D (11.32)—As would be

No. 547] GENETICAL STUDIES ON ŒNOTHERA 407

expected, this cross was very similar to that just described
(11.35, grandiflora B X biennis D). The 195 plants
brought to maturity fell into two classes distinguished by
precisely the same groups of characters as in the previ-
ous cross. Class I with red papillate glands, etc., was
represented by 11 plants; Class II by 184 plants. Cer-
tain small differences of foliage, apparent only on close
inspection, together with the fact that no plant seemed
as favorable for my purposes as 11.35La, determined the
selection of the other culture (11.35) as the one upon
which to base further studies.

3. grandiflora I X biennis D (11.37). The rosettes of
this cross (Fig. 14) were more like those of Lamarckiana

Fic. 14. Mature rosette of an F, hybrid, grandiflora I x biennis D (11.37).
A type of rosette very similar to that of Lamarckiana.

than in any of my hybrids so far obtained, but the mature
plants were disappointing. The strain grandiflora I has
a strongly petioled leaf, and the stems bear dense clusters
of flowering side shoots. These characters appeared in

408 THE AMERICAN NATURALIST [VoL. XLVI

pronounced form in the F, hybrids rendering them less
favorable for my studies than the crosses with the strains
grandiflora B or D. The culture was interesting since
it presented two classes of F, hybrids as in cultures 11.35
and 11.32, separated by identical groups of characters.
Out of 144 plants brought to maturity 4 plants were of
Class I and 140 plants were of Class II.

A culture of the reciprocal cross of these types, biennis

D X grandiflora I (11.39), was started late in the season
after the appearance of the paper of De Vries (11) on

double reciprocal crosses, but the culture was unable to
reach maturity. By September 15, out of a total of 171
rosettes, 31 had sent up side shoots of sufficient length
(2-4 dm.) to determine the coloration of the papillate
glands; 21 of these plants had red glands on a green
stem, and in 10 plants the glands were uncolored.

4. grandiflora D X Tracyi (11.33)—The F, hybrids
of this cross were remarkable plants 2.5-3 m. high,
developing from transitory rosettes and with a luxuriant
foliage of narrow leaves. As noted before, the plants
were very far from Lamarckiana. It is a cross that is
likely to occur in nature where the two species grow
together as at Dixie Landing, Alabama, and this possi-
bility must be reckoned with in analyses of the two
species. @Œnothera Tracyi has the red coloration of
the papillate glands characteristic of biennis D. The
40 plants of the culture fell sharply into two classes dis-
tinguished by characters similar to those of the crosses
between biennis D and grandiflora. There were 3 plants
of Class I, with red glands, etc., and 37 of Class II.

The cultures of the F, hybrids as recorded above have
given data that can not be satisfactorily discussed until
similar crosses have been made with parents that are
beyond question homozygous. My strains grandiflora B

and D have now been carried in pure line for three gen- —

erations and have proved uniform, but the parent plants
biennis D and Tracyi were from wild seed. It does not
seem probable that the latter types are heterozygous, for

t

No. 547] GENETICAL STUDIES ON ŒNOTHERA 409

the reason that both have the habit of close pollination,
but, nevertheless, the forms are not above suspicion. As
the data stand the most interesting features shown by
the F, hybrids are: (1) The absence of dominance on
the part of any character of the parents, including the
coloration of the papillate glands, (2) the sharp differ-
entiation of two classes of hybrids (twin hybrids?)
clearly distinguished by groups of characters, and (3)
the absence of marked patroclinous or matroclinous con-
ditions in either of these two classes. The differentia-
tion of two classes of hybrids has been a new phenomenon
in my experience with F, hybrids of Œnothera which
have previously proved remarkably uniform except for
the extremes represented by a few plants somewhat ap-
proaching the respective parents.

6. HYBRIDS IN THE F, GENERATION FROM THE F, HYBRID
Pruants 10.30La andn 10.30Lb.

In my paper of a year ago (Davis, ’11, pp. 211-217)
two plants were described and figured which among the
hybrids up to that time most closely resembled nothera
Lamarckiana. They were the F, hybrid plants designated
10.30Za and 10.30Lb, the result of the cress grandiflora
B X biennis A. Cultures from self-pollinated seed of
these two plants were grown in the season of 1911, giving
considerable data on the behavior of such a cross in the
second generation. :

These F, hybrids were not grown with the expectation
that any of the plants would be taxonomically the same
as Lamarckiana, since it had been apparent for some
months that the strain biennis A is not so favorable a
biotype as many others to cross with grandiflora in the
hope of obtaining Lamarckiana-like hybrids. The
hybrids were grown to satisfy my keen desire to observe
the behavior of such a cross and in the hope that its study
would prove helpful in the planning of future work.

The mass of the F, hybrids held in their characters

410 THE AMERICAN NATURALIST [Vou. XLVI

within a certain wide range of variation, but a large
number of very striking types appeared, which, if they
come reasonably true in a third generation, must be
regarded as new forms of specific rank. These variants
constituted in some cases classes entirely distinct from
all of the other plants and in other cases were extreme
types of segregates connected by intermediate forms more
or less closely with the mass of the hybrids. Although the |
two F, parent plants of these cultures, (10.30Za and
10.30Lb) were sisters, their progeny was far from similar.
Each F, hybrid plant gave rise to its own peculiar set
of variants and the character of the plants constitut-
ing the mass of the cultures likewise differed greatly.
The present description of the cultures will be very brief,
but a more full and illustrated account may follow if
studies of the more interesting of the variants in a third
generation make the publication of such an account seem
desirable,

1. THe F, Generation (11.41) FROM THE HYBRID PLANT
10.30La

This culture was grown from the contents of 26 cap-
sules containing in all about 3,300 seeds. From these
1,505 seedlings appeared in the pans within 5-7 weeks,
and 1,451 grew into well-developed rosettes. It was
possible to select from the rosettes a group of 141
which were smaller than the average and had strongly
etiolated leaves; these rosettes developed a peculiar
group of dwarfed plants apparently constituting a clearly
defined class in the culture. The mass of the rosettes
presented a remarkable range, from many that resembled
very closely the rosettes of grandiflora, distinguished by
having broader leaves cut at the base (see Davis, ’11,
Fig. 6), to much smaller rosettes with narrow leaves
somewhat biennis-like in form. Between these extremes
was an assemblage of intermediates so various in char-
acter that I was unable to make any satisfactory classi-
fication further than the selection of the 141 etiolated

No. 547] GENETICAL STUDIES ON ŒNOTHERA 411

types. It may be said, however, that the green rosettes
inclined more towards the female parent of the cross,
grandiflora B, than towards the male, biennis A. The
plants set in the ground numbered 1,310 green rosettes
and 141 etiolated.

The green rosettes were set out with the more grandi-
flora-like at one end and the more biennis-like at the
other. In general the plants at the more grandiflora-
like end were at maturity considerably larger 1.5-2 m.
high, than those towards the more biennis-like end where
a small group of plants (about 20) remained small, 6
dm.—1.2 m. high. Among the larger plants were many
(about 50) with flowers as large as or larger than those
of the grandiflora parent, but these could not be sepa-
rated as a class since the variation on individual plants
was considerable and they intergraded with great per-
fection into the mass of the culture. The smaller plants
at the more biennis-like end were variable, but all had
flowers 2-4 times larger than those of the biennis parent.
The larger flowers had petals measuring 44.5 em. long
(those of grandiflora B are about 3.3 em. long); these
flowers were therefore as large as the largest-flowered
types of Lamarckiana. The large flowers also, as a-rule,
presented the grandiflora relation of the stigma to the
anthers, i. e., the stigma lobes were well above the tips of
the anthers. Thus in flower size the culture showed a
clear segregation, but inclined markedly towards the
larger size of the grandiflora parent, in many cases sur-
passing this flower in its measurements.

The foliage of the culture presented a range of varia-
tion that defied my attempts at classification. That of
some of the larger-flowered plants closely resembled
grandiflora, and a few of the smaller plants had a foliage
approaching that of biennis, but the leaves throughout
the culture as a whole were larger than those of the
parents of the cross and generally distinctly crinkled.

The mass of the culture was fairly close to the F,
hybrid plant, 10.30La, from which it came, but there was

412 THE AMERICAN NATURALIST [Vot XLVI

much variation from this plant in the measurements
and proportions of the organs with the greater tendency
towards the grandiflora parent type. There was, how-
ever, no constancy apparent and plants with flowers
similar to grandiflora might present a foliage of a dif-
ferent type, most frequently in the form of larger and
strongly crinkled leaves. A few plants were so close to
grandiflora that taxonomically they would probably be
included in the range of this species, although they could
not be considered identical with the strain grandiflora B;
there were no segregates so close to biennis A.

A count was made with the assistance of Mr. H. H.
Bartlett of the plants with green stems (biennis-like) and
of those with red stems (grandiflora-like). This was ex-
tremely difficult for the reason that the mass of the cul-
ture presented intermediate or mottled conditions similar
to the F, hybrid plant 10.30La. ‘However, of the 1,310
plants, 195 were classified as red-stemmed and 192 as
green-stemmed; the red were markedly nearer the more
grandiflora-like end of the culture and the green nearer
the biennis end. The expected number according to
simplest Mendelian ratio should have been about 327
plants of each color. Considering that anthocyan is so
variable in its appearance, and consequently most un-
satisfactory for a study of color inheritance, these results
are by no means against Mendelian expectations. No
two persons independently would make the same count in
such a culture, for the results all depend upon what con-
ditions of the plant are defined as intermediate’ or
mottled. |

The culture presented a number of remarkable types
which no taxonomist would think of identifying with
either biennis, grandiflora, or the F, hybrid plant 10.30La.
These will not at. present be described in their many
characteristics, but will be briefly listed.

11.41¢, a type with conspicuously crinkled leaves, repre-
sented by about 170 plants, apparently intergrading with
other forms of the culture.

No. 547] GENETICAL STUDIES ON C2NOTHERA 413

11.41bk, represented by 13 plants, narrow leaves,
flowers very light yellow, anthers sterile.

11.41s, a single plant similar to 11.41bk except for
broader leaves.

11.417, representing the class separated as 141 etio-
lated rosettes. The plants of this class as they developed
gradually outgrew the etiolated conditions of the younger
stages, but remained dwarf. They varied greatly among
themselves in flower size and foliage, and a dozen differ-
ent types could have been selected.

The above list does not include the numerous types of
segregates which clearly differed specifically from the F,
hybrid plant 10.30La. These forms, being of hybrid
origin, are of course marked segregates presenting in
varying combinations and degrees characteristics of the
parents of the cross, bignnis and grandiflora.

2. Tue F, Generation (11.42) FROM THE HYBRID PLANT
10.30 Lb

From the contents of 5 capsules, containing about
1,200 seeds, a culture of 992 seedlings appeared in the
pans within 6-8 weeks and grew into well-developed
rosettes. From these a group of 147 rosettes were
readily selected for their uniformly small size and
narrow leaves. They gave rise to a definite class of
dwarfs which remained absolutely distinct from the rest
of the culture; the plants were not etiolated or in other
respects similar to the class of dwarfs separated in the
culture (11.41) from the sister hybrid, 10.30La. The
rosettes constituting the mass of the culture presented a
wide range of form similar to that shown in 11.41, with
the extremes approaching the rosettes of the biennis and
grandiflora parents. Between the extremes was the
similar large assemblage of intermediates of varying
degrees which made a classification of the rosettes im-
possible. There was apparently no clearly marked
tendency of the rosettes in the culture as a whole to
resemble either of the parents.

414 THE AMERICAN NATURALIST [Vou. XLVI

The plants set in the ground numbered 833 large
rosettes and 90 of the dwarf. This culture at maturity
proved much more varied than the former (11.41). The
plants at the more grandiflora-like end of the culture were
in general considerably larger (1.7-2 m. high) than those
at the biennis end (1-1.5 m. high), but there were many
exceptions to the rule. The foliage proved exceedingly
diverse. Occasional: plants (about 20) presented a
foliage similar to grandiflora although never matching
the parent strain in all respects; a smaller number of
plants (about 5) presented a foliage resembling that of
biennis. The foliage, excluding the exceptional types,
ranged from lanceolate leaves to broadly elliptical or
ovate leaves with well defined crinkles.

Many plants bore flowers as large as and larger than
those of the grandiflora parent and with a similar rela-
tion of the stigma lobes to the tips of the anthers. The
larger flowers equaled in size those of the largest-flowered
types of Lamarckiana. Other plants bore flowers ap-
proaching in size those of the biennis parent, but never
so small. There was no fixed relation of the larger
flowers to the plants more grandiflora-like in foliage or
of the smaller flowers to those more biennis-like, although
larger and smaller flowers were in some instances found
on plants approaching the respective parents.

The mass of the culture was fairly close to the F,
hybrid plant 11.30L6, varying from it in habit, and in the
form and measurements of the leaves and flower parts.
There was not so strongly marked a tendency on the part
of the plants towards the grandiflora parent as was
exhibited by the former culture. The flowers, as indi-
cated above, showed the same progressive advance in
size, many of them being larger than those of grandiflora
and none so small as those of the biennis parent.

A count of the plants made with Mr. Bartlett led to a
classification of 104 red-stemmed and 90 green-stemmed
in the group of 833 plants. The remainder of the culture
presented intermediate or mottled coloration similar to

No. 547] GENETICAL STUDIES ON C2NOTHERA 415

the F, hybrid plant (10.3020) from which they came.
The red-stemmed plants were most numerous towards the
more grandiflora-like end of the culture and the green-
stemmed towards the more biennis-like end.

A larger number of remarkable new types appeared in
this culture than in the former (11.41), types specifically
different from either biennis or grandiflora, or the F,
hybrid plant 10.30Zb. The list is as follows:

11.42e, a type represented by 4 plants with small and
narrow leaves, and medium-sized flowers.

11.427, leaves small, irregularly toothed, flowers
medium-sized, capsules large (about 3.3 em. long) ; repre-
sented by several plants and intergrading with the mass
of the culture.

11.429, medium-sized plant with large flowers and Jarge
crinkled leaves; a type rather common and intergrading
with the mass of the culture.

11.42), a single plant with almost linear leaves, flowers
very small (petals 6 mm. long), anthers sterile. In its
foliage this plant seems. very close to De Vries’s
‘‘mutant’’ Œnothera elliptica.

11.421, medium-sized flowers; remarkable for its broad,
entire, and very much crinkled leaves.

11.42r, the class separated as 147 small rosettes with
narrow leaves of which 90 were saved and grown to
maturity. The mature plants, 3-4 dm. high, rarely
branched, bore medium-sized flowers, and constituted a
clearly defined class of dwarfs.

There are not included in the above list the numerous
segregates with characters in various combinations ap- .
proaching one or the other of the parents of the cross.
Some of these types clearly differed specifically from the
F, hybrid plant 10.30Lb.

A consideration of these two cultures (11.41 and 11.42)
of hybrids in the second generation will bring out certain
important conclusions that may be summarized.

1. In the immensely greater diversity exhibited by the
F, generation over that of the F, is clearly shown a

416 THE AMERICAN NATURALIST [Vou. XLVI

differentiation of the germ-plasm expressed by the ap-
pearance in the F, plants of definite tendencies in dif-
ferent directions towards the parents of the cross. This
sems to the writer the essential principle of Mendel-
ism and does not necessarily involve the acceptance of
the doctrine of unit characters and their segregation in
either modified or unmodified form.

2. Certain characters of the parent species have ap-
peared in the F, segregates in apparently pure condition,
but the very large range of intermediate conditions indi-
cates that factors governing the form and measurements
of organs (if present at all) must in some cases be con-
cerned with characters so numerous and so small that
they can not be separated from the possible range of
fluctuating variations. If this is true such characters
seem beyond the possibility of isolation and analysis and
the unit character hypothesis for these cases has little
more than a theoretical interest.

3. Both cultures certainly showed a marked progres-
sive advance in the range of flower size, the largest
flowers having petals somewhat more than 1 em. longer
than those of the grandiflora parent. There was a
similar advance in the size of the leaves and the extent of
their crinkling. These progressive advances would
seem to demand on the unit character hypothesis either the
modifications of the old or the creation of new factors.

4. The absence of classes among the F, hybrids (except
for the dwarfs) further works against the unit character
hypothesis as of practical value in the analysis of a
hybrid generation of this character. It should be remem-
bered, however, that there were in this cross no sharply
contrasted distinctions of color, anthocyan coloration
proving most unsatisfactory for the purposes of a gene-
tical study.

5. Both cultures presented a class of dwarfs, very
different from one another and from the F, hybrid plants
10.30Za and 10.30Lb. This phenomenon of nanism has

No. 547] GENETICAL STUDIES ON CNOTHERA 417

appeared also in all previous F, generations that I have
grown and presents a most interesting subject for study.

6. The extreme variants of the cultures would rank as
new species different from either parent of the cross and
from the F, hybrid plants. Their germinal constitution
(provided it is stable) on the hypothesis of unit char-
acters apparently must involve the modification of old
factors or the creation of whole sets of new factors in
large numbers, a degree of complication unfavorable to
the hypothesis.

7. The progeny of the F, hybrid plant 10.30La was
very different from that of 10.30Lb although these two
plants were sisters. What would have been the compli-
cations if F, generations had been grown from a hundred
or a thousand sister plants!

7. THe PROBABLE CoMPOSITION OF THE CULTURES OF
CARTER AND COMPANY FROM EVIDENCE FURNISHED
BY THE SHEET IN THE Gray HERBARIUM

Brief reference was made in a former paper (Davis,
"11, p. 228) to a very interesting sheet in the Gray Herba-
rium of Harvard University, which shows a plant with
characters in part those of @nothera Lamarckiana, but,
in the writer’s opinion more largely those of grandiflora.
This sheet bears notes in the handwriting of Dr. Asa
Gray to the following effect—in ink, ‘Œ. Lamarckiana,”’
‘Hort. Cantab. 1862,’’ and ‘‘from seed of Thompson,
Ipswich’’; in pencil and probably of a different date
‘‘said by English horticulturists to come from Texas.”’
It was the habit of Dr. Gray at that time to use herbarium
labels marked Hort. Cantab. and this fact, together with
the absence of other writing on the sheet indicates that
the plant was grown in the botanical garden at Cam-
bridge, Massachusetts.

The date, 1862, suggested a possible genetic relation-
ship of this plant in the Gray Herbarium to the cultures
of Messrs. Carter and Company of London about 1860 of
plants which were grown for the market under the name

418 THE AMERICAN NATURALIST [Vou. XLVI

(Enothera Lamarckiana and which seem to have been the
starting point of the Lamarckiana now under cultivation.
The only descriptions which we have of these cultures
are the very unsatisfactory accounts in The Floral Maga-
zine, Vol. II, Plate 78, 1862, and in ‘‘ L’Illustration Horti-
cole,’’ Vol. IX, Plate 318, 1862, both accompanied by the
same figure of an impossible Hnothera.

An inquiry was at once started to determine the mean-
ing of the note ‘‘from seed of Thompson, Ipswich,’’ and,
thanks to the courtesies of correspondents, the matter
now seems clear. There seems to have been no botanist
or horticulturist of the name of Thompson in Ipswich,
Massachusetts, at that period from whom Dr. Gray could
have obtained this seed. The reference is almost certain
to have been to William Thompson of Ipswich, England,
who died several years ago. William Thompson ‘‘was a
seedsman in a small way of business, but a most enthu-
siastic cultivator with correspondents in all temperate
countries and the introducer of numerous herbaceous
plants more particularly annuals and biennials.’’ He
corresponded with Kew as early as 1860.

It is ina high degree probable that William Thompson,
with his interest in novelties, obtained from Carter and
Company their new @nothera and that the seed sent to
Dr. Gray was either directly from this source or from
plants cultivated by Thompson. It will be remembered
that Carter and Company stated that they received their
seed from Texas, which accords with Dr. Gray’s note
‘‘said by English horticulturists to come from Texas.”
If this interpretation of the history of the sheet in the
Gray Herbarium is correct the plant was very close
indeed to the cultures of Carter and Company, possibly
not more than one or two generations removed, since
these plants were probably cultivated as biennials. This
sheet then gives evidence on the composition of the cul-
tures of Carter and Company immensely more valuable
than the obviously-inaccurate plate of The Floral Maga-

No. 547] GENETICAL STUDIES ON CGENOTHERA 419

zine and the description that tells little of value for the
- problem under consideration.

The following is a description of this sheet in the Gray
Herbarium:

1. Stems and Foliage——The stem bears long hairs
arising from papille (glands) which are colored red as
in Lamarckiana and are about as numerous as in that
species. A detached leaf (Fig. 15, a), about 18.5 em. long
with sinuate margins, slightly lobed below, and with some
evidence of former crinkles, suggests by its shape (al-
though too small) the basal leaves of Lamarckiana. The
upper foliage is similar to a broad-leaved type of grandi-
flora, the leaves being short-petioled and not so nearly
sessile as in Lamarckiana.

2. Inflorescence.—The inflorescence has longer inter-
nodes than in Lamarckiana and consequently is not so
compact; in this respect it resembles grandiflora. The
bracts are broad at the base, slightly toothed, and per-
sistent, becoming lanceolate leaves on the fruiting
branches as in grandiflora.

3. Buds.—The buds (Fig. 15, b) are about 9.5 cm. long,
not stout and 4-angled as in Lamarckiana, but with a cone
circular in section as in grandiflora. Sepals apparently
green, their tips attenuate as in grandiflora and project-
ing 1 cm. beyond the folded petals; pubescent, with long
hairs arising from papilla among much shorter sessile
hairs as in Lamarckiana.

4. Flowers.—The flowers are somewhat larger than
those of any grandiflora known to me. Petals about 4.5
em. long, as long as those of the largest forms of La-
marckiana., Stigma lobes about 8 mm. long, close to 9
mm. above the tips of the anthers, in these respects agree-
ing with both Lamarckiana and grandiflora. :

5. Capsules.—The capsules, about 3 em. long, are of
medium thickness and similar to those of grandiflora;
they are not so stout as the capsules of Lamarckiana.
The persistent leafy bracts and long internodes give the

420 THE AMERICAN NATURALIST [Vou. XLVI

fruiting branches a marked resemblance to the conditions
characteristic of grandiflora.

This plant then presents grandiflora characters in the
upper foliage of the plant, in the longer internodes of the

”

`G. 15. Sheet in the Gray Herbarium of Harvard University. An M@nothera
grown by Dr. Asa G at Cambridge, Massachusetts, in 1862 and probably
derived directly or nie ‘tly from the cultures of Carter and Company of
Pone which were pragas under the name Lamarckiana. The specimen on
this herbarium sheet has important characters of grönulora. indicating a rela-
iocielite to this species

inflorescence and the persistent bracts, in the form of the
buds (4-angled) with attenuate sepal tips, and in the
longer capsules of medium thickness. The plant resem-
bles Lamarckiana in the red coloration of the papille on

No. 547] GENETICAL STUDIES ON ŒNOTHERA 421

the stem at the base of the long hairs, and in the form
and size of a large detached leaf (probably basal). The
form of the flowers is essentially similar to either grandi-
flora or the large-flowered types of Lamarckiana; their
size is very close to that of the latter plant, which it also
resembles in the pubescence of the sepals. In this mix-
ture of characters the most important are, in the writer’s
opinion, distinctly grandiflora-like and indicate a. close
relationship to some strain of grandiflora.

The short description in The Floral Magazine, Vol. II,
1862, quotes the following from Carter and Company.
‘We received, about four years ago, some seed from
Texas unnamed. When we had flowered it, we sent some
blooms to Dr. Lindley, who pronounced it to be Œnothera
Lamarckiana, a species, we believe, introduced into Eng-
land by Mr. Drummond. Its height is between three and
four feet; it blooms the first year, is a very hardy
biennial, and: is superior to any other @/nothera in the
size and number of its blossoms, which measure four
inches in diameter.’’ Of the characteristics noted in this
quotation, all of which fit Lamarckiana, the most impor-
tant is the statement that the plant is ‘‘a very hardy
biennial,’’ although this is somewhat weakened by the
remark that ‘‘ it blooms the first year.” :

Now grandiflora is clearly annual and the @notheras
of the south and southwestern United States are, as far as
the writer is aware, generally annual or perennial. A well-
defined biennial habit, characterized by the development
of largeand short persistent rosettes, is an adaptation to
the short seasons of northerly climates. The question arises
whether the plants raised by Carter and Company are
represented or could have been represented in the Texan
flora. We know that Texas and the southwest generally
have some large-flowered @notheras and it may be that
the climatic conditions of certain mountainous regions
would favor the development of a biennial habit. How-
ever, botanical exploration has not yet brought forward
any plant of the Lamarckiana type. In the absence of

o

422 THE AMERICAN NATURALIST [Vor. XLVI

corroborative evidence we can hardly at. present accept
as beyond doubt the statement that the seeds of Carter
and Company were from Texan plants.

It is a confused situation with a number of possible
explanations which are not worth a discussion until we
have more evidence at hand. This evidence may come
through other herbaria sheets comparable to the one in
the Gray Herbarium, although examinations by my corre-
spondents of the collections at the British Museum, at
Kew, and at Cambridge University indicate that there is
nothing at these centers. Evidence may also come from
field studies in America, which should be pushed by
western and southern botanists in a position to observe
(Enothera throughout the season. The writer sowed
about 200 seeds (60 years old) from one capsule on the
sheet in the Gray Herbarium, but there have been no
germinations after being four months in the seed pan; the
experiment, a forlorn hope, needs no apology, considering
the importance of the problem—the composition of the
cultures of Carter and Company.

The most important point for the writer’s hypothesis
of the hybrid origin of @nothera Lamarckiana as a cross
between biennis and grandiflora is the strong evidence
that the cultures of Carter and Company contained forms
with characters in part grandiflora-like and in part
biennis-like, for it must be remembered that Lamarckiana
in its hardy biennial habit and in other characters resem-
bles the latter species. It is possible that the plants were
hybrids at that period (1860) for if the sheet in the Gray
Herbarium is representative of the composition of these
cultures, the plants were very different from the present
day Lamarckiana. Even if the plants of Carter and
Company came from Texas as a form related to grandi-
flora there would of course have been abundant oppor-
tunities for hybridization with biennis before De Vries,
a quarter of a century later, began his studies.

No. 547] GENETICAL STUDIES ON (ENOTHERA 423

8. FURTHER CONSIDERATIONS ON THE POSSIBLE ORIGIN OF
(Hnothera Lamarckiana as a HYBRID oF
O. biennis and O. grandiflora

Among the most interesting of the results from the
cultures of the past summer (1911) have been those con-
cerned with the behavior of the hybrids of biennis and
grandiflora in the second generation. In the progeny of
the F, hybrid plants 10.30La and 10.30Lb there was clear
evidence of an advance in the size of the flowers which
carried them in many plants beyond the size of the
grandiflora parent and as far as the larger-flowered
forms of Lamarckiana. The petals of these F, hybrid
plants measured about 2.2 cm. ; those of the largest of the
F, flowers measured 4.5 em.; grandiflora B has petals
about 3.3 em. long and those of the largest-flowered forms
of Lamarckiana measure about 4.5 em. There was a
similar advance in the size of the leaf and the extent of
its crinkling, although the leaves did not reach the condi-
tions characteristic of Lamarckiana. This is clearly pro-
gressive evolution and doubtless can be maintained
through selection, though there may be in later genera-
tions constant retrogressive variation. :

This demonstration of a progressive advance in flower
measurements in the F, generation disarms those critics
(as Gates, 711) who in prematurely discussing my F,
hybrid plants 10.30La and 10.30Lb apparently failed to
appreciate the basic principle that the characteristics of
an F, hybrid are shown by what it will do in the F, gen-
eration and not by what it may seem to be. No experi-
menter in touch with present-day genetics would hope to
get a perfect Lamarckiana type in the F, generation from
a cross between biennis and grandiflora, for the reason
that the characters of these species could neither blend
nor appear as dominant or recessive in such a manner as
to give this type. But with a proper selection of parent
biotypes there is good reason to be hopeful that from the
F, generation of such a cross forms will appear with

424 THE AMERICAN NATURALIST [Vor. XLVI

characters so shifted and modified as to match very
closely those of Lamarckiana. ,

My hypothesis does not of course demand that La-
marckiana arose with all of its characters fully present.
as the result of a simple cross. There have been abund-
ant opportunities for repeated crosses of a most varied
character during the long period in which this plant has
been under cultivation. That the plant hybridizes readily
in nature is evidenced by the complex assemblage of
varied types found in such localities as the sandhills of
Lancashire, England, where Lamarckiana-like forms are
growing wild in great numbers together with types of
biennis. The fundamental feature of the hypothesis is
the belief that Lamarckiana exhibits the behavior to be
expected of a complex hybrid and that its characteristics
are those likely to come from a mixture of germ-plasms
from types of biennis and grandiflora.

I shall not, in order to perfect a synthesis of Lamarck-
iana-like hybrids, depart from my plan of experimenting
with wild races of biennis and grandiflora, for the most
important feature of the line of experimentation under
way is the study of the behavior of species hybrids. Such
a study offers the opportunity of treating experimentally
one of the most vital problems of genetics,—the possibili-
ties and methods of the progressive evolution of specific
characters through hybridization.

The studies have not yet proceeded sufficiently far to
justify a detailed comparison between the various forms
of F, hybrids and the ‘‘mutants’’ of De Vries. It is clear,
however, that the F, generation will give numbers of
types that differ from F, hybrid plants in the same
manner as the ‘‘mutants’’ differ from Lamarckiana.
Some of the variants have characters so different from
the F, parent hybrid plants that they would stand in
taxonomic practise as new species, others have differ-
ences of such a nature that they might be called progres-
sive, or retrogressive varieties. The chances are im-
mensely against obtaining a Lamarckiana-like hybrid

No. 547] GENETICAL STUDIES ON ŒNOTHERA 425

which will produce the same series of variants as the
‘‘mutants’’ obtained by De Vries, for the reason that no
' two F, hybrid plants in so complex a cross as that
between biennis and grandiflora are likely to give exactly
the same set of variants. This principle is strongly indi-
cated in the diverse F, progeny of the two sister plants
10.30La and 10.300.

It is, naturally, important for my hy pothesis that
a like hybrids continue to give in successive
generations similar variants after the manner of La-
marckiana, but, as noted above, it is not necessary that
they be taxonomically the same forms (i. e., gigas, rubri-
nervis, nanella, ete.), or that they be produced in the same
proportions. Upon these points we shall sooner or later
have definite data.

The results from the work of the past summer (1911),
described in this paper, have strengthened the writer’s
hypothesis in the following respects:

1. A definite advance has been made in obtaining more
favorable F, hybrids by employing a biotype of biennis
(biennis D) with characters closer to those of Lamarck-
tana than the characters of biennis A and B previously
used. l
2. The F, generations from the hybrid plants 10.30La
and 10.30Lb have shown that large numbers of variants
may arise from a cross between biennis and grandiflora,
some of apparently new specific rank, others exhibiting
variations of either a progressive or retrogressive
nature, and very many presenting in various degrees of
complexity a segregation of the characters of the parents.

3. The progressive advance in flower and leaf size over
that of the parents of the cross indicates that selection of
the sort common among gardeners (i. e., the choice of the
largest and most vigorous plants) might readily estab-
lish a race surpassing in these and other respects both
parent plants. It is exactly this sort of selection which, —
in the writer’s opinion, has established the ea
forms of Lamarckiana. |

426 THE AMERICAN NATURALIST [Von. XLVI

It is to be regretted that the terms mutant and muta-
tion are being used so variously by different workers in
the fields of genetics. ‘‘Mutations’’ have been described
which are obviously from heterozygous parentage, in
some cases small differences, as of color or measurement,
in other cases very large differences. In contrast to
these have been described ‘‘mutations’’ from possible
homozygous parentage. Germinal variation due to the
mixing of different germ-plasms, and consequently from
heterozygous material, is a common phenomenon and
easily defined. Germinal variation in homozygous ma-
terial presents an equally clear concept.: The two types
of variation should be sharply distinguished for one of
‘the most important fields of genetical research is the
study of possible germinal variations in homozygous
stock and the conditions under which they may occur.

In our understanding of germinal variations (muta-
tions) as distinguished from somatic variations (fluctua-
tions) a great advance has been made towards a clear ap-
preciation of the problems involved. The problems cen-
ter in the study of germinal variations which are of two
types since they may be due to the mingling of germ
plasms (amphimixis), or to environmental influences.
Amphimixis is of course responsible for heterozygous
conditions now better understood as a result of Mende-
lian studies. Relatively little is known, however, of the
conceivable effects of environment as a source of germi-
nal variation. It may be doubted whether the specific
terms ‘‘mutation’’ and its alternative ‘‘fluctuation,’’ as
commonly used, are well chosen since the concepts are so
clearly expressed by the designations germinal and
somatic variations.

It is clear that De Vries regarded the ‘‘mutants’’ of
Lamarckiana as variants from a type representative of a
wild species and as nearly homozygous as most well-
defined species. That the Lamarckiana with which De
Vries worked was strongly heterozygous, in fact a hybrid
of biennis and grandiflora, is the hypothesis for which I
am trying to present as much evidence as possible. By

No. 547], GENETICAL STUDIES ON GENOTHERA 427

this hypothesis, if finally accepted, most of the
‘‘mutants’’ of De Vries are likely to fall into the class of
variants due to the mixing of different germ-plasms. If
the word mutation, in the sense of De Vries, is to have
a meaning more precise than that of variant it must be
kept for the type of variation from homozygous stock.

That germinal variations may occur in homozygous
material seems to the writer more than probable, and it
is possible that many small variations are of this type.
The trend of experimental investigation, however, dis-
tinctly indicates that large variations (ranking as salta-
tions) are rare if present at all in homozygous material,
and consequently can not be important factors in organic
evolution. The variations considered by Darwin are
chiefly either the small variations in relatively homozy-
gous forms, or the large and small variations from heter-
ozygous stock.

UNIVERSITY OF PENNSYLVANIA,

February, 1911
LITERATURE CITED

Bartlett, sA H., ’11. @Œnothera Tracyi sp. nov. Rhodora, Vol. XIII, p.

209, 1911.
Davis, B. M., 711. Some Hybrids of Œnothera biennis and O. grandiflora

that reacinbte O. Lamarckiana, AMER. NAT., Vol. XLV, p. 193, 1911.
De Vries, Hugo, ’05. Ueber die Dauer der Mutationspériods os Gnothera

Tiitii. Ber. deut. bot. Gesell., Vol. XXIII, p
De Vries, Hugo, ’07. On Twin Hybrids. Bot. Gaz., Vol. p a p. 401,

1907.

De Vries, Hugo, ’09-’10. The Mutation Theory. Chicago, 1909-10.

De Vries, Hugo, 711. ene nfo AT Bastarde von @Œnothera
biennis L. und O. muricata L. Biol. Centbl., Vol. XXXI, p. 97, 1911.

Gates, R. R., ’10. Tho Earliest Description of Ginothera Lamarckiana.
Beicnce, Vol. XXXI, p. 425,

Gates, R. R., ’lla. Early aaee a Records of the CEnotheras.
Proc. Towa Acad. Sci. for 1910, p. 85, published in 1911.

Gates, R. R., ’11b. Mutation in Œnothera. AmER. Nat., Vol, XLV, p.

1

+» VOL č
il, A. + Shull, = H, a Small, J. K., 05. Mutants
and Hybrids of the Œnotheras. Carnegie Inst., Pub. 24, 1905.
MacDougal, D. T., Vail, A. M., and Shull, G. H., ’07. Mutations, Varia-
tions and Riclationships of üis Œnotheras. Carnegie Inst., Pub. 81,
1907.

EVIDENCE OF ALTERNATIVE INHERITANCE
IN THE F, GENERATION FROM CROSSES
OF BOS INDICUS ON BOS TAURUS

DR. ROBERT K. NABOURS

KANSAS STATE AGRICULTURAL COLLEGE

. THe common domestic cattle of India appear to be a
distinct species (Bos indicus). They are mainly charac-
terized by a large hump on the fore shoulders, short
horns, large drooping ears, extensive dewlap and sheath
(Fig. 1). There are several varieties, or breeds, but they
are so commonly hybridized that it is exceedingly difficult
to ascertain which are the pure strains and which hybrids.
In this respect they are probably analogous to. Bos
taurus. In size they vary greatly, ranging from very
diminutive breeds to those the largest individuals of
which weigh upwards of 2,000 pounds. The males are
considerably: heavier than the cows. The colors also vary
considerably, the most common being creamy buff, brown,
ashy gray, red, black and white, and blends of these.

They appear to be highly resistant to the cattle dis-
eases of tropical and subtropical countries, and they are
immune to the attacks of cattle ticks, that is, ticks do not
remain attached to them and suck their blood (Figs. 1
and 9), and they are said to be less liable to suffer from
the effects of the bites of insects than any of the breeds of
Bos taurus.

They are very gentle and docile. In India the males
are used as beasts of draught, and are yoked to the plow,
being the main animal used for tilling the ground. They
are very agile, being able to travel thirty or more miles
a day, carrying a heavy burden, or drawing a cart with a
considerable load on it. Recently they have been intro-
duced into Jamaica in considerable numbers, where they

428

No. 547] ALTERNATIVE INHERITANCE IN BOS 429

ty =

Fig, 1. suas bull (Bos indicus) When in good condition,
is bull weighs 2,000 pounds.

are used on the banana plantations. It is generally re-
ported that the strong draught oxen of Spain have been
derived from crosses between Indian cattle and the native
Spanish cattle, but the evidence is at best only anecdotal
and nothing whatsoever seems to have been recorded of

430 THE AMERICAN NATURALIST [VoL. XLVI

the inheritance behavior of the progeny of any of the
crosses.

The hump of some of the largest specimens may weigh
as much as fifty pounds, and it is esteemed by English
residents of India as a delicacy for the table. From the
Persian province of Gilan, on the Caspian, the humps,
smoked, of a small breed are shipped to parts of Russia
where they are in much demand as a delicacy. The meat,
products of these cattle, on the whole, are said to be
unexcelled. Some of the breeds give milk that is ex-
cessively rich in quality, but it does not appear that any
of them produce it in large quantities.

It appears that Brahma (Bos indicus) cattle were first
brought to the United States in 1853 by Mr. Davis, of
South Carolina. These cattle were subsequently taken
westward and their progeny distributed throughout the
southwest and parts of Mexico. In southern Texas and
parts of Mexico there are many native cattle that are
said to carry the blood of these cattle and other Indian
stock which were secured from menageries and circuses.
The common brindle cattle of these regions are said to be
descendants of the Indian on native cattle. Wherever
these part Indian cattle are found there is a general im-
pression among stockmen that they are thriftier and
larger than the native stock and more resistant to the
ravages of diseases, ticks, and insect pests.

In 1906 Mr. A. P. Borden, of Pierce, Texas, imported
about thirty head of Brahma cattle, mostly young bulls,
and since then he has been crossing them quite exten-
sively on native Texas cattle and on grade Durhams and
on grade Herefords. For an account of the experiences
in importing these cattle and the beginning of the experi-
ment, the reader is referred to Mr. Borden’s interesting
paper, ‘‘Indian Cattle in the United States,’’ The Ameri-
can Breeder’s Magazine, Vol. 1, No. 2.

In September, 1911, Mr. Borden kindly permitted me
to visit his herds for the purpose of studying them and
making photographs. The following preliminary ac- —

No. 547] ALTERNATIVE INHERITANCE IN BOS 431

count of three of the herds is given, because they tend to
show definite inheritance results. The study of the herds
and the work of making the photographs were greatly
hampered by the absence of the head herdsman, the rain,

ə, ı hybrids T Bos indicus on Hereford. The heifer on the right
weis sax ,000 pounds at twelve months of age. The bull on the left weighed
1,450 ip ipi at twenty-six eb of age. Photo furnished by Mr. A. P. Borden.

and the very short time at my disposal. I plan to make
a further study of them at an early date.

Whenever the Brahma cattle have been crossed on
grade or pure Herefords, the color characters of the
latter are on the whole dominant in the F, progeny (Figs.

i Hybrids from Bos indicus on Hereford and Durham. rae bull in the
center is from peters mother and weighed about 1,400 pounds 22 months
of age. All are tick-fre

432 THE AMERICAN NATURALIST [Vou. XLVI

3 and 4). The F, progeny from crosses on grade Dur-
ham show the Durham color and other characters to be
dominant. The progeny from crosses on native Texas
cattle of unknown constitution are very variable. In
some cases the F, progeny from the latter resemble the
Brahma greatly. However, this apparent dominance of
the Brahma in many cases, when crossed on native cattle,
is probably altogether due to the fact that a considerable
proportion of the native cattle already have Brahma
characters in them—their immunity to ticks and other de-
sirable qualities having favored their perpetuation since
the early introduction of the Davis and other Brahma
stock.
Herp No. 1

This herd consists of twenty-five or thirty F, cows
(Fig. 5, adults) from a white Brahma sire on grade Dur-

‘1G, 5. Fy mothers from Bos — on grade Durham and a few te Here-

for oa The white calf, on the right, is from F, half-brother of the co The
pure Durham calf, on left, and seed in center are from Durham sire.

,

hams and a few grade Herefords. Twenty-four of these
F, cows were mated to their Brahma-Durham F, half-
brother, and eighteen F, calves were produced. Of
these, six are white and resemble the grandfather white
Brahma. One of the calves is shown in Fig. 5, on the
right. The remaining twelve F, calves from these
crosses of hybrid X hybrid resemble the Durham mostly,

No. 547] ALTERNATIVE INHERITANCE IN BOS 433

though there is evidence of the Hereford characters in a
few. A few of these F, cows were mated to a registered
Durham bull. Only two of the F, calves from these crosses
were observed. They are shown in Fig. 5, on the left and
in the middle. The one on the left is a pure Durham in
all respects. The one in the center is a typical Brahma.
Durham hybrid. A downpour of rain and the scattering
of the herd prevented further observations.

Herp No. 2
The F, cows of this herd are the product of a Brahma
male on grade Hereford and grade Durham in about

Fic. 6. The cows are from Bos indicus on impure ea and the calves
from these F, cows bred back to Bos indic

equal proportion. These F, cows show the color char-
acters of the Hereford and Durham quite distinctly,

Fic. 7. The yearlings are from Bos indicus x high-grade Hereford Fi
hybrids bred back to Bos indicus.

though there is slight evidence of the hump of the
Daku sire, and the dewlap is somewhat enlarged (Fig.

434 THE AMERICAN NATURALIST [Vou. XLVI

6, adult). (The brothers of these cows were not ob-
served.) The F, calves of this herd are from a Brahma
sire on these F, cows. My notes fail to show whether or
not the sire of the F, calves is the same as that of their
F, mothers. The sire of the F, calves appears to be in-
termediate in color between the white and brown Brahma.
The calves are of two distinct types, about one-half of
them having the Brahma characters and the other half
bearing the characters of their hybrid mothers (Fig.
6, calves). In the figure (Fig. 6) a good type of the
hybrid resembling the mother is the sucking calf at the
right, while several apparently pure Brahmas are shown
in the foreground.

Herp No. 3

The F, cows of this herd are the progeny from a
Brahma sire on high grade Hereford cows. The F,
calves and yearlings are from another Brahma sire on the
Brahma-Hereford hybrid. On counting these thirty-two
calves and yearlings, it was found that seventeen of them
resemble mostly the sire and grandsire Brahma while
fifteen come nearer to the type of the Aybnd mothers |
(Fig. 7, calves and yearlings).

CONCLUSIONS CONCERNING THE INHERITANCE BEHAVIOR

It appears that the color patterns of Herefords and
Durhams are dominant in the F, generation. However,
the hump, large sheath and dewlap of the Brahma show
slightly in the Brahma X Hereford or Durham F,
progeny. It is clear that in the F, generation, pure
Brahma and pure Durham are segregated. Indications
are that when the parent strains are pure the segregation
follows the simple law of alternative (Mendelian) in-
heritance. However, the conditions of the experiment,
the lack of full knowledge of the constitution of the
parents and the inadequate observations prevent any
positive conclusions as yet concerning the ratios.

No. 547] ALTERNATIVE INHERITANCE IN BOS 435

Immunity To THE Texas CATTLE Ticks
I am able fully to confirm Mr. Borden’s statements
(loc. cit.) that the pure Brahma cattle and the hybrids
are perfectly immune to the Texas cattle tick. Fig. 8
shows the ordinary conditions of the native Durham or
Hereford cattle, while Figs. 9, 1 and 3 show the condi-

Fic. 8, A _ tick-infested Hereford Fic. o A sie indicus cow free
fre ticks.

Each of TE cows has suckled a calf during sa summer and both
e been together in the same pas

tions of the pure Brahma and ee all running
together on the same range. I was not able to ascertain
definitely the inheritance behavior of this character (im-
munity to ticks) in the F, progeny. However, it is ex-
pected that data on this point will be available in the
spring or summer of 1912.

SIZE AND PROLIFICNESS

The statement by Mr. Borden (loc. cit.) that the
hybrids running on the range average about 50 per cent.

436 THE AMERICAN NATURALIST [Vou. XLVI

larger than the ordinary native range cattle is fully con-
firmed by my observations. The hybrids shown in Fig.
4 (center), which had had no other advantage than the
range conditions, weighed 1,400 pounds at two years old.
The hybrid heifer, shown in Fig. 3 (at the right), which
had run on the range, weighed 1,000 pounds at twelve
months old. The bull, Fig. 3 (on the left), weighed 1,450
pounds at 26 months of age. These weights appeared to
me to be more than 50 per cent. greater than the average
of the native cattle at the same age kept under similar
circumstances.

A pure Brahma bull will put seventy-five to eighty
cows with calf each season, while the native or even high-
grade Hereford or Durham will impregnate only twenty-
five or thirty cows.

I desire to express my gratitude to Mr. Borden for
courtesies shown me while studying and photographing
the herds and for reading and correcting the manuscript,
and to Professor T. J. Headlee for the arrangements
which made the trip possible; I am further indebted
to Professor Headlee for suggestions while writing the
paper and arranging the illustrations.

SHORTER ARTICLES AND DISCUSSION

ON THE INHERITANCE OF TRICOLOR COAT IN
GUINEA-PIGS, AND ITS RELATION TO GAL-
TON’S LAW OF ANCESTRAL HEREDITY*

In 1889 the late Francis Galton formulated a ‘‘law of ancestral
heredity’? which he believed would prove to be of general
applicability in as much as he had already applied it with
gratifying results to two widely different categories of cases, viz.,
the height of man and the color of dogs. The law as stated by
Galton is:

The two parents contribute between them, on the average one-half,
or (0.5) of the total heritage of the offspring; the four grandparents,
one-quarter, or (0.5)?; the eight great-grandparents, one-eighth, or
(0.5)*, and so on.

The validity of Galton’s law has since been seriously called in
question, though neither of his two cases has yet received a
wholly satisfactory explanation, but sufficient progress has been
made in the study of heredity to show the unsoundness of the
basic principle on which Galton’s generalization rested. In
reality there is no such thing as ‘‘ancestral’’ inheritance; for we
inherit from our parents only, not from our more remote an-
cestors. It is true that a knowledge of the more remote ancestry
may help us to understand better what we have inherited from
our parents, but it is clear that such knowledge of the ancestors
will not enable us to predict the character of their descendants in
the precise manner outlined by Galton. In a specific case, the
inheritance of albinism in mice, I made in 1903 a test of the
comparative value of Galton’s law and Mendel’s law in predict-
ing the character of offspring, with the result that Mendel’s law
gave predictions closely according with observation, whereas
predictions based on Galton’s law proved wholly unreliable.

As regards height, however, and other size characters, Gal-
ton’s law is quite as good a basis for predicting the result of
particular matings as is Mendel’s. Indeed it is not clear that
either of them is applicable to such cases as human stature.

‘From the Laboratory of Genetics of the Bussey Institution, Harvard
University.

437

438 THE AMERICAN NATURALIST [Vor. XLVI

The other case studied by Galton, the inheritance of the tri-
color condition in Bassett hounds, no one up to the present time
has attempted to explain on any ground other than that taken

y Galton in his law of ancestral heredity, although it is now
generally conceded that his law does not apply to color in-
heritance in general. :

The tricolor condition seen in dogs is found in very few other
animals, one of which however is the guinea-pig, the tricolor
variety of which I have had an opportunity to study for some
years. The observations made on tricolor guinea-pigs throw
light on the facts observed by Galton regarding tricolor dogs.
They show wherein Galton’s interpretation fails and what ad-
vantages a Mendelian interpretation has as applied to this case.

The tricolor condition of guinea-pigs is one of long standing.
Cuvier reports it as depicted by Aldrovandus, who made the
first scientific description of the guinea-pig, soon after 1550.
This color variety had probably been in existence for a long
time before the discovery of America. The natives of Peru still
rear it for food in the recesses of their adobe cabins, as their
ancestors have doubtless done for untold centuries. Neverthe-
less it does not breed true, and can not be made to do so.

The tricolor animal is white marked with irregular but distinct
blotches of black and yellow. Tricolors produce besides tri-
colors young which are black-and-white or yellow-and-white, but
never in my experience those which are wholly free from white.
In other words they breed true to spotting with white, but not to
spotting with black and yellow. The black-and-white as well as
the yellow-and-white offspring of tricolor parents may produce
tricolor young. Indeed any one of these three conditions is able
to produce both the others. See Table. Nothwithstanding the

TABLE SHOWING THE KINDS oF YouNG PRODUCED By A RACE OF GUINEA-
PIGS CONTAINING TRICOLOR INDIVIDUALS
B-W, black-and-white; Y-W, yellow-and-white; T, tricolor

Parents 1S
T B-W Y-W Per Cent. T
T... 8 1 5 57

TXS B-W 9 3 5 53
TS Y-W 13 2 9
B-WX B-W. 1 3 2 17
B-WX Y- 6 | 1 1 75
Y-WX Y-W i i 9 45

oe a $ | P 31 52

No. 547] SHORTER ARTICLES AND DISCUSSION 439

fact that neither the black-and-whites nor the yellow-and-whites
produced by tricolors breed true, there are races of black-and-
white and of yellow-and-white guinea-pigs which do breed true.
It remains to explain why the others do not. A black-and-white
animal which breeds true may be considered to possess some
chemical substance necessary for the production of color (which
we call a color-factor) distributed irregularly throughout its
coat. Wherever this substance is wanting no color is formed
and a white area results. The specific factor for black (prob-
ably an enzyme) is however everywhere present in the coat so
that wherever color forms the color is black. Such races as
this breed true.

The yellow-and-white animal which breeds true may likewise
be considered to have an irregularly distributed color-factor, but
to lack entirely in its coat the black factor. Hence the color,
wherever formed, is yellow.

Yellow races also exist which do not bear spots of white, but |
which have spots of black. In such animals the color-factor is
evidently uniform in distribution, whereas the black factor is
irregularly distributed.

Now the tricolor race is a yellow one spotted both with white
and with black, i. ¢., it results from irregularity in distribution
through the coat of two different chemical substances, the color
factor and the black factor. These two factors are known to be
independent of each other in heredity. See Castle (1909). It
is therefore not to be supposed that they will commonly coincide
in distribution. If the black factor extends over all the colored
areas, the animal will be black-and-white. If the black factor
falls only on areas which lack the color factor, it will produce no
visible effect, and the animal will be yellow-and-white. If,
finally, the black factor falls on some of the colored areas but not
on all of them, those in which it falls will be black, the others
yellow, and the uncolored areas of course white. Hence a tri-
color will result. But the gametic composition of these tricolors
Will not be different from that of the black-and-whites or red-
and-whites produced by the same race, since all alike will be
characterized by irregularity in distribution of the same two
factors. A tricolor race on this hypothesis should be unfixable,
as has up to the present time been found to be true.

The same line of explanation will answer equally well for the
case of the Bassett hounds studied by Galton. These Galton

440 THE AMERICAN NATURALIST [Vou. XLVI

classifies as either ‘‘tricolor’’ or ‘‘non-tricolor’’ (‘‘lemon-and-
white’’). A third variety, ‘‘black-and-tan,’’? is mentioned by
Galton, but disregarded in his statistical treatment. He does not
state whether it may or may not possess white spots, but I
strongly suspect that the ‘‘black-and-tans’’ produced by spotted
Bassett hounds would also be spotted, in which case they would
correspond with the category black-and-white of guinea-pig
races containing tricolors. For the ‘‘tan’’ feature of black-and-
tan dogs is in reality due to a pattern factor as distinct from
black as is the agouti pattern of guinea-pigs. Hence so far as
spotting is concerned Galton’s disregarded ‘‘black-and-tans’’

were probably black-and-whites which happened to possess the
‘‘tan’’ pattern (light spot over eye, light belly and legs below).
If so, the behavior of the dogs accords in every point with that
of the guinea-pigs as regards color inheritance. For Galton
observed that neither tricolors nor non-tricolors breed true, but
each sort may produce the other, though it produces more of its
own kind. The same is true for guinea- pigs; see Table. It is
desirable that any one having the opportunity should look into
the breeding capacity of Bassett hounds which are black-and-
white or ‘‘black-and-tan.’’ If they can produce ‘‘tricolor’’ and
“‘lemon-and-white young,” the parallel between the breeding
capacity of tricolor guinea-pigs and tricolor dogs will be com-
pletely established. W. E. CASTLE

BUSSEY INSTITUTION,
HARVARD UNIVERSITY

BIBLIOGRAPHY
Castle, W. E.
1903. The Laws of Heredity of Galton and Mendel, and Some Laws
Governing Race Improvement a Selection. Pree. Amer. Acad.
of Arts and Sci., Vol. 39, pp. 223-242.
1905. Heredity of Coat harioti i in Guinea-pigs and Rabbits. Carnegie
nstitution of Washington, Publication No. 23.
Castle, W. E., in collaboration with H. E. Walter r, R. C. Mullenix and S. Cobb.
1909. Studies of Inheritance in Pe Carnegie Institution of Wash-
ington, Publication No.
Cumberland, C
1901. The Guinis -pig or Domestic Cavy. 100 pp. illustr. London, L.
Upcott Gill
Galton, F
1889. Natural Inheritance. Macmillan Co., London and New York.
1897. The Average Contribution of Each Several Ancestor to the Total
Heritage of the Offspring. Proc. Roy. Scc. London, Vol. 61,
pp. 401-413,

VOL. XLVI, NO. 548 | AUGUST, 1912

The American

Naturalist

MSS. intended for publication and books, etc., intended for review should be
sent to the Editor of THE croatian NATURALIST, Garrison-on-Hudson, New York.

Articles containing

tion are especially w

One hundrea reprints o

inher Mal ase will > supplied at cos
bscriptions d advertise

eren price is yadi E a year

ts additional.

ork bearing on the problems of _— evolu-
welcome, and will “i given preference in publicatio
of epee gigs are supplied to authors a of charge.

should be sent to the dgio The
Foreign postage

The charge for single portia is
a page.

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

THE SCIENCE PRESS

NEW YORK: Sub-Station 84

Lancaster, Pa.

Entered d-class matter, April 2, 1908, at the Post Offi

Garrison, N. Y.

at Lancaster, Pa., under the Act of

Congress of March 3, 1879.

THE BULLETIN—For bargains in Ethnolograph-
ical and Pre-historic Specimens. Books on Natural
History, Science, Travel, Voyages, etc. See THE
BULLETIN post free for 3 cent stamp.

4 Duke St., Adelphi—London—England

TENTH EDITION.
THE Memesi

COMSTOCK PUBLISHING p praen N. Y.
a a ah,

- Reptiles, Amphibians

and Fishes-

z WF Paia

University of Colorado
Mountain Laboratory

Fourth summer session June 24 to Au-
gust 2, 1912. Courses in Field Biology

(animal and plant), Ecology, Systematic

‘Botany, Biology of Ponds and- Streams. i

Suitable work for undergraduates and grad- a

uates. Daily Field excursions. Pe T
The laboratory is at Tolland, olo., in

hia | the sedi Mountains at an altitude

T T and expenses moderate.

THE
AMERICAN NATURALIST

Vout. XLVI August, 1912 No. 548

A CASE OF POLYMORPHISM IN ASPLANCHNA
SIMULATING A MUTATION

PROFESSOR J. H. POWERS, Px.D.

UNIVERSITY OF NEBRASKA

Durtne the fall of 1909 the writer accidentally dis
covered a case of heterogenesis in the genus Asplanchna
which bore all the earmarks of a bona fide mutation. The
two forms concerned were quite sufficiently different to
be classed as distinct species, and even as strongly
marked species. The transition was sudden and com-
plete, without apparent intergradation. The transition
was also in one direction only, and it could not be con-
sidered as in any sense due to an immediate cross, be-
cause the reproduction of the Asplanchna, aside from
resting eggs, is wholly parthenogenetic.

Subsequent study of the same species, during the fol-
lowing spring, as it appeared in several different loca-
tions, supplemented by extensive experiments, showed
that the phenomena in question were not those of typical
mutation, but are rather to be classed as striking in-
stances of polymorphism. As, however, their interest
depends in large measure, for the writer at least, upon
the ease with which they may be mistaken for mutation, I
i will first describe the facts as they appeared in the
original collection of material.

This collection was made about October 1, in the
remnant of a vile pool on the alkali flat west of the city
441

442 THE AMERICAN NATURALIST [ Vou. XLVI

of Lincoln, Nebraska. This portion of the flat has for
years been used as a dump, having been filled in several
feet in depth mainly with compost. Two years previ-
ously a heavy summer flood had excavated a cavity, the
size of a village lot, to the original alkali bottom, and in
this cavity the pool remnant, to the depth of about one
foot, remained. The water was dark brown with alkali
and the essence of compost, but it suited Asplanchna
exactly, for while almost no other plankton was present
the Asplanchnas were so abundant as to almost touch one
another; the sweep of a yard with a 20-inch dip-net
brought up a double handful of the strained animals.

My original study was confined to material from this
single collection because, a day or two later, a heavy rain
flooded the pool, killing the Asplanchna to the last
individual,

The collection in question contained two distinct types
of Asplanchna. The dominant form, outnumbering the
other several hundred to one, was of a large humped
type, closely akin to Asplanchna amphora. I shall, how-
ever, postpone a detailed discussion of the species, as it
is intricate, and unfortunately involves some controversy,
while the interest in the phenomena which I wish to
describe is not specially dependent upon it. Reference
to Fig. 1 will show fairly well its general appearance, and
I need but add at present that it was of very large size,
frequently measuring 1,500» in length.

Sparsely distributed among the mass of these animals
was a mammoth rotifer of related yet very different type.
It was saccate, or, more truly, campanulate, in form. Its
robust body, without humps, somewhat exceeded in length
that of its more slender companion, while its enormous
ciliated wreath or corona was extended beyond all limits
that the writer ever expected to see in a rotifer. Its

transverse diameter frequently equaled the entire length -

of the animal, This massive giant swam about with a
typical though rather slow Asplanchna movement, vibrat-
ing or flapping first one portion and then another of the

No. 548] A CASE OF POLYMORPHISM 443

ciliated corona in a very conspicuous fashion. When it
met one of the other rotifers in a head-on collision it was
much less inclined to contract or turn a quick somersault
than was the companion type. As to the specific,
varietal, or other classification of this large campanulate
Asplanchna I could find nothing. I concluded that it
must be an as yet undescribed species. Its distinctness
was further borne out by the results of detailed study.

The characters hitherto most extensively used for
specific definition in the genus Asplanchna are size, body
form, type and development of nephridium, and the form
of the trophi. As to the nephridia of the two forms in
question, they -proved, indeed, sufficiently different.
Although built upon the same general type, their size and
the number of flame cells bespoke wide separation.
Rousselet, who is the first authority upon the classifica-
tion of rotifers, in his last article! discussing the species
of Asplanchna expresses his belief that the number of
flame cells is constant in each species, although he admits
that it is very difficult to make strictly accurate counts of
these delicate organs. He assigns to Asplanchna
amphora about 40 flame cells, and this is a much higher
number than has hitherto been found in any other species.
My preparations permitted much better counts than may
be made on the living animal, but still I am not certain
of exactness. I found, however, that even the humped
rotifers frequently bore about 50 flame cells on a side,
and in many instances I found 55. In the campanulate
type the number was astonishingly greater. Not only
was the nephridial tube bearing the tags longer, stretch-
ing from the knotted portion near the bladder in a wide
diagonal curved line to the lateral edge of the corona, but
it was very thickly studded throughout the greater por-
tion of its length with the flame-bearing tags. The
number was certainly variable in larger and smaller
Specimens, but it was frequently about 100, and in a
_ "On the Specific Characters of Asplanchna intermedia,’’ Hudson, Jour.
Queck. Mic. Club, Second Series, Vol. VIII, pp. 7-12.

444 THE AMERICAN NATURALIST [ Vou. XLVI

number of instances, by thoroughly conservative counts,
I reached the surprising number of 115. The nephridial
development in this giant rotifer is therefore in propor-
tion to its large size and massive tissues. The trophi
also proved sufficiently different and of a size hitherto
quite unknown in the genus. A glance at Figs. 3 and 4,
showing the trophi of the two types drawn to the same
scale, will show at once their great difference. A detailed
discussion of their minor features will show them to be
still further apart.

In general, then, to judge by the structure of the
females, the two types seemed very distinct, and several
weeks of painstaking study, including the examination of
several thousand individuals, left me with very little
doubt that I was dealing with what we ordinarily class as
distinct species.

It was while examining some of my first mounted slides
of the large Asplanchna that I came suddenly on the evi-
dence of apparent mutation. As is well known, the
young of Asplanchna are highly developed before birth,
and in the material which I was studying this chanced
to be true to an unusual degree. My mounted slides, too,
were perfectly transparent, giving views not only of all
the organs of the adult, but of every organ and almost
every histological detail in the structure of the unborn
young. Examining these young, I noticed at first noth-
ing peculiar; the young within the humped rotifers bore
humps even more marked than those of the adult, it being
one of the recognized characters of A. amphora that these
body extensions reach a maximum development in the
young at about the time of birth. The first young noted
in the campanulate type were also, in all respects, essen-
tially like the parents.

I was, therefore, astonished to suddenly hit upon 4
large campanulate rotifer containing an unborn rotifer,
not of its own, but of the other, the humped type, Fig. 2.
The word mutation was the first thing that framed itself
in the midst of my inarticulate consciousness. But of

No. 548] A CASE OF POLYMORPHISM 445

course I neither accepted at once the suggestion nor the
apparent facts before me. Mutation, although among
things proven, is among the things proven to be rare, and
is therefore to be accepted in any given case only after
fullest verification. Moreover, facts themselves are fre-
quently deceiving. Perhaps the young humped As-
planchna did not belong within the body of its seeming
parent. Perhaps it was some accidental inclusion forced
through some invisible aperture in the side of the larger
organism. In short, was it really in the uterus? Per-
haps it had been swallowed by the adult campanulate and
had managed afterward to tear its way through the
digestive tract, which latter organ had then healed and
allowed the prey to continue development in the body
cavity. |

While I was asking myself these questions and others,
I was also assiduously seeking for more examples, for if
the phenomena were regular more examples would cer-
tainly be found.

A number were indeed discovered, and every possible
explanation was eliminated except that they were, as they
appeared to be, young humped rotifers which were pro-
duced by normal parthenogenetic development within the
giant campanulate type. I was able to discern, for some
of them, that they were within the uterus, and in every
way in normal position. The only slight sign of ab-
normality was that they usually seemed to be a little
small for the parental organism producing them; they
filled less of the body cavity than did corresponding
young that were developed true to type. But upon
second thought this was but one more sign of the com-
pleteness of the heterogenesis. These humped young
were developing of exactly the size that they would have
developed within the body of a humped parent; within
the body of a campanulate they naturally left room, so to
Speak, to spare. The internal organs, too, of these
heterogenetic young bore out, in detail, the conclusion
Suggested by the external appearance. The lighter

446 THE AMERICAN NATURALIST [Vou. XLVI

musculature, the narrower ovary- or shell-gland, were
obviously characteristic of the smaller humped type.
Most definite of all, however, was the evidence of the
trophi. These are universally recognized as yielding
characters of specific rank. They are wholly developed
before birth, and they were very different in the two
types in question. I used, therefore, the utmost pains to
ascertain the nature of the trophi in these heterotypic
young. The result was without question; the trophi in
every case indicated a complete and sudden transition
from the campanulate to the smaller humped type. This
was true both in regard to size and in details of structure.

I concluded, then, after the study of every organ which
I could make out with clearness in the material at my
disposal, that these atypic young in the large campanu-
late Asplanchna represented, not an ordinary variation,
but the sudden production of one type by another and
quite distinct one.

But a number of questions suggested themselves at
this point. In what number were these mutants, if such
they really were, being produced? Was the production
of one type by the other occurring in both directions, or
only in the one in which I had chanced to find it? And,
again, could transition specimens or any signs of grada-
tion between the two types be found among the adults,
or individuals of any age, in the material at hand?

As to the first question, I examined carefully 270
mounted adults of the giant campanulate type. Nearly
all were in reproduction, but only 90 of them contained
young sufficiently advanced in development and suffi-
ciently well placed to allow a certain judgment as to their
type. Of the 90 unborn young nearly one fourth (22)
were indisputably of the humped type, while 68 were just
as certainly of the parental or campanulate type. Noth-
ing should be made here of the ratio of one fourth and
three fourths. The phenomena in question are not the
result of a cross, and, moreover, the determination hinges
upon several factors; for one thing, upon my caution,

No. 548] A CASE OF POLYMORPHISM 447

when dealing with the earlier embryos, in rejecting any
case which was not quite certain. In other words, had I
been working with live material and rearing the animals
to a proper stage, a considerably larger proportion than
one fourth would probably have turned out heterotypic.

To answer the second question—i. e., whether this
atypical reproduction were occurring in both directions
or in one direction only—I first made a preliminary
examination of several thousand individuals of the
humped form, a high percentage of which contained well-
developed embryos. I found no indication of any de-
parture from normal reproduction. I then examined
critically one thousand individuals, recording results in
each case. Of this one thousand individuals, 419 con-
tained young sufficiently developed and well placed to
admit of safe determination. All were plainly of the
parental type; nor did I find a single case in all this
material that even suggested strong variation toward the
larger humpless rotifer.

Transitional specimens, however, could, if present, be
more certainly found among adults, or at least among in-
dividuals after birth. This was my third query. Did
such exist? Throughout my entire examination I sought
for them, but with little success. At certain stages in
development each species approaches a little nearer to
the other species. Thus the young of the larger, cam-
panulate type are not only smaller, but have slightly
bulging sides, suggesting the possibility of humps.
Their corona, too, although very heavy and broad, is less
disproportionately so than in the fully grown adult. But
in no case was it difficult to determine the relationship of
such specimens; other characters placed them at once,
especially the enormous and differently formed trophi,
which are fully developed before birth. Very old adults
of the smaller humped type might possibly, at first glance,
be taken for variants toward the larger form, in that
their humps become smaller and less sharply marked off
from the general body wall. In no other character, how-

448 THE AMERICAN NATURALIST [Vou. XLVI

ever, do they approach the campanulate type. Putting
aside such stages, which, with a little practise, fall
naturally into place, the amount of further variation did
not prove to be great. The humps, both lateral and
ventral, on the smaller. type, did show a considerable
amount of variation in relative size. And in numerous
cases it was of interest to note that such variations were
plainly hereditary; i. e., an extra large female with, say,
relatively small humps, would contain a large young in-
dividual that was obviously of the same type.

But these minor variations were to be expected; as
well as their tendency toward transmission, but they had
no evident connection with the type of change of which I
was in search.

Of true transitional types I discovered but two possible
instances. These two individuals were quite alike, but
differed notably from any others of the thousands
studied. Each was, all in all, indisputably a humped
rotifer, but decidedly beyond the ordinary maximum size
and with disproportionately small humps. The corona
of each was much broader and its cells much heavier than
in any other humped individuals observed; the body
walls and musculature also were much heavier, taking the
deep stain shown otherwise only by the campanulate.
All of these characters approached more or less the
campanulate type. Most striking, however, were the so-
called ovaries. In each case they were of the heavy type
shown by the larger rotifer. The nephridia did not
admit of careful observation. The most crucial organ of
all, however, the trophi, agreed wholly with the humped
type in form, and, though large, did not exceed the
maximum size found in the ordinary type. Studying
these two individuals with care, I came at last to the con-
clusion that, although in a sense they might be called
transitional, they were after all only variants of the
humped type—variants pointing toward, but not actually
leading to, the production of the campanulate type.

All in all, then,.my search for signs of genetic rela-

No. 548] A CASE OF POLYMORPHISM 449

tionship between the two types seemed to leave the evi-
dence for mutation sharp and clear: the transition from
the giant humpless Asplanchna to the smaller humped
type was sudden and complete; it involved many char-
acters; it was being effected in many individuals simul-
taneously; it was a transition in one direction only; in-
termediate types and fluctuating varieties seemed, in all
the material studied, no more in evidence than was to be
expected in any two species of one genus. At the close
of this part of my study it seemed that I had found an
ideal case of mutation, agreeing in every particular with
the definitions of the concept put forward by its origi-
nator, DeVries.

However, just as I was finishing the study of the pre-
served material at hand, and deeming that I had reached
the above conclusions with full certainty, I was suddenly.
halted by coming into possession of more living material,
and material so surprising that my well-worked-out re-
sults again assumed the character of problems, and I
resolved, if possible, to subject the whole to the test of
experimental method and further observation.

It was about the middle of March, 1910, that the
Asplanchna began hatching in a large covered aquarium
jar in which I had placed some of the resting eggs the fall
before. Between the last of March and the 10th of
April the species also began to appear gradually in two
ponds, in which, fortunately, the available food organisms
for many weeks were wholly different. Observations
were thus begun upon the species under diverse condi-
tions, even from the start. And, as they were continued
with but few and short interruptions through the spring,
summer, and well into the succeeding fall,? the variety of
conditions under which the development of the species
was followed included very wide extremes. Thus, during
a cold period in early April the species multiplied in
Shallow ponds upon which ice was forming nearly every

*Since the above was written nearly all of the eo recorded have
again been observed during a second season.

450 THE AMERICAN NATURALIST [Vou. XLVI

night, while heavy prevailing north winds kept the water
in continual turmoil. On the other hand, in July and
August, I was able to study the species in both permanent
and temporary ponds during periods of the hottest
summer weather.

Side by side with these observations, and correlated
with them as far as possible, a large number of culture
experiments were carried out.

The general method was as follows: The ponds were
visited several times a week, and such observations as
possible made on the spot. Then large quantities of the
animals, together with their food organisms, were
brought to a basement laboratory and placed in large
aquarium jars. These mass collections were sometimes
obtained by merely dipping up the water and the organ-
isms it contained in their natural degree of dispersion; at
other times the organisms were much concentrated by the
addition of large numbers taken with suitable nets. Flew
results were more interesting than the observation of the
resulting phenomena in parallel aquaria in which the
animals had been placed, now in small numbers, and
again in various degrees of concentration bus to almost
actual contact.

The material in these mass cultures was critically ex-
amined and followed from day to day. Then, when-
ever interesting phenomena seemed to be occurring,
either in the ponds or the mass cultures, or in both, iso-
lation experiments were begun with single individuals,
or with a definite number carefully selected and exam-
ined one by one.

To my satisfaction, the animals proved remarkably
adapted to experimental treatment. They withstand a
great variety of environmental influences, providing only
that two conditions are rigidly met. In the first place,
a copious food supply must be furnished; and second,
the fluid medium (it may be saturated with such ingredi-
ents that one hardly speaks of it as.water) in which the
animals have developed must be left practically un-

" No. 548] A CASE OF POLYMORPHISM 451

changed. This latter condition limits the range of pos-
sible experiment, but I could not once succeed in getting
Asplanchna to live and reproduce in any other than its
native fluid ;? even dilution of one fourth with water from
some other source seemed in every case to preclude a suc- `
cessful culture. In nature, too, every rain which diluted
to an appreciable extent the vile ponds in which this
Asplanchna flourishes results in the death of the entire
stock, which is only replaced by the hatching of new in-
dividuals from resting eggs.

As to food, many organisms were tried, and a number
proved available. Best of all, for the smaller forms of
the Asplanchna, Paramecium proves an ideal food.
Other rotifers come next, of which I used especially Hy-
datina and Brachionus; while, surprising as it would
appear, the crustacean, Moina paradoxa, although a
large-sized member of its group, provides, under certain
circumstances, an available food supply for this over-
grown and voracious rotifer. The adult Moinas are not
eaten, but young and even half-grown individuals may
be regularly eaten by the mammoth campanulate form I
have described. The humped rotifers ingest the crusta-
cea with difficulty, but succeed in doing so at times to a
considerable extent. j

Omitting further reference to methods, and refraining
from any attempt to describe the course of experiments,
I will briefly give their results, prefacing only that I make
no general statement which is not based on the outcome of
at least ten separate experiments or as ample observa-
tions in nature, or on both.

The first of these results which I will state, although
nearly the last that I demonstrated, is the error of my
original conclusion in regard to the relationship betwen
the two types already described. The campanulate and
humped types do reproduce each other reciprocally, al-
though with very different frequency in the two direc-

* This does not apply to cultures started from sepen eggs. These may be
hatched and the young developed in varied m

452 THE AMERICAN NATURALIST ~~ [Vou. XLVI

tions. Moreover, transition forms (in every feature ex-
cept the trophi) do oceur between them under given
though unusual conditions. The phenomena thus fall
more naturally under the rubric of dimorphism than
under that of mutation; although, as will be apparent
later, the phenomena with this rotifer bring the two con-
cepts into closer relation than they have perhaps ever
been brought before.

The second general result of work with the living ani-
mals was the surprising discovery that the species is not
only markedly dimorphic, but trimorphic, possessing a
third form—a smaller saccate type which differs from
both the humped and the campanulate types quite as
much as these differ from each other, and which, curi-
ously enough, differs far more, in external appearance
at least, from these other forms of its own species, than
it does from an allied but distinct species, Asplanchna
brightwellt,

I shall return to this latter point of specific distinctness
again. But by way of description I will here state that
this sacecate type, as I shall call it, although showing a
considerable range of fluctuating variability, is a rather
primitive but typical Asplanchna. In size it averages
much smaller than either of the other types dealt with,
reproductive individuals ranging from about 500. to
1,200 in length. In outward form it resembles closely,
now A. brightwelli, now A. priodonta. The corona, when
the animal is seen from the end, is always circular, as 1s
typical for the genus, never showing the strong dorso-
ventral flattening which is the invariable condition in
the humped and campanulate types. Correlated with
this cireular corona is the general cylindrical shape of the
animal. When placed on a slide in a shallow drop of
water these saccates characteristically rest on their sides
instead of their dorsal or ventral surfaces as do both of
the larger types. The trophi are of the same type as are
those of the humped Asplanchna, except that they are a
little smaller and proportionally a little more robust.

No. 548] A CASE OF POLYMORPHISM 453

The internal organs are of the same general character,
although more compact and filling more of the body space
than in the two larger forms, two of them, however, be-
ing notably different in development.

First, the nephridia. I have not studied these in
stained preparations as I have in the other types, but,
using living material, I make out a decidedly variable
number of flame cells; they range from as low as twenty
to nearly forty. It is at least plain that the number
averages much less than in the humped form, and but a
fraction of that in the: campanulate. The difference
would seem to be plainly correlated with the general dif-
ferences in the size of the organisms.

Singularly in contrast to the smaller sis pitidte is the
development of the contractile vesicle or bladder. This
is a small organ in both the humped and campanulate
types, but in the much’ more diminutive saccates it be-
comes very much larger, filling, when expanded, a large
part of the body cavity. I have examined considerable
numbers from diverse sources and reared under different
conditions, to make sure that this contrast was not acci-
dental. I find no exceptions to it. Like the number of
flame cells, this matter of size of the contractile organ,
relative to that of the body, has been hitherto considered
a specific difference.

A final peculiarity of the saceate type, although a var-
iable one, is its tendency toward an excessively rapid
rate of parthenogenetic reproduction; it is usually
crowded with embryos. Four, five and six are frequent ;
sometimes I have counted nine; while in many individuals
they are so closely packed that counting is out of the
question. These numerous young are also frequently
born at a disproportionately small size, even compared
with the diminutive parent, thus still farther increasing
the extremes of size which the species presents. It seems
probable that this smaller and less developed type is
itself, to some extent, a product of this rapid rate of
multiplication. I have found it multiplying at a slower

454 THE AMERICAN NATURALIST [Vou. XLVI

rate—i. e., producing but one or two embryos; but such
instances were always due to degeneration or growth un-
der unfavorable conditions. I may add by way of con-
trast that the humped form most typically produces but
one young at atime. Thus, in the material of my orig-
inal study, among one thousand individuals I found but
sixteen showing the more or less simultaneous develop-
ment of two embryos, and none showed more than two.
The campanulates regularly develop several embryos at
once, four and five being common numbers; but in their
capacious bodies there is no overcrowding, and certainly
no curtailment of nutrition.

As to the occurrence, the definite relationship, and
the causes of the production of the three types of the
species, I have ascertained most of the facts, though
not all.

The saccate form is, I think, unquestionably the only
form that emerges from the resting egg. In all cases—
seven in number—in which I have examined temporary
ponds within a few days after formation and found this
rotifer just appearing, or in which I have caught the spe-
cies evidently close to its first appearance, I have found
either nothing but the saccate type, or numerous saccates,
together with transitional and humped types. In the
majority of these cases the saccates have been rapidly
displaced by the humped type, which, in nature at least,
very rarely reverts to the production of the smaller form.
As soon as the species is in full swing, so to speak, count-
less numbers may be examined for weeks without a sac-
cate occurring. All of these phenomena have been par-
alleled in my aquaria, and I have also repeatedly observed
the birth of the humped type from the saccate parent.
This is usually in instances where the high numbers of de-
veloping embryos in the saccate parent have given place
to few or to but one. Usually, too, the parent is a rather
large saccate and the humped young a rather small ex-
ample of its type. But occasional instances have been
noted where the humped young, a moment after birth or

No. 548] A CASE OF POLYMORPHISM 455

as soon as it had assumed complete expansion, appeared
fully a third larger than its saccate parent. The act of
giving birth to such heterogenetic issue is a decidedly
prolonged and painful process for a rotifer, very differ-
ent from the ordinary, sudden expulsion, well known in
Asplanchna. .

Turning to the consideration of the actual cause of the
transition from the saccate to the larger humped form,
I have not discovered it definitely. The cessation of
the rapid rate of reproduction, already spoken of, seems
a partial proximate cause, though it itself I can not ex-
plain. It seems in part a mere matter of the number
of generations following hatching from the resting egg.
But a copious food supply and generally favorable con-
ditions probably favor the change. In early spring the
saccate form maintained itself for several weeks in a
pond where, in July, after drying up and being refilled
by rain, the saccate form began to produce the humped
form sparingly within four days, and was very soon sup-
planted by it. In another pond of purer water and seem-
ingly less well adapted to the species, the saccate form
appeared scantily in early May, and struggled on spar-
ingly for three weeks, feeding mainly upon the flocculent
masses of blue-green alga, but somewhat upon Brach-
ionus; it then disappeared entirely, having produced no
form but its own so far as I could discover. Yet in isola-
tion cultures from this same stock, when fed on Para-
mecium, I raised, after several generations, a number of
rather small humped Asplanchna.

While I have seen little evidence, thus far, that either
of the larger types give rise to the saccate form in nature,
yet in two small cultures I have produced a population of
from one hundred to several hundred small saccates by
the degeneration of the humped stock. In one case the
colony was a very old one and the degeneration possibly
due to this fact; in the other it was plainly the result of
diluting the culture medium with tap water. The small
saccates, even when they had dwindled to a half dozen,

456 THE AMERICAN NATURALIST [Vou XLVI

before total lapse, still bore developing embryos in their
diminutive bodies.

The typical case of relationship, however, is, plainly,
for a few generations of rapidly developing saccates to
be succeeded, after a few transitional forms, by a con-
tinuous and ever-increasing population of the humped
type.

As to the occurrence of the humped type, this is, prac-
tically, as just stated. Although never hatching from the
resting egg, it soon supplants the smaller and earlier
form and then continues to multiply, in favorable local-
ities, until put an end to by drought, rain, or the exhaus-
tion of the food supply. I do not believe that it can mul-
tiply indefinitely, however, for in even my largest aquar-
ium jars, the species invariably seems to die out after
several crescendos and decrescendos of development.
The largest cultures have lasted about five months. That
their death is not due to the accumulation of metabolic
products in the unchanged water seems probable, because
in one case a new generation hatched from resting eggs
soon after the death of the first, and showed no signs of
weakness, although growing in the same long-used
medium. |

The occurrence of the giant campanulate form is a to-
tally different matter from that of the other two types.
It never occurs alone. It never occurs in newly hatched
cultures or young stocks. It rarely occurs in great
numbers, and never does it do so, so far as I have ascer-
tained, in nature when feeding upon what may be called a
normal food. On the other hand, in old stocks that have

* This holds good under all ordinary conditions. Since the above was
written, the experiment has been made of hatching out hundreds or thousands
of resting eggs in a very small space—e. g., a three-oz. bottle, i in almost
pure tap water. The results of this experiment are very striking. aced
thus from the start almost without nutrition, a few of the more vigorous
animals begin cannibalism at once, with the result of a an early appearance
of both the humped and the campanulate types. In just what generation
they thus appear is not known, but very probably in sa second and thir
respectively, The development and saltation oceur with but few individuals,
the majority starving to death or serving for food only.

No. 548] A CASE OF POLYMORPHISM 457

become numerous, I find this giant type invariably pres-
ent. The original collection which I studied was about
typical in this respect. I have not ascertained the exact
proportion which the campanulates bear to the humped
form in any given case, for experiment shows that it is
capable of variation within wide limits; but I think that
in nature they would rarely equal one to one hundred.®

What was the relation of these two types? Weeks of
all but continuous work were necessary to decipher it.
The fact that the two types were so frequently associated
indicated probable connection, but did not demonstrate
it; the fact that the campanulate always appeared only
long after the development of the humped form was again
suspicious, but proved nothing. Isolation cultures of the
campanulates soon showed, as my mounted material had
done, that they regularly produced not only their own
type, but a considerable proportion of the humped type as
well. But the reverse process is at least so rare that I
have been as unable to find it among my living material, as
I was among that which I had mounted. Scores of isola-
tion cultures, begun with few or single individuals, fed
on diverse types of food, and nearly all very successful,
remained negative in result, with the exception of one
of my very first attempts. I was trying out the rotifer,
Hydatina, as a possible food for the humped Asplanchna,
and placed half a dozen of the ordinary type in a watch-
glass with many of the smaller rotifers. Soon, among
the young produced, there appeared one that rapidly
developed into the giant type; but subsequent trials, car-
ried on as long as this special food supply lasted, were
negative in results.

Over against these failures, however, a large number
of my mass cultures yielded at once all but conclusive
evidence that the larger type was somehow derived from
the smaller. These cultures were started as follows:

*During the recent summer one pond in which the species developed
copiously from spring to August showed at the latter period a much higher
number—about one campanulate to 20 humped.

458 THE AMERICAN NATURALIST [Vou. XLVI

From the surface of a pond several feet in depth, and at
a point several feet from the shore, a large quantity of
the humped type were collected with a small net. This
material and the entire pond, so far as my investigations
had gone at this time, had shown none of the larger form,
and the care with which the material was taken from the
surface rendered it improbable that a resting egg of the
larger form should contaminate it, the resting eggs of all
these rotifers being heavy and sinking at once.° This
material was placed in clean transparent jars of several
liters, although beside each larger culture was usually
placed a smaller one in a tall thin vessel which permitted
careful scrutiny of the entire contents with a low-power
lens. It was almost impossible to crowd the animals to
the point of injury, though perhaps every other visible
organism would die. Now in all of these crowded mass-
cultures, whether large or small, whether fed, half-fed,
or starved, the campanulate form made its appearance
within at most a week. Sometimes a single individual,
either young or fully developed, would first be discov-
ered; more frequently a number would be present before
noticed. But in every instance the sudden appearance of
the large form was followed by its very rapid multiplica-
tion, coincident with a still more rapid diminution in
numbers of the humped form. The latter were eaten up
by the former, even adults of the humped type falling
victims to the prodigious ingesting power of the campan-
ulates. This all but complete displacement, in the
course, say, of three weeks, of one form by the other,
was a striking and almost astonishing spectacle. I have
not observed it in nature, however crowded the species
becomes in its natural situation. But in my culture —
dishes it was the regular occurrence in case the individ-
uals became sufficiently numerous.

*It has later been ascertained that, F certain circumstances, some of
the resting ova may again rise and fl at; in a windy pool, however, they

would very soon be blown to the esate and age Yet the foree of the
experiments cited is reduced.

.

No. 548] A CASE OF POLYMORPHISM 459

The suggestion of cannibalism as the cause was thus
all but conclusive; but I was able to press the proof one
step farther. My isolation cultures, started with one or
a few individuals, invariably failing, I decided to try
cultures started with a larger number of individuals,
each one of which was first subjected to examination un-
der the microscope. 240 individuals were thus isolated
and placed in a single stender dish, care being taken to
reject any that deviated never so little from the normal
humped type, especially as to extra size. In the rapidly
multiplying culture thus started there appeared within
a week several typical campanulates, and the whole sub-
sequent course of the culture duplicated the develop-
ments which had taken place in the mass cultures already
described. This is as near as I have come to actually
observing the production, by the humped Asplanchna,
of its larger humpless congener. I have not witnessed
the birth of the one from the other or seen it in uteri.
The demonstration of fact is, however, sufficiently com-
plete, and the farther conclusion is sufficiently plain, viz.,
that the transition in this direction between the two types
is a relatively rare one. Not every humped Asplanchna
possesses the power, whether this depends upon size,
ingestive reaction or digestive capacity, to produce the
larger form. Indeed, this power would seem to reside in
but very few individuals.

As soon as I had reached even the tentative conclu-
sion that cannibalism must be the cause of this marked
heterogenesis, I set about to determine, if possible, why
this should be the case. Was there some magic in the
cannibalistic diet as such? Was it merely a case of rich
nutrition? Or was the result perhaps due to the mere
size of the ingested food organisms?

Excessive feeding with most of the usual food organ-
isms of the species was plainly without result. So long
as the Asplanchnas were relatively few or greatly out-
numbered by their food organisms they glutted them-
selves to repletion without unusual consequences in re-

460 THE AMERICAN NATURALIST [ Vou. XLVI

production. Thus in the case of one pond, an algal-
feeding Brachionus developed until it exceeded in num-
bers anything that I had ever before observed under sim-
ilar conditions. The surface water was so filled with
them that one observer pronounced it as thick as good
tomato soup. In this medium the humped Asplanchna
gorged itself during days of rapid multiplication. Every
stomach was packed with the smaller rotifers; yet no
change in type resulted. Finally the Asplanchnas liter-
ally ate up the Brachionus, but themselves disappeared
a short time after this was accomplished without having
produced any of the giants. Yet condensed material
taken from this same source and placed in large culture
dishes did produce the cannibals some time after the
Brachionus had been devoured. It follows from this in-
stance, as well as from a number of others equally con-
clusive, that mere feeding on the flesh of rotifers, so to
speak, is not sufficient to cause the change.

The same proved true of over-feeding on Paramecium
and several other protozoa that were greedily eaten.

I was the more surprised, therefore, to discover one
other food organism, and one only, so far as my observa-
tions extend, which is fully competent to bring about the
same result. It is the Daphnid-like crustacean, Moina
paradoxa. It seems almost incredible that a rotifer
should feed upon this robust entimostracan, even the
young of which are born of a size which would seem be-
yond the utmost stretch of a rotiferean stomach. More-
over, this Asplanchna does not by any means invariably
attack Moina. It may disappear, seemingly from starva-
tion, in a pond where Moina is present and rapidly mul-
tiplying. This has also happened repeatedly in my cul-
tures, and I have placed numbers of the young Moinas in
watch-glasses together with the Asplanchna without hav-
ing them eaten. Yet in other very similar cases the
young Moinas are ingested by the Asplanchnas—by the
humped form, and rarely by the larger saceates. The
very youngest Moinas may even at times be ingested by

No. 548 | A CASE OF POLYMORPHISM 461

the smaller saccates. It seems probable that the vigor
of the Asplanchna stock, and, I think, the period in the
reproductive cycle (number of generations since the rest-
ing egg), have much to do with the refusal or attack of
this oversized food organism. Be this as it may, how-
ever, the habit of eating the young Moinas, once started,
seems to rapidly become fixed in not only the individual,
but its progeny and the whole subsequent stock; although,
weeks later, perhaps, this feeding reaction may again be
lost, and the last rotifers starve to death in the presence
of abundance of young Moinas. Such starving stocks
may be again revived, when almost perished, by the sup-
ply of a smaller food organism.

Now the feeding upon this crustacean has, to all ap-
pearance, the same effect in bringing about the produc-
tion of the large campanulate type as has cannibalism. I
have not proven it with the same precision, but I have
observed the transformation of mass cultures from the
smaller to the larger type under circumstances that sug-
gested the all but certain conclusion that Moina-feeding
was the initial cause. I have also reared large numbers
of the campanulates, for weeks at a time, by using the
young Moinas as food. Under these circumstances the
humped rotifers all but disappear from the cultures. A
few always remain, because they are continually being
produced, in minor numbers, by the campanulate type.
But as very few of the humped young are able to acquire
the habit of Moina-feeding, they starve to death without
reproduction or fall victims to their mammoth progen-
itors.

In one instance, and one instance only, I have followed
the transformation of the Asplanchna to its largest type,
through Moina-feeding, in nature. The case interested
me much, being very different from any other instance
that I have followed, and showing to an astonishing de-
gree the protean possibilities of the species. As usual,
the first appearance of the species was in this instance
by scattered individuals of the saccate form. Their food

462 THE AMERICAN NATURALIST [ Vou. XLVI

consisted chiefly of Brachionus and Hydatina. These
organisms were not numerous, however, and were ex-
hausted just as the countless hordes of the young Moinas
appeared. Under these circumstances a wholesale tran-
sition to Moina-feeding occurred, even while the As-
planchna was in its earlier phases of development. The
smaller humped forms had just begun to appear when
the Moina-feeding began, both by these humped individ-
uals and by not a few of the larger saccates. The species
became for a few days indescribably chaotic. So far as
body form was concerned, transition stages could be
found between every possible type and variation. This
is the only instance in which I have found the saccate
form giving rise to the campanulate without the humped
intermediate, but it certainly did so to a considerable ~
extent in this case. The outcome of Moina-feeding in
this pond was the establishment, after about nine days,
of a new equilibrium for the species. The giant campan-
ulates had become in this instance, and this instance only,
among my observations in nature, the preponderant type,
reproducing their own form in the main, but also a few
slender, long-humped individuals. Of these latter a very
few managed to ingest the Moinas and to reproduce, while
the majority showed empty, shrunken stomachs and no
developing embryos. All traces of saccate and transi-
tional forms had disappeared.

I will note, in passing, that the campanulate form which
is produced by the Moina-feeding is not quite identical
with that procured by cannibalism. The size reached is
even a little larger, and the animals have the appearance
of being even more powerful in general musculature;
but the flaring corona never becomes quite so extreme;
the animals never assume quite the bell-like form which
the most majestic cannibals present.

(To be concluded)

HARDINESS IN SUCCESSIVE ALFALFA
GENERATIONS

L. R. WALDRON

DICKINSON, NortH DAKOTA

In 1908 Mr. Charles J. Brand, of the Department of
Agriculture, inaugurated an experiment in alfalfa to de-
termine, among other things, the relations that different
strains of alfalfa have to the cold of severe winters. The
writer aided in this investigation..

For this purpose 68 regional strains of alfalfa were
assembled from the various alfalfa regions of the world,
many of them being foreign in immediate origin. The
alfalfas were planted in hill and drill rows, and during
the season of 1908 the rows were thinned so that accurate
countings could be made. There were in the neighbor-
hood of 80 plants to each strain. The winter of 1908-09
was particularly severe to alfalfa, and as a consequence
most of the strains were sadly deleted. :

Twelve of the 68 strains were entirely killed, no living
plants remaining. Twenty-eight of the strains killed out
over 90 per cent., and over 60 per cent. of the strains
killed out over-80 per cent. There were but 3 of the 68
strains that killed out less than 10 per cent. The killing
of the American alfalfas was severe as indicated by the
fact that the 9 strains from Utah killed over 90 per cent.,
while the 3 Montana strains killed over 65 per cent.

The hardier strains were those of more recent foreign
origin. Two strains of the Grimm alfalfa had an aver-
age killing of less than 5 per cent., thus being the hardiest
im the nursery, N eglecting the 12 strains that killed out
entirely, the average killing of the nursery amounted to
17.51 per cent., using each strain as a unit.

' Charles J. Brand and L, R. Waldron, ‘‘Cold Resistance of Alfalfa and
Some Factors Influencing It,’’ Bulletin. 185, Bureau of Plant Industry,

- S. Department of Agriculture.
463

464 THE AMERICAN NATURALIST [Vou. XLVI

During the summer of 1909 the seed produced by the
living plants was saved from each strain separately. In
the spring of 1910 a sowing was made of the original seed
that had sown the first nursery, which had been desig-
nated as nursery A. This sowing was known as
series 201. In addition another sowing was made in
1910, known as series 202, from the seed collected
from nursery 4 in 1909. This sowing was from seed se-
cured from plants that had survived the severe winter of
1908-09. In addition a number of rows were sown with
seed from nursery A plants, selfed during the summer of
1909.

This second experiment, carried on by the writer, con-
sisted of a number of duplicate rows seeded (a) with seed
from original geographical sources, and (b) rows seeded
with daughter seed obtained from the plants surviving
the winter of 1908-09. These two seedings comprised
112 150-foot rows, each row containing up to 150 plants.
During the summer of 1910 the rows were thinned and
accurate countings were made. In the spring of 1911,
after growth was well started, a determination was made
of the number of dead plants, and in addition the living
plants were gauged as to their vigor, on a basis of 1 to 10,
the best plants receiving the highest markings.

The data obtained indicated the apparent increase of
hardiness among the different strains. A possible
source of error was the effect of vicinism in nursery 4
in producing hardiness in the progeny plants constitut-
ing series 202. Limited space prevents presenting the
evidence which would lead one to think that this error
was slight in effect and probably nearly negligible.

Only a brief summary of the results can be presented
in this article, a detailed account of which must be left
to a future publication. In the first place, let us regard
the comparative results of the winters of 1908-09 and
1910-11 upon the strains in nursery A, and upon those
of series 201. These two sowings were from the same

lot of seed, and had they been representative samples one 2 —

No. 548] HARDINESS IN ALFALFA 465

would have expected that they would have fared rela-
tively the same in the two winters. Deviations that
might appear could be charged to the differences of the
two winters, as we may suppose that alfalfa becomes ac-
climated in different ways to different types of cold.

The relation of the killing of the plants in nursery A
and in series 201 is presented in the correlation table of

x

Means of Duplicate Strains, 1910—11
11 16 21 26 31 36 41 46 51 56 61 66 71
15 20 25 30 35 40 45 50 55 60 65 70 75

Means of Original Strains, 1908—09.
i

-=
(S)
bo
_
j
p=
—

2
a 1 Lt

or

p

ou

pi

_
OOrRONWD RN ENR RE RR ORF CONF

O45 0° 8h SAS 2 Ss 2 a Be
Fig. 1. Correlation of the means of various strains of alfalfa. Original
nursery A seeding subject, duplicate of thes strains in 201 series relative.
Coefficient of correlation, + 0.62 +0.06. Standard deviation ‘of Y, 24.85 + 1.57.
Standard deviation of X, 19.62 + 1.24.

Fig. 1. The means of nursery A are the average per
cents. of killing for the different strains, and the means
of the duplicate seeding, series 201, are the correspond-
ing per cents. for that nursery. In the table, the nursery
A means appear as subject, and the means of series 201
appear as relative. The table shows good correlation
amounting to + 0.62 + 0.06, thus indicating that the two
series of alfalfa were affected relatively the same by the
two winters. This fact is perhaps more strikingly shown
when the two series of means are platted on coordinate

466 THE AMERICAN NATURALIST [Vou. XLVI

paper. The ups and downs of one series correspond
very closely with the ups and downs of the other series.
The mean killing of series 201 was 27.43 + 1.75 per cent.
as against 77.51 + 2.21 per cent. in nursery A, indicating
that the second winter was the milder one.

The relation between the killing experienced by series
201 and series 202, during the 1910-11 winter, is ex-
pressed by the correlation table of Fig. 2. In this case
the 201 series is subject and the 202 series is relative.
The mean of the 202 series amounts to 6.43 + 0.66. The
correlation in the table amounts to + 0.46 += 0.07. As

x
Means of 202 Series
ELCIO (230-96

0 31

5 10 15 20 25 30 365
O° S16 ¢ 7 9
6 1018 1 4
1558.6 9
$16 20| 2 1- i 4
B21 25| 2 13 5
226 30| 2 3
Sal 35|3 J 4
> S 36 40 1 1-1 3
© 41 45/1 1 2
646 50|1 1 1 114
è 51 55 1: í 1 3
= 56 60 1 1 2
61 Gi i 1 3
66 70 0
71 75 1 1

24 7 1 1 56

Fig. 2. Correlation of the means of various strains of alfalfa. Duplicate
seeding of nursery A (series 201) subject, progeny of the hardy plants of nur-
sery A (series 202) relative. Coefficient of correlation, + 0.46 = 0:07. Standard
deviation of Y, 19.62+1.24. Standard deviation of X, 7.35 = 0.46.

indicated by the means of the two series, 27.43 and 6.43,
there was a remarkable apparent increase in hardiness.
This is partially expressed in the correlation table, but
in the table the most pronounced changes do not work
for a strong correlation.

For instance, one Utah strain killed out 59.6 per cent.
in the 201 series, while its offspring killed but 6.2 per
cent. in the 202 series. This indicates a weak correlation
but a great increase in hardiness. There were but 3
instances of the 202 series killing more severely than the

No. 548] HARDINESS IN ALFALFA 467

901 series. Two of these instances were Turkestan al-
falfas. The third one was due to the fact that the 202
row was an outer row, thus not afforded protection by
an adjacent row.

In our theoretical discussion of the data we can
scarcely more than present the problem. Within the
limits of the pure Medicago sativa, as pure as it exists
to-day, the fact is patent that there is a wide range of
diversity in hardiness among the different strains of
alfalfa, dependent in greatest measure upon their geo-
graphical origin. Strains of Medicago sativa that have
been grown for long periods in cold climates, e. g., Mon-
golian alfalfas, are found to be hardy during the severe
winters of this country. Upon the other hand the strains
of this species which have been grown for long periods in
hot countries, e. g., Arabia and Peru, were found to be
exceedingly tender in cold districts. Thus at Dickinson,
North Dakota, it has never been possible to bring alive
through the winter a single plant of the Arabian alfalfas,
and only under exceptionally favorable conditions has it
been possible to winter any of the Peruvian alfalfa
plants.

The diversity is so great between the Arabian and the
Mongolian alfalfas that we must consider that hardiness
has actually been added to the alfalfas which have become *
hardy, or that hardiness has been lost to alfalfas that
have become tender.2 It is not unlikely that changes
have been brought about in both directions. The simple
problem is, has this change come about through the
“law of ancestral inheritance,” or must the change be
accounted for by distinct mutations occurring within any
particular strain of alfalfa? Or is it possible that
changes have come about through conformity to both
methods?

*The fact is not lost sight of that an increase of hardiness may possibly
be brought about by a recombination of certain (at present unknown)
morphological characters physically responsible in different ways for the
‘Presumably complex character of hardiness. See Nilsson-Ehle, ‘*Kreuzungs-
untersuchungen,’? Lunds Univ. Ars., Bd. 5, Nr. 2, p. 114. Also a forth-

coming article by the writer in the American Breeders’ Magazine.

468 THE AMERICAN NATURALIST [ Vou. XLVI

The belief that any strain of alfalfa is made up of
many substrains with various steps of hardiness has
been arrived at largely by a priori methods. We find
that there are distinct morphological types of alfalfa
within a strain which breed true, and that with other
plants there are physiological types within any strain or
variety likewise breeding true. It is reasonable to sup-
pose that the same is true in regard to hardiness in al-
falfa.

One of the Utah strains of the 201 series killed 42.8 per
cent. from a total of 76 plants, and at the same time this
strain of the 202 series sown from seed secured from
three mother plants killed but 3.5 per cent. from a total
of 131 plants. We have established at once a compara-
tively hardy alfalfa from one quite tender. With this
Utah alfalfa, as with several others, it is difficult to avoid
the conclusions that the strain is made up of many bio-
types, relative to hardiness, which show their indepen-
dent character even when no precaution is taken against
interbreeding

We have also experimental evidence upon this point.
Alfalfas that were selfed in 1909 were from both hardy
and tender strains of alfalfa. A selfed Mexican plant
had progeny that showed absolute hardiness during the
winter of 1910-11, while the mother strain killed 24.5
per cent. Others of the selfed alfalfas acted in the same
manner, producing offspring behaving radically differ-
ent from the behavior of the parent strain. In some
cases progenitors of non-hardy strains were selected from
hardy strains.

It seems likely then that any regional strain of alfalfa
as far as hardiness is concerned, is made up of biotypes
with different cold resistant qualities. An alfalfa of
this nature when moved to a colder region loses the rep-
resentatives of the tender biotypes leaving the hardy
ones for propagation. But this explanation accounts in
no measure for the absolute changes of hardiness which
some alfalfas must have undergone to have allowed the

No. 548] HARDINESS IN ALFALFA 469

species to accommodate itself to extreme climatic con-
ditions.

While it is evident that there are alfalfas in existence
to-day which do not contain elements of hardiness suffi-
cient to allow them to live through the severer winters
of the United States, this statement apparently does not
hold for the alfalfas that are grown in the New World,
with the exception perhaps of the alfalfa known as the
Peruvian.’

Inasmuch as the South American alfalfas grown in
hot regions contain elements capable of surviving very
severe winters, it is apparently not necessary to assume
recent mutative periods in the species. Inasmuch as
the alfalfa plant grown in the New World contains ele-
ments of hardiness allowing it to persist through periods
of severe cold, it is reasonable to assume that this ele-
ment of extreme hardiness may be dissociated from the
elements of lesser hardiness, even in the tenderest
strains, in such a manner that it can be carried from
generation to generation. This practical result has not
yet been attained, not because such a result is theoret-
ically impossible, but because no systematic attempt has
been made. It is evidently a saner thing to commence
practical breeding with those forms nearest to the one
desired.

* Charles J. Brand, ‘‘ Peruvian Alfalfa,’’ Bulletin 118, Bureau of Plant
Industry, U. S. Department of Agriculture. The Peruvian alfalfa is con-
sidered varietally distinct. This and the Arabian alfalfa have not as yet
been found to possess elements capable of amelioration to the»severe con-
ditions of the north. Whether these two forms are retrograde mutants
derived from the old alfalfa stock is pasillo to determine at this time.
One would be inclined to believe that such is the case.

The historically and long-continued growth of. alfalfa in hot régions
almost precludes the hypothesis of the loss of cold-resisting biotypes, which
would obtain were the alfalfa strain transferred to a hotter region.

SOME FEATURES OF ORNAMENTATION IN
FRESH-WATER FISHES

HENRY W. FOWLER

THE ACADEMY OF NATURAL SCIENCES OF PHILADELPHIA.

Tue fishes known as the minnows (Cyprinide) and
suckers (Catostomide) are familiar types to almost
everybody. Certain of them during the spawning-season
exhibit peculiar dermal excrescences or tubercles, these
often rendering their appearance quite striking if con-
trasted with that of other seasons. These tubercles have
long been known in many forms, though they do not seem
to have been elaborately studied in connection with the
relationships of the various groups in which they appear.
Considerable difficulty may be encountered by those wish-
ing to make studies of them, not only as they are seasonal
and the fishes often difficult to secure at the required
time, but also as they crumble and rub or drop off in al-
coholic preparations, sometimes not even leaving the
usual scars. It is hardly possible to arrange the dis-
tribution of the tubercles into distinctive groups of sig-
nificance. Perhaps when enlarged and few in number on
the head they contrast strongly with other cases, such as
when very numerous and universally distributed. It
may also be added that in a great many of the fishes be-
longing to the families under consideration no tubercles
appear at any season, and this is often generic as well
as specific.

As now understood the minnows embrace the largest
family in ichthyology, having about 210 genera and 1,500
species in North America, Europe, Asia and Africa.
They are mostly fishes of great similarity, small in size,
weak, and among the most difficult in which to distinguish
species. In this respect our local forms are very fre-
quently no exception. Passing over the different genera

470

No. 548 |

ORNAMENTATION IN FISHES

EXPLANATION OF PLATE

The figures are merely outlines, showin

All are drawn with the accompaning

a Campostoma anomalum
g Chrosomus erythrogaster.
- Pim es promelas

. Not:
io otropis whipplii ponpe E RE,
ma Notropis cornutus.
Notropis chalybæus.

H
Oycleptus pin, beg

. Moxostoma macro

g the ac aon of the tubercles.

line representing

agra ce
Rhinichthys atronasus.
ybopsis kentuckiensis

472 THE AMERICAN NATURALIST [Vou. XLVI

and species from the Middle States region it has been
“my opportunity to study in the present connection, the
following conditions are found.

The stone roller (Campostoma anomalum) occurs in
the Mississippi Valley. In spring or early summer the
adult males sometimes have the entire upper surface of
the head and body covered with small tubercles, though
on the fins I have only found them extending on the rudi-
mentary dorsal rays.. Though mostly attended with higk
coloration, red being more or less conspicuous at most of
the bases of the fins, this is not always the case. Others
were found full of spawn, without tubercles, but in high
color.

The red-bellied dace (Chrosomus erythrogaster) is one
of the most brilliant of all our common freshwater fishes,
occurring mostly west of the Alleghanies. The adult
spawning male has the entire head and body almost
everywhere covered with minute tubercles in the greatest
profusion. No females with tubercles have been secured.

The fat-heads (Pimephales promelas and P. notatus)
are remarkable for their black heads, in the case of
spring males, besides having only a few large conspicuous
tubercles on the muzzle. The former species occurs only
west of the Alleghanies while the latter is found in all the
river-basins of Pennsylvania. The tuberculate males
seem to be the exception, or of short duration in that
condition, as among hundreds of examples of the latter
but few were found.

The fall fish (Semotilus bullaris) and creek chub (5S.
atromaculatus) are two well-known fishes. The former
occurs only east of the Alleghanies, and the fully adult
male is with brilliant rosy sides and mostly rosy fins.
Though reaching a length of nearly two feet, examples
three inches long have been taken with fully developed
eggs. The only tuberculated examples were all over a
foot in length, and had only their muzzles densely
covered with small tubercles. No nests were ever found
made by the small fish of three or four inches in length.

No. 548] ORNAMENTATION IN FISHES 473

The nests discovered were attended only by large fish.
The creek chub occurs in most all our mountain brooks,
and reaches a maximum length of ten inches, though
males five inches long have tubercles. These tubercles
are féw, but quite conspicuous, and occur usually as a
pair over each eye and one in front of the latter. Large
examples have the scales on the hind part of the back
thickened, or but slightly tubereulated. No tuberculated
females of either species were secured.

The true dace (Leuciscus elongatus and L. vandoi-
sulus) are represented by the former in streams west of
the Alleghanies and by the latter in streams east of the
Alleghanies. In both species the spring males have
crimson sides and the top of the head minutely, but incon-
spicuously, tuberculate. In the eastern species occasion.
ally a female will be found with tubercles on its head
above, and also similar brilliant coloration to that of the
male.

A few species of shiners (Notropis) present cases of
tuberculated males. The silver fin (N. whipplii analo-
stanus) of our eastern streams has the body and basal
Portions of the fin-rays largely covered with small
tubercles in the adult male. The adult female is occa-
sionally tuberculate. . The red fin (N. cornutus), found
in almost all of our streams, has the adult male very
gaudy in spring, its tubercles appearing very conspicu-
ously on the muzzle, front and predorsal region. Occa-
sionally a tuberculatefemaleisfound. Intheiron-colored:
shiner (N. chalybeus), often brilliant in the spring, the
males sometimes have the muzzle, front and predorsal
with many rather large tubercles. No females with
tubercles were found.

The male of the long-nosed dace (Rhinichthys cata-
racte) when fully adult has its snout, top of head, entire
back and rudimentary dorsal rays minutely tuberculate
in the spring, The adult male of the black-nosed dace
(R. atronasus), one of our most abundant fishes, has the
front and predorsal region minutely tuberculate in spring

474 THE AMERICAN NATURALIST [Vov. XLVI

and early summer. No tuberculate females of either
species have been found.

The crested chub (Hybopsis kentuckiensis) is remark-
able for the adipose-like frontal crest of the adult male in
the spring, somewhat suggestive of the hooded seal.
This species reaches ten inches in length, and examples
half that size were taken with tuberculate muzzles, though
without the crest.

The suckers are represented by 14 genera and about
77 species, mostly in North America, though several occur
in eastern Asia. The following notes will show the con-
ditions in those I have examined.

The carp-suckers (Carpiodes) occur in our waters
mostly west of the Alleghanies. In several species (C.
difformis, C. velifer and C. cutisanserinus) all have
tuberculated muzzles or snouts in the spring males. In
fact the last of these was so named by Cope for the
fanciful resemblance of this very character to ‘‘goose-
flesh.’’

The black horse (Cycleptus elongatus) has the body
finely tuberculated over its entire upper surface in the
spring. The upper surfaces of the paired fins are also
tuberculated. Only large or old examples were
examined.

Our common fine-scaled sucker (Catostomus commer-
sonnii) is the only one of its genus east of the Alleghanies
I have found tuberculate. It occurs all over our Middle
Atlantic States. The tubercles appear within a very
great range of age. Examples three inches long were
found with well developed milt and roe, like those nearly
two feet in length. The smallest examples with
tubercles, which were only on the lower caudal peduncle
surface, lower caudal lobe and anal fin, were four inches
long. Others, nearly of the maximum size attained by
the species, were found with exactly the same arrange-
ment of the tubercles. Examples with the entire upper
surface of the body and head, besides the dorsal and
paired fins tuberculated, were always over a foot in

No. 548] ORNAMENTATION IN FISHES 475

length, though seldom more. Further, I have every rea-
son to believe these small fish were also spawning with
the large ones, as I captured specimens of similar dis-
parity in the same waters in the spawning-season. That
many males spawn without ornamentation also appears
likely. No tuberculated females were found.

The chub sucker (EHrimyzon sucetta oblongus) has
quite pronounced sexual differences, the spring males
showing usually three large tubercles on each side of the
snout, the anal rays tuberculated, and usually the last
rays more or less incised. These characters only appear
in males over five inches, and until the maximum size
(eleven inches) is obtained.

The red horse (Moxostoma), with its numerous species,
often exhibits the anal fin tuberculated (M. macrolepi-
dotum). I have not obtained recent material, however,
for these studies.

In conclusion the following may be stated.

First. The disposition of the tubercles in the orna-
mentation of breeding males is not always complete in
the known design sometimes attained by certain in-
dividuals.

Second. The development of the tubercles may obtain
in comparatively small or young individuals provided
they are sexually mature. Tubercles occur, so far as I
have observed, in the very young, especially in those
species which undergo considerable transformation in
certain apparent generalized characters, such as the
development of the lateral line, change in coloration from
ii or spotted color-pattern to a more uniform tint,
ete,

Third. The tubercles are subject to individual char-
acteristics in being retained or lost for a greater or
shorter period after the spawning season.

Fourth. The adult male alone in most species exhibits
the tubercles, and though occasionally also present in the
female of some species the latter seldom assumes equal
evelopment in other respects, such as bright pigment,

c.

476 "THE AMERICAN NATURALIST [Vou. XLVI

Fifth. Sexually mature fishes spawn without the
tubercles appearing, and in some cases frequently.

Sixth. The effect of different waters, like those of the
sluggish warm lowland cedar-swamps where dark and
blackish, as contrasted with clear cold rapid upland
streams, does not seem to materially change the degree
of development.

Seventh. Many species have the inner pectoral rays
furnished with tubercles, and these no doubt are of
assistance to the males in clasping the females during the
spawning operations.!

t For the spawning in detail of some species see Reighard, in Science,

XVII, April 3, 1903, p. 531, and Bull. Bur. Fisheries, XXVIII, 1908, pp.
1111-1136, Pls. 94-120.

THE FORMATION OF CONDENSED CORRELA-
TION TABLES WHEN THE NUMBER OF
COMBINATIONS IS LARGE

DR. J. ARTHUR HARRIS

CARNEGIE INSTITUTION OF WASHINGTON

AFTER the principles of any method of research are
laid down by those who have the genius or the good for-
tune to make fundamentally new contributions, there
always remains much to be done in the refinement, simpli-
fication, or adaptation of methods to render them most
practically applicable in the routine of investigation.
This is especially. true in the modern higher statistics,
where, at the very best, the labor is excessive.

One of the most onerous of the statistical processes is
the determination of correlation in cases in which each
individual measurement must be weighted by comparison
with a series of others. In an earlier number of this
journal! a method was described for the rapid formation
of the heavy intra-class and inter-class? correlation and
contingency surfaces by the use of a machine permitting
simultaneous multiplication and summation. Methods
of dealing with such correlations without the formation
of tables will be published later. But abstract formule
in the hands of inexperienced calculators are apt to lead
to erroneous constants, which in the absence of the orig-
inal data can never be corrected. Again, the validity of
the correlation coefficient as a measure of interdepend-
ence depends largely upon linearity of regression. Hence,
tables should be given whenever possible. The purpose
of this note is to show how, in the case of relationships

***On the Formation of Correlation and Contingency Tables when the
aa of Combinations is Large,’ AMER. Nat., Vol. 45, pp. 566-571,

* These terms will be clear from their context in this note; they will be
more precisely defined later.

477

478 THE AMERICAN NATURALIST [ Vou. XLVI

involving a very large number of combinations, the chief
advantages of the correlation (but not the contingency)
surface may be even more easily realized than in the
method already described.

By condensed correlation tables are to be understood
those giving the (weighted) frequencies for a first char-
acter x and the first (and where necessary also the sec-
ond) rough moment about 0 as origin of the associated
array of the y character.? From such a tablet r may be
quickly obtained® and the means of arrays calculated for
linearity of regression tests.

In principle, the formation of these reduced tables is
very simple. Let x, y, 2, ---, be measures on the indi-
viduals of the same or associated classes. Let there be
n, P, q, ***, of these individuals. Then if n, p, q, =, I(x),
=(y’), =(2), ee z(a"), 3(y”*), 3(2°), ae (where 3 indi-
cates a summation within the class and the dashes indi-
cate that the measures are to be regarded as deviations
from 0) be again summed for each of the component
measures, seriated by grades, the four columns—grade
of ‘‘first individual,” weighted frequency, and the two
rough moments about 0 for associated individuals—thus
secured for each character either constitute the desired -
table or one from which it may be easily derived.

The arithmetical routine will be determined largely by
the nature of the records. Roughly, two cases are possi-
ble: n, p, q, ---, are small, m is small or large; n, p, q,°*")
are large, m is small, m being the number of classes or
groups of classes.

Suppose n, p, q, ---, small, say 4-20. The best method

*In direct intra-elass correlations z and y are measures of the same kind;
in ¢ross intra-class correlations de are different; in inter-class relation-
ships they may be the same or differ

*For example, Table X of Bonne, Vol. 8, p. 61, gx or Table II
nec ales Table I of the Amer. Nart., Vol. 44, p. 695, 1

The Arithmetie of the Product Moment we pater fot the Coeff-
desk a Correlation,’?” AMER. NAT., Vol. 44, pp. 699, 0.
€ Cases where both the numbers within n clea ae the number of classes

are large are very rare because of the great labor required in making the
observations.

No.548] CONDENSED CORRELATION TABLES 479

is to write the values of the first character under consid-
eration—designated for convenience as the subject—
down the side of a separate sheet for each class. Oppo-
site each entry is then written n, 3(a’) and 3(2”*), p, 3(y’)
and 3(4”), q, 3(2’) and 3(z2’") and so on, according to the
relationships desired. Thus, the measure used as the
_ subject and the number and summed first and second
powers of deviation of the individuals of the relative
array may be for the same or different characters or
classes, depending on whether direct or cross, intra-class
or inter-class correlation is to be computed. In any case,
the number and moments are only once determined for
each class—their repeated entry on the sheet is merely
rapid clerical work.

This done, the sheets are clipped into strips by subject
entries, the strips seriated according to the subject, and
the class numbers and moments summed for éach grade
on the machine.

For inter-class correlations, the resulting table is cor-
rect, embracing as it does, say, S(pq) entries. For intra-
class relationships, say for x, the entries are too high by
S(n), S(x) and S(x?) since it comprises S(n*) entries
when only Sn(n—1) are desired. Hence, the actual fre-
quency for each subject grade must be subtracted from
the weighted frequency, and the products of the actual
frequency by the grade and by the square of the grade
must be deducted from the first and second summed
moment column, respectively.

When the number of individuals per class, n, p, q, 18
large (e. g., 25 or over) another procedure is desirable.
The classes of the subject character are seriated (in
transverse rows) in a table of vertical columns captioned ©
by the grades. Opposite each row is entered n, =(«’) and
3(a"), p, 3(y') and 3(y”), q, 3(2) and 3(2”), +, for all
characters to be correlated. The associated (weighted)
values for each subject grade are quickly gathered by
multiplying up and summing simultaneously the fre-

480 THE AMERICAN NATURALIST [Vou. XLVI

quencies in each column of the subject seriations by the
opposed entries in the relative (number and summation)
columns. Again, the result is the desired table or one
from which it may be derived.

Illustrations will make the methods most clear. Table
I shows the frequencies for the different grades of radial
asymmetry’ of quinquilocular fruits gathered from 34
individuals of Hibiscus Syriacus in the Missouri Botan-
ical Garden in the fall of 1907. Table II gives the seria-
tions for the locular composition’ of the same fruits.
The last two columns of Table I and the next to the last
two of Table II give the first two summations for each
individual.®

radial asymmetry is at standard deviation of the number of ovules

per a. about the mean number ee ovules per fruit. See Biometrika,
Vol. 7, pp. 476-479, 1910, an “fall dise

At the head of this table the soctislaata z asymmetry are for condensa-
tion en to only two places. In all the calculations, however, they have

used to six places. Their values and their true squares as ‘ased in the
eee e are:

Asymmetry a a?
00 .00
400000 .16
489897 .24
632455 .40
748331 56
800000 64
427 .80
979795 .96
1.019803 1.04
1.095445 1.20
1.166190 1.36
000 1.44
1.264911 1.60
1.356466 1.84
1.600000 2.56

* Expressed here simply as the number of locules per i an ‘‘odd”
number of ovules. Cf. Biometrika, Vol. 7, pp. 483-487, ae

°” The last two columns of Table II give the peal Pe of Table I for.
convenience in determining the cross intra-class tables. When the cross —
intra-class tables are to be formed with asymmetry as the subject the
=(c’) and Z(e”) column may be added to Table I. Here it is omitted for
convenience in publication.

481

CONDENSED CORRELATION TABLES

FSS 9ZS9EF'ZS | ZOT — —

ma Ort O
—
_
mto
snl

A = <1 14 14 ro mi ot A
eae ee a, A

DNNN
=

87'S |£0£928' PF! TOT sie sei
0SE919" 6€| 66 EE ar

te
ho ie E 8 v1 BO) OP OF

kami

TOT o oa

a a a

at dit, tly tee

LOST6LFE | OOT esi TR
ZOOTPE'LZ| 66 = a
PYESE |ETSHSS OS | 86 on Z
$99 |EPOZ6L'ST| 46 po te
OF'0Z SLPILY FE} 46 Bi Eis

ELE ta tl dit ER bene

ae a a
SN OCN

pnn Dh iin te
e]

N U m T Oe u SS
3 ay ae a ae ee cae |

re

CON st OO NN

mrt

Z6'LZ senior 10r | — —

FCFSOT'OS | ZOT a0 =
9E8TZ9'SF| 66 sae =o ye
012089" SZz| OOT 5 oe

re
Lan

4 St OJ ON Y

rt OO OD HN Ht
nse
i oe

eo

ee el Fo bt pee Sue ee E A

len] [[ [aan | lalarl|alalaanad| | |anranan

“apeta EL iar a pee bil bai
al] loan[anataalfal | | || | laslalaan lade

wa Ebb TEAR ea EET

lll leet [tare] |
ANANHINAMOOHAD

Tos

4
MUANMDMMOMTOHHOHOORAN
A nN 6 es

N N R O m m i e HA

PT8 |TS60FS' FF | 66 a a

No. 548]

e | oe | w | eo | se | o
| SIPMA ENPFAFpUT Jo suopepaoq pavparjg—AyomuLdsy perpe |

| 09°T | oe'T | 93'T | oz'T | 9t'T | 60°T | TOT | 16°

s

(mz | (ox | w

STVAGIAIGN] A4 AUWLAWWASY ‘IVIGVY JO SNOMVNWAg ANV SNOLLVINAS
I WIdvVib

TABLE

II

SERIATIONS AND SUMMATIONS FOR LOCULAR COMPOSITION BY INDIVIDUALS

m Locular Composition—Number of
È “ Odd” Locules per Fruit N Le) | 2°) S(a') 3(a’)
0 I 2 3 4 5
Pree | oe tos ee OE 99 | 168 | 466 44.540951 | 28.24
2 196 (25 £184 GS 99 | 131 | 351 91.411571 | 17.28
3 | 46| 361i ii 99 77 F Mi 24.561697 | 12.16
4 | 67 | 30 D e ak 87 42 60 15.792043 7.38
oO 10 | 241-83 124-2 1 224 | 654 50.586565 | 31.76
6-1 98-18) 48 | 24 11-13 | 106: {220° | 612 51.819299 | 32.00
T Vth St o 1i 2) 200. 1.7102. | 250 25.580710 | 12.80
e 130191 Sh 22 37 124 9 225 | 691 45.621836 | 26.32
O i a (811 47) 29 1 3.402 | 191 | 533 50.105424 | 31.04
O era e elor 7 = 4 10s 131. | Bt 40.556715 | 24.64
Horer ar 7r a 101 | 104 40 46.631638 | 27.92
12 EINI 2s) oot 0 99 | 199 | 521 51.558133 | 31.44
13° Sn r24] 971 0l 3.9 97 118 84 34.471473 | 20.40
14 | 59| 33 ta ee 97 44 58 15.792043 6.64
15 9) l9 r242: 8| å 98 | 211 | 599 50.254513 | 33.44
ie HAS Sab Te fed od | 99 96 02 27.941002 | 14.64
1 St eo on ie Fa | 100 168 | 427 34.791567 | 19.28
18 50 b-26 ta eB oe 99 TI 15148 24.159111 3.04
Io 266 Ist oR TI le 99 55 | 113 17.750439 9.36
W 72 | 21 |B 1 bee te oo 38 66 12.719177 6.24
I Ae 87 to &} 45 4-100 4106 |= 250 30.618961 | 17.76
22 Al | 38} 907 16 | 4i 3 | 100 | 132 | 368 29.498695 | 16.00
B34 Sa 355) 641 12 Sr e 100. | DO 988 33.917659 | 19.04
m 1 32 1 10-) 28 | a7 6 | 8 99 | 150 | 410 39.675350 A4
D EAT 136) 28 bd | 5b - 101} 160 + 8798 44.876305 | 25.28
26 | 26 231 14'|°10 | 3 98 | 165 | 475 37.794451 | 22.16
2 be | 31 1. 10 | Be ik = 00 57 89 20.164872 9.76
28. |38 1 30| i i 100 97 42 287 29.402270 | 14.80
29 & | 27 |31 20| 16 96 | 189 | 491 52.288755 | 32.56
30 | 44 | 975-4 O12 98 96 | 216 31.425564 | 18.72
st U1 IA | SF) 44 19] 5 | 100 | 231 | 717 49.34444 30.16
oe | 28 | 33 | 24 Ie 4] 3 | T10 | 138 | 32 41.749890 | 26.24
ss 80 1-26) 17 1-441 5T 157 | 475 42.901029 | 29.04
34 tI 251 2 t2119: 1 1 10 | 227 | 6590 52.456526 | 31.84
TABLE III
LOCULAR COMPOSITION
0 1 2 3 4 5

i ios — ae a S 49

.400000 - — 30 ee = 187 —

.489897 2i ai 420 | 306 — -=

D .632455 neo ion we l i a

3 748331 —— — 179 Gi —

z] .800000 45| 101 e an 12 4

E .894427 — se es es 1| —

2 .979795 14 12 = 5 2

Be 1.019803 ae ak 6 1i —— a

= 1.09544. —— — T 1 — —

E 1.166190 a 8 ioe a

m 1.2 = eto oal

1.264911 o 1 a a e rae

1.356466 — how 1 1 ae pai

1 + nes ies pee ae a

Totals, 1,098 | 886 | 688 | 461 | 205 55

483

CONDENSED CORRELATION TABLES

No. 548]

FS°C69' TL 8Z6S°SE6'TZT 9se'ose CL'¥SL | L092°ZES'T €68'€ 96°9TF'SL cece TLT eal 6IL'8EE

T

96° FZF'T CSO8s See's cers SYP 96ST'S ce Pr'6ZP'T LOPP OPES O06F'S ppo 9

96'9FF'S OZ8F' 86'S 9TE‘0% 0G EP PE6L'06 e07 9T'06F‘S FEL9'EL0'6 1ZS‘0Z ppo F

G1 609'TT 89FG6'SSS'6I FIO'SP SF 0ST 696° L196 TOF 09°6EL‘TT OSPE OSF'6T cL0'Or ppo £

ZS’ LOL‘9T FITE'SES'LS 620°89 OF 09% SLLL' SOP 889 G6'L96'9T 6880'6E7'83 292'89 ppo g
8Z°6LT'ST TILES SST'TE 69S‘ L8 9S TES’ E916 PEP 988 $9 OTF'ST FES LESTE 8PhPF'S8 ppo T
OF PZE'ST OSE0'6EF'SE 0ZE‘SOT OSPF TAISI 860'T 02 69¢'8T TSSE'06F'ZE SLF'60L ppo 0

oL D [B10], u p? 0L M [810 u 3? T80L 7 [BIOL u | uni somo astnooT
[Quy FUJIO A peyonped oq 0} son[e A SINJU A S019) |
AULINWASY ANV NOMISOdWOD UV'INDO
A WIV

Po'Z69' TL 626986 TST 9SE'SEs GL VGL LOOL’GES'T 68'S 96'9TF'SL CESS LLT'SS1 | 612'8ee

80°37 L9G6'8E Zor 99g 0009°T I PIPZ Ł999°0F Sot 000009'T
07'£9 T962°L6 P6L 89'E 6611'S Z 88°99 0609°00T 961 99P9SE'T
FP 6G SOPS'SP TOT og'r 6F97'I ji PO'TE PSOT'Og OT TI6F9U'T
OF SFT OOLS° ZFS 86F oZ 0000°9 g 09°S¢T TOT8S'SPZ £09 000008'T
ELSES OTL] S68 106 PESI LS6F'OI 6 96'093 1992°SOF O16 O6L99T'T
8769S 9F09'G6 S61 OFS 6061'S G 89'T9 Co6L'F6 L61 SPPrs60'T
88'ZEP EOST LIL 069'T 8O°LT L988" LT LY 9908F OL88' FEL LOL‘T €08610'T
00'9LL 9089'983‘T GSG'E 89'TE GEES SE EE 89°08 LETO'STE'T 987'E 862616
6£'000'T SZ18°629'T SSL‘ OF'08 Z886' EE 8E ZZ'OS0ʻI 2008°899'T 06L'E LEPV6S"
98666 9889" L99'9 000°9 89°SOL 0009°66T Z9T FO'S0L'F 9EET L829 ZOT‘OT 000008°
82 LEF'S GEST SF6'8 0L9'0% POLIT SLOP 9ST 603 ZE'PPS'S L¥89°660'6 6L8'0Z TEESPZ'
82° LZPS ThPS Ses" 2 eho‘ ST 09°EL LZILE°9TT PST 9¢°919'F 6919°899‘2 LZP'ST Ccrceo®
SE'SSS‘LT G809°682'6% 008ʻT4 $c FLT 6999'SSE 962 99962‘ LT SELZ°S19'6S 93932 26868F'*
GE OFS 6T ITF6'ZS8'SE T69'06 GL’OFT 0008°99€ L16 ¥0' 28861 OIFL GSTs 809'16 00000F'
89°6ST'ST 6E8'ZS1'SE LEZ‘ LOT 007000 0000°000 280'T 89°6ST'ST E6E8'SST'ZE PZE'SOT 000000"
z? PIPL M IOL u | gz? T80L 9 ROL | u | gz? [BIOL 7? OL u | Lnewusey

IQR SuryIoM | peynpeqd oq 0} sƏNJBA | Sen[BA 8501)

AMLIAWWASY ANV AULANWASY
AI WTavil

484

THE AMERICAN NATURALIST

class tables.?°

The columns under ‘‘

(Vot XLVI

From I and II, the machine quickly compiles four work-
ing tables—a direct intra-class for asymmetry, a and
another for locular composition, c, and two cross intra-

gross values’’

in
TABLE VI
ASYMMETRY AND LOCULAR COMPOSITION
Gross Values Values to be Deducted | Working Table
A
n | Total ¢ | Total c’2 [Total o|Totale® n | Total c’ | Total of

.00 | 108,324 | 117,335 288,699 | 1,087 245 | 1,225 | 107,237 | 117,090 | 287,474

Al 91,608 | 127,391 332,887 1 141/8 | 3,122 90,691 | 125,913 329,1

.48 72,52 117,550 318,254 726 | 1,758 | 4,434 1,8 ,792 13,820

.63 18,427 | 30,358 81,952 1 7 1,291 18,243 | 29,879 ,661

.74 20,879 36,577 100,993 209 | 448 9 20,67 36,129 100,007

.80 16,162 26,180 70,5 162 169 393 16,000 26,011 70,167

.8 nri 54 18,09 38|: 4 53 3:10 ,508 18,040

.97 3,2 5,014 13,422 833 42 142 8,252 4,972 13,280
1.01 1,707 3,040 8,460 ITA 45 123 1,690 2,995 337
1.09 19 379 1,06 2 | 5 13 195 37 1,052
1.16 910 ,600 A 9 19 41 901 1,581 4,365
1.20 503 | 1,024 2,954 H s 5 98| 1,019 2,949
1.26 102 191 523 $ | 1 1 101 190 522
1.35 196 422 1,198 2 | 5 13 194 417 1,185
1.60 | 103 131 5 1 | 0 0 102 131 311

|
|

338, 719 | 473,741 | 1,243,777 | 3,393

4, 740 | 12, 442| 335,326 | 469,001 ti, 231,335

fal 990,949 | FUTVA o

TABLE VII
LOCULAR COMPOSITION AND LOCULAR COMPOSITION
‘Los. | Gross Values Values to be Deducted Working Table
Comp.
| n | Total c’ Total «°? n | Total c’ Total e? n Total c’ Total r:
| | ewer
0 109,418 | 117,758 288,576 | 1,098 | 0,000 | 0,000 | 108,320 117,758 288,576
1 448 | 118,815 305,713 886 886| 87,562 | 117, 929 304,827
2 TOTA 111,775 302,47 688 | 1,376 | 2,752| 68,079 110,399 ,
a 46,075 78,15 213,853 461 | 1,383 | 4,149) 45,614 76,768 209,704
4 20,521 37,538 105,540 205 0 | 3,280} 20,316 36,718 102,260
5 | 5490| 9704|) 27620) 55| 275| 1375| 5435| 9,429| 26,245
ee
338,719 | 473,741 | 1,243,777 | 3,393 | 4,740 (12,442 | 335,326 | 469,001 | 1,231,385.

Tables IV-VII give the results.
p=q, a total S(p?) =S(q?) =S(pq) entries, whereas in

S{p(p—1)1 =
S{q(q—1)], and in the cross intra-class Sipit

the direct

intra-class

relationships

ee are desired.

r the relationship between radial asymmetry and locular compe

These contain, S

sition, ay other for the correlation between locular composition and
asymmetry.

he same end result, and Beas one need
be found unless the linearity of both regressions is to be t :

Of course, both give

since

*

No.548] CONDENSED CORRELATION TABLES 485

From these gross values must be deducted, therefore,
the actual frequency for each grade of the subject and the
product of the frequency by the first and second power
of the grade in the case of direct intra-class correlation,
or the frequency of the grade and the sum of the first and
second powers of the values of the relative character in
the same fruit in the cross intra-class correlation. Data
for these are given in the table showing the correlation

-for asymmetry and locular composition of the same fruit,

Table III. The second set of three columns in Tables
IV-VII gives the quantities so calculated from Table ILI
to be deducted. The final three columns are in each case
the working tables.

The first and second moments for the (weighted) popu-
lation A and o are given by the totals of the two final
columns. Or those for the subject character may be cal-
culated (and a check for the accuracy of the totals
secured) from the grade of the subject and the weighted
frequency column."

From our working tables, indicating by S a summation
from our final tables, we determine by the methods of
Amer. Nar., Vol. 45, pp. 693-699, 1910, these values:

| For Asymmetry
S(a’) =121,938.5928, Aa= .363642,
S(a’?) = 71,692.2400, ==.285093.
For Locular Composition
S(c’) = 469,001, Ac==1.398642,
S(e”) == 1,231,335, Oc = 1.309906.
For Asymmetry and Locular Composition
Table IV, S(a,/a.’)—= 48,818.9505, r= .1637,
Table VI, S(a,/c.’) =192,072.3309,'* r= .1716,
Table V, S(c,’a9’) =192,072.3308,” r= .1716,

* Of course in pfactise, the second population moment may be calculated
by S[(n—1)2(2")], S[(p—1) 29], S{(q—1)2(2")],.-., thus
obviating the labor of forming the third columns, which are included here
for completeness of illustration merely.

"The difference of .0001 is due to the necessity of lopping off the last
two places of the six decimals in the asymmetry coefficient in the one case
while they can be retained in the other. Of course, it is of no practical
Significance.

486 THE AMERICAN NATURALIST [ Von. XLVI

Table VII, S(c,’co’) =763,048.0000, r—.1861.

While primarily illustrations of method, these results,
if they are substantiated by further work, seem to me of
considerable biological interest. They show not only that
individuals of H. Syriacus differ in the radial asymmetry
and in the locular composition of their fruits, but that
when an individual bears fruits above the average
asymmetry, it also produces fruits above the average in
number of ‘‘odd’’ locules. Apparently, this cross corre-
lation is as high as either of the direct correlations.

Two biological interpretations are possible. (a) The
production of radially symmetrical ovaries and those
with a high number of odd locules depends upon the same
morphogenetic tendencies of the primordia,!* which give
rise to the fruit. (b) There is in Hibiscus an intra-indi-
vidual selective elimination similar to that demonstrated
in Staphylea, the intensity of which differs from indi-
vidual to individual in such a way as to bring about
(statistical) correlation for characters originally uncor-
related.

The discussion of these points falls outside the scope
of the present note where the data serve merely as a
random illustration of a very rapid method of carrying
out the routine of a widely applicable statistical pronti

COLD SPRING HARBOR,
April 25, 1912

3 In the individual fruit radial asymmetry and locular composition are
necessarily associated (cf. Biometrika, Vol. 7, pp. 491-493, 1910). In
Staphylea, correlations of r==.22 to r==.33 have been noted. Table III
be gives r= .527 for asymmetry and locular composition of the same

coer in all these relationships Re is not linear, and the corre-
lations abe be interpreted with cautio

“ Biometrika, Vol. 7, pp. 452-504, 1910; Science, N. S., Vol. 32, pp. 519-
528, 1910; gje Na f. Ind. Abst. u. Vererbung, Vol. 5, Pn. 273-288, 1911;
Pop. Sci. Mo. Vol. 78, pp. 534-537,

SHORTER ARTICLES AND DISCUSSION

PRODUCTION OF PURE HOMOZYGOTIC ORGANISMS
FROM HETEROZYGOTES BY SELF-FERTILIZATION

Has any one worked out formule for determining the rate at
which organisms become homozygotie through continued self-
fertilization? The problem is of interest in various connections,
but principally perhaps with relation to the ‘‘pure line’’ work.
Johannsen worked, for example, with self-fertilizing beans that
are held to be homozygotie in all respects. I have often heard
the questions raised: How probable is it that such plants really
are homozygotic? Is it indeed possible that they have reached
a purely homozygotice condition?

I have not come across a working out of this matter, and
finding it necessary to deal with the problem in connection with
studies of inheritance in Paramecium, it will perhaps be useful
if I put on record the results.

The principles which underlie the matter are the following:
(1) In self-fertilized organisms, all characteristics that become
once homozygotic, remain homozygotic forever after, since there
is no method in self-fertilization of introducing a gamete that
is diverse in this respect; (2) characteristics heterozygotically
represented become homozygotie in a certain proportion of the
offspring. - The problem becomes essentially this: in what pro-
portion do the heterozygotie characters become homozygotic, and
how great a proportion of all the organisms will therefore have
become thus homozygotic after a given number of self-fertiliza-
tions ?

Suppose that we begin with an organism in which all separable
characters are heterozygotically represented.

1. Consider first a single pair of such alternative characters,

which we may call { = The gametes produced will be A, a,

*At the moment of receiving the proof of this note I receive also the
masterly paper of East and Hayes (‘‘Heterozygosis in Evolution, and in
Plant Breeding,’’ Bulletin 243, Bureau of Plant Industry, June 5, 1912),
in which this matter is dealt with and a general formula given. ‘‘The

487

488 THE AMERICAN NATURALIST [VoL. XLVI

A, and a, and when these combine in all possible ways, they give
zygotes AA + Aa 4- aA + aa; that is, two homozygotes and two
heterozygotes. Thus, the self-fertilization of such an organism
gives 14 the progeny homozygotic (with respect to this charac-
teristic); 14 heterozygotic. If we let x= =the proportion of
homozygotes, y the proportion of heterozygotes (with respect to
one character), then after the first self-fertilization

x= Y¥, of all.

y =l% of all.

Now, after the next self-fertilization, of course the homozy-
gotes x remain pure, so that half of all the progeny are still ho-
mozygotes on this account. The heterozygotes y of course again
break up, in the way already set forth, one half into x, the other
half remaining y. Since y included half of all, this will give 4
of 4% (= of all) as x, % of % (= of all) as y.

So the total proportion for the homozygotes x becomes after
the second fertilization

s= y% + (%)’?=%,
while
| y= (14)? =, :

This process is repeated after each fertilization, so that if
there are n fertilizations in succession, the total number of homo-
zygotes x, becomes

a= + (%)?+ (%)*.. . up to (%4)*.

i ; 2 1 i
This expression reduces to % = ye? where n is the number

of fertilizations. l g
_For the heterozygotes, y, on the other hand the formula is
simply

y= (14).

These then are the formule in case we deal with but one pair
of characters. They express (1) the proportion of all the organ-
isms that will be homozygotic (or heterozygotic as the case may —
be), after a given number n of fertilizations; (2) also they of
course express the relative probability for a given case, as to
whether it shall be homozygotie or heterozygotic.

2. When we are to deal with two or more pairs of characters,
the problem may be attacked in two ways. One is by the gen-
eral principles of probabilities; the other is by analyzing the
case of two or more characters in the way exemplified above.
The two methods give the same results. . oa

The first method is far the simpler. It is merely an appli-
cation of the principle that when we know the probability for i

No.548] SHORTER ARTICLES AND DISCUSSION 489

each of two or more things separately, the probability that all
of them shall happen is the product of the separate probabil-
ities for each. Now, we know that the probability æ for the
homozygotic condition with respect to one character is

2” —1

po

oe o&

For two characters it is then

2” — 1 2"—1 iv
p) x (= > or ( Qn y
ae 2"—1)\°
For three characters it is of course uae and in general,
for any number m of characters, the probability æ for pure ho-
mozygotes (or the proportional number of pure homozygotes) is

(EFE)

By similar reasoning, the proportion of all the organisms that

will be heterozygotie with respect to all the m characters is
y = (1%).

With two or more characters, there will be of course a consid-
erable number of the organisms that are homozygotie with re-
spect to some characters, heterozygotic with respect to others.
If we call the proportion of these z, then

e=1—(x+y).

And if we let v be the total proportion that contains any
heterozygotic characters (so that v=y + 2), then

J 2n pen 1 m gas u (2" — 1)"

v=1- (4) =— p.
E- z

These formulæ may readily be deduced algebraically, or veri-
fied, by a detailed analysis of a case of two or more characters.
It may be worth while to indicate the method followed, by taking

up the simpler case of two pairs of characters. Call these { x

nd s - The gametes formed are AB, Ab, aB, and ab. When

these combine in all possible ways (as indicated in the diagrams
given in Bateson’s Mendelism), these give the following results:

14BAB +. 14AbAb + 1aBaB + labab + 24Bab +
2AbaB + 24BAb + 24 BaB + 2Abab + 2aBab = 16.
It will be observed that of the entire 16, the first four are pure
homozygotes, the second four are pure heterozygotes (heterozy-

490 THE AMERICAN NATURALIST [Vou. XLVI

gotic with respect to both characters) ; while the last 8 are mixed
(homozygotic with respect to one character, heterozygotie with
respect to the other). Letting z= pure homozygotes, y = pure
heterozygotes, z==mixed,:we find thus that

= 4, y=, e-= ¥f, of all.

Now, by an analysis of the sort already given, it will be found
that at the next self-fertilization, remains x; y breaks up, 14 of
these ‘becoming z, 14 becoming z, and 14 remaining y; z breaks:
up, 14 of these becoming x, 14 remaining z.

Now, when we recall that before the second fertilization æ was
1; Y, 14, and z, 14 of all, we see from the above that after the
second fertilization

2 — 1y
2=4 + (%X%) + UXW =A= (=H),
y=(%4 X uU) =Yo= (%2)”,
z= (V2 X Y2) + (12 X U) == (%)* + (4).

These are the same formule for x and y that were obtained
by the other method (since here n and m are each 2). This
method however gives in addition a direct formula for z.

It is easy to verify the formulæ for three pairs of characters,
though of course the conditions become here somewhat more
complex.

We may now summarize our formulæ, and show the results
they give in certain examples.

Let =the proportional number of organisms that are pure
homozygotes (with respect to all the characters con-
sidered),

y =the proportion that are heterozygotiec with respect to
all the characters considered,

z=—the proportion that are mixed,

v= the proportion that have any heterozygotic characters.

Then, if nthe number of successive self fertilizations and
m = the number of pairs of characters,

we ( ae y, (1)

y= (14), (2)
z=1— (z +y), (3)
Y= a ie (4)

Examples—(1) Suppose that there have been eight self-fer-
tilizations, and that we are dealing with 10 pairs of characters.
What proportion x of the organisms will be homozygotie with

No. 548] SHORTER ARTICLES AND DISCUSSION 491

respect to all the 10 characters? What proportion will be
homozygotic with respect to any given one character? To any
two or three?

Taking first the case for the entire 10 characters, by formula
(1)

8 10

“= (- z) = (a 3 = log. 9.9830020 = .961617.

Thus, out of 100 individuals, somewhat above 96 would be
pure homozygotes; or by formula (4), but one in 26 would be
heterozygotic in any respect (v= .038383).

With respect to any one character formula (1) gives

pa 1 Bao
C= (=)= aT a .9960937 5,
so that all but 4 in 1,000 would be homozygotes with respect to
that character.

In the same way we find that with respect to any two char-
acters the proportion of homozygotes would be .9922; with re-
spect to three, .9883; with respect to four, .9845, ete.

(2) Suppose that there are 20 pairs of characters, and that
‘there have been 20 self-fertilizations, Then

oP TN / 1048 SIO
t= (=>) = (x oas576) = log. 9.9999957 = .999998.
That is, of a million individuals, all but two would be pure
homozygotes.

It thus appears that if the number of separably heritable
characters is not very great (say not above 100), while the or-
ganism has been self-fertilized for many generations, it is to be
expected that practically all of the organisms will be homozygotic
with popet to all their characters, they will be ‘‘pure homo-
zygotes

H. S. JENNINGS

YELLOW AND AGOUTI FACTORS IN MICE NOT
‘‘ ASSOCIATED”

In a recent number of the AmerIcaN Naruraist Mr. Sturte-
vant! suggests that these two color factors may bear to each other
the relation which Bateson has called ‘‘repulsion’’ or ‘‘spurious
allelomorphism’’ and which Morgan now includes with ‘‘coup-
ling’’ in a more general category, ‘‘association.’? The supposed

* Sturtevant, A. H., ‘‘Is there Association Between the nye and —
Factors in Mice??? Am. Nar, Vol. XLVI, No. 546, p. 368, 1

492 “OTHE AMERICAN NATURALIST [ Von. XLVI

relation is such that the characters involved fail to pass into the
same gamete even though they may be present together in the
parent zygote. That yellow and agouti in mice are not in gen-
eral so related is shown conclusively by experiments which will
be more fully described elsewhere, but which may be briefly
summarized in the following table:

Parents Offspring
Mating
Both Yellow Yellow Agouti Black or Brown

185 894 X895 1 1 5
10 502.24 X502.5A 10 2 1
397 3,908 X875 4 2 2
273 2,049 X875 5 1 1
159 786 X784 5 3 ul
446 Unmarked 4,054 8 6 2
500 Unmarked 4,152 2 3 1
545 — x4,1 4 2 1
467 chins 9 i Pa 1
519 Cuniacked sn oo 6 2 1
543 2 1 3
397 3,908 y pia 4 2 2
240 1,828 X 1,829 10 1 6
113 562X 563 6 3 2
173 1,074X 563 2 1 2
Total 78 31 31

In these experiments yellow animals bred inter se have produced
non-yellow young half of which are agouti and half of which
are non-agouti. It seems therefore to be wholly a matter of
chance whether a yellow animal heterozygous in agouti trans-
mits that character with yellow or apart from it. Sturtevant’s
contrary conclusion is due in part .to his reliance on the insuffi-
cient numbers observed by Morgan and in part to his overo a
certain of the results reported by Miss Durham. For, in a
tion to the category of matings of yellow mice cited by Sturte-
vant, she reports matings of sable (yellow) mice inter se which
produced 17 sable (yellow), 8 yellow, 5 agouti, 4 black, and 2
brown young, a result in harmony with that which I have de-
scribed.

In the matings reported by Miss Durham in which yellow
parents produced only yellow young and agouti young, it seems ~
probable that one or both of the yellow parents was homozygous
in agouti. The same was probably true in the similar experi-
ments of Morgan. This would explain why all the noae
young were agouti marked.

As further evidence that yellow and agouti are wholly
independent characters may be cited experiments of my ow?

in which yellow animals evidently heterozygous in agouti were | is

No. 548] SHORTER ARTICLES AND DISCUSSION 493

mated with brown animals which invariably lack agouti.
There were produced 15 young, of which 7 were yellow, 5 agouti
and 3 black or brown. Evidently the yellow parent transmitted
non-yellow (black or brown) in 5 cases associated with agouti,
and in 3 cases not so associated. On Sturtevant’s hypothesis all
non-yellow young should have been agouti.
C. C. LITTLE
LABORATORY OF GENETICS,
BUSSEY INSTITUTION, HARVARD UNIVERSITY,
June 17, 1912

PHYSICAL ANALOGIES OF BIOLOGICAL PROCESSES

Two schools or methods of thinking of heredity and other
general problems are recognized among biologists. Some hold
that all biological phenomena are to be explained in terms of
physical and chemical properties of unorganized matter. Others
are inclined to believe that the activities of living matter repre-
sent agencies or relations not shown in the inorganic world. The
first view has been called materialism, the second vitalism.

These distinctions are not as important as sometimes supposed,
because of our inadequate knowledge of the properties of matter,
whether organic or inorganic. The materialistice view may be
said to have a practical advantage in encouraging the investiga-
tion of the physical and chemical phenomena of the organic
world, but vitalism may claim at least an equal advantage in
permitting the recognition of facts that lie on the other side of
the biological field, where the analogies of physics and chemistry
find little or no application. Thus the specific constitution or
speciety of living matter, the fact that organisms maintain their
existence and make evolutionary progress only in groups of
individuals united into specific networks of descent, involves the
recognition of a condition or property quite foreign to the usual
conceptions of the physicist or the chemist. Yet this universal
condition of speciety must be considered as a general basis or
background for any strictly biological study of the organic
world. We may count, weigh, measure or analyze the bodies
and activities of organisms from as many other standpoints as
we please, but it is idle to draw general biological conclusions |
from any merely mathematical or physical data. The true
biological significance of statistical and physical facts has to be
determined by biological analysis. es,

494 THE AMERICAN NATURALIST [ Vou. XLVI

desirable and legitimate. Even an incomplete and inadequate
analogy may be very useful for descriptive purposes. But when
the elaboration of an analogy interferes with perception of the
fact it is supposed to elucidate the limitations of the method
become apparent. This danger may be illustrated by a recent
attempt to define biological evolution in physical terms.

In the minds of most of us the term ‘‘evolution’”’ is associated
probably more closely with the biological than with the physical
sciences. Yet the concept is essentially physical in character,
and is definable in exact terms probably only in the language
of physics. For in its last analysis we may define evolution as
the history of a material system undergoing irreversible trans-
formation. To the physicist, therefore, the study of evolution
is essentially the study of irreversible changes, and the law of
evolution is the law of increasing entropy, or, more generally, of
the increasing probability of the successive states of any rea
material system. j

Whether such definitions have any practical advantage as aids
to further investigation may be questioned. Exact terms are
of little use in biology unless they are also concrete, that is,
unless they convey an idea of something that can be seen, Or at
least imagined. Physicists are much more tolerant of mathe-
matical and metaphysical abstractions. The habit of using ab-
stract conceptions often leads to the announcement of wonderful
discoveries that resolve themselves, ón closer inspection, into
purely metaphysical manipulations of terms. Thus it is con-
sidered one of the notable services of a ‘‘supreme biologist” that
he should have identified chemical reactions with organic
tropisms, tropisms with instincts, instincts with morals, and then
predicted a chemical analysis of morality.? :

Such inferences only show that the vocabulary has been thrown
into solution, not that any concrete insight has been gained.
Physical terms can be set into formule as mystical as a wizard’s
incantations. Sleight of hand is discredited, but verbal miracles
are still performed with ‘‘scientifie principles.’ When 80-
ealled physical science runs to seed in antinomies it becomes
plain that we are back again in the vicious circles of meta-
physical deduction.

To describe biological evolution as a process of irreversible
transformation seems especially inappropriate because it is of

1 Lokta, Alfred J., ‘‘ Evolution in Discontinuous Systems,’’ Journ. Wash.
Acad, Sciences, Vol. II, No. 1, 1912, pp. 2-4.

21t The Chemistry of Morals,’’ Current Literature, 52: 180, February,
1912.

No. 548] SHORTER ARTICLES AND DISCUSSION 495

the very nature of biological changes to be reversible, or recur-
rent in each generation. Only by limiting the idea of evolution
to changes in the underlying mechanism of transmission could
the specification of irreversibility be made to apply.

Nor is there justification for considering biological evolution
under the caption of discontinuous systems. Variations, in the
sense of changes of expression of characters, are often discon-
tinuous, but this does not mean that evolution is discontinuous.
Some biologists have supposed that new characters constitute
new species, but in reality we have to think of the old species
as developing the new characters instead of the characters
originating the species. As long as species are considered only
in a statistical or biometrical sense, the existence of different
species will appear to rest on evidence of mathematical discon-
tinuity. But for any truly biological purposes such discon-
tinuity must be considered as a result of evolution, rather than
as a condition or cause of evolution. All characters, in the
evolutionary sense of the word, are first presented as differences
among the members of species. Such differences are always
variable, or alternative in expression, which is another way of
saying that they are reversible; that is, they appear and dis-
appear, or are expressed in various degrees, in the different in-
dividuals belonging to the same specific group.

The phenomena of mutation represent discontinuity among
members of the same species, rather than differences between
Species. Natural species seldom have the same kinds of differ-
ences as mutative variations. The theory that species originate
by mutative changes of characters is well calculated to deceive
physicists and biometricians who are not familiar with the facts
of diversity in natural species. Ortmann diagnosed the diffi-
culty with the mutation theory by saying that DeVries does not
know what a species is. Familiarity with species is seldom con-
sidered as a necessary qualification for the study of evolution
and heredity. Indeed, most of our physiological and mathe-
matical biologists have dismissed species as something too 1m-
definite for their purposes. They prefer to begin with i
definition that can be stated in exact terms and turned into
figures or formule. The casual choice of words like irreversible
or discontinuous becomes fraught with a vast importance, seldom
mitigated by any appreciation of the fact that most of one
anguage of biology is merely descriptive and comparative, and
hence to be understood only in relative senses.

If we cut through a tree top it will appear that the branches ee
are discontinuous, but if we follow individual branches oa, le

496 THE AMERICAN NATURALIST [Vou. XLVI

effect of discontinuity is lost. For some purposes the process
of growth might be described as continuous or gradual, and
for other purposes as discontinuous, for there are daily inter-
ruptions or variations in the rate of growth. Darwin compared
the evolution of species to the growth of branches on a tree.
Theories that would supplant Darwin’s conception of continuity
in the evolution of species are not based on equal familiarity
with the facts. Constructive evolutionary progress comes by
gradual changes in the characters of species, not by saltatory
transformations of one species into another. That albinism and
other defects appear as mutations and show Mendelian in-
heritance does not destroy or even conflict with the evidences of
continuity in the evolution of species.

The many different applications that can be made of such
terms as continuity or discontinuity show how little is gained by
the choice of any particular statement of biological facts as a
basis for deduction or mathematical elaboration. That new
facts can be learned by syllogizing is no longer believed. Is
more to be gained by turning syllogisms into mathematical
formule? Even when physical analogies can be more definitely
drawn the biological relations of the facts seldom permit any
complete application of mathematical methods of thought.

Taking evolutionary deviations into account, the recurrent life
cycles of organisms could be considered as spirals and this might
appear to bring them within the range of mathematical treat-
ment. Yet no regularity of form could be ascribed to such
spirals, for evolutionary intercalations of new characters are not
merely additions at the ends of definite series, but are likely to
intervene in any part of the life cycle. Moreover, such inter-
calations are made without throwing the remainder of the cycle
out of adjustment.

From the historical standpoint evolution may be presented as
a series of transformations, but from the standpoint of heredity
the results must be treated as permanent and coexistent, not
merely as successive reactions. Each member or part of the
eycle must be supposed to carry the determinants or potentiali-
ties, the powers of reproducing every other part. And we
know further that these potentialities continue to be carried in
latent or recapitulated form, even after they have ceased to
come into normal expression. In other words, the piological
system retains latent possibilities of reversion, the maps, as it
were, of courses of development long since abandoned. Even —
though the characters be considered as irreversible in the sense
of having permanent transmission they are subject vever 3

No.548] SHORTER ARTICLES AND DISCUSSION 497

to endless vicissitudes of internal and external relations that
influence the expression of characters.

Thus the life-cycle spiral is not a single line, but is resolved
into a vast multiplicity of lines, tracing back to all the different
ancestors. The expression relations of the characters can be
analyzed, in a measure, by breeding experiments and by compar-
ing behavior under different conditions, but the nature of the
system is such that its most permanent and stable adjustments
are those that are farthest removed from the influences of the
environment. Hence the futility of mathematical treatments
that attempt to combine environmental vicissitudes with the
entirely inconsistent facts of biological evolution. Instead of
evolution representing a law of increasing probability of suc-
cessive states, the contrary would be more nearly true. The
more specialized the organization the less the probability of
passing to another state. Thousands of species are extinguished
to one that develops a higher type of organization. Instead of
evolution representing a summarized result of environmental
influences, it is rather to be considered as a history of ways of
avoiding such influences. There are no facts to show that evolu-
tionary progress is caused by environmental agencies or by in-
ternal mechanisms. The causes of evolution lie in the structure
of the species.

The crowning complexity of the biological system is that the
life-cycle spirals do not remain separate and distinct from each
other, but are thoroughly interwoven to form a continuous net-
work of descent for each species or group of interbreeding indi-
viduals. It may be that a mathematical formula could be made
for a spiral wire mattress or a Turkish rug, but this would afford
only a faint analogy for the complications that attend the evolu-
tion of a species. To describe a species as a network of descent
may be only a figure of speech, lacking altogether in mathe-
matical exactness, but it is a way of pointing out a concrete
biological fact. When the facts are essentially complex any
statement that conceals or disregards the complexity is to that
extent specious and misleading. :

Many of the results of evolution can be described in simple
terms of quantity or sequence, but to consider such statements as
definitions, or to suppose that they confer any special license for
mathematical elaboration, is to disregard the true nature of the
facts and problems of biology. For purposes of physics or
chemistry individual plants or animals may be considered as-
machines, but for purposes of evolution the whole species repre-
sents the machine in which the patterns of new characters are

498 THE AMERICAN NATURALIST ~ [Vou XLVI

woven. If mathematical elaboration is to serve any useful
purpose in showing how evolutionary progress is made the nature
of the machine, the specific organization or speciety of the
organic world, must be recognized.
O. F. Cook
WASHINGTON, D. C.,
March 15, 1912

ON FAIRNESS AND ACCURACY IN SCIENTIFIC
REVIEWING

TO THE EDITOR OF THE AMERICAN NATURALIST: Aal one who
takes a turn at the critical hoe with the object of ridding the
biological field of some of the noxious products of fertile imagina-
tions untrammeled by quantitative facts must expect just the
sort of attack which appears in your recent issue (AMER. NAT.,
March, 1912, p. 165).

1. Dr. Spillman has not felt the purpose, methods or results
of my paper worth statement. Instead he illustrates by it the
‘“‘noticeable degree of correlation between positiveness of state-
ment and inaccuracy of statement.’? And for the reason: ‘‘in
Dr. Harris’s paper he represents me as having cited the fact
[sic] that these genotype norms [sic] form a frequency curve
[sic] as proof of the genotype hypothesis [sic].’’

One excuses the minor inaccuracies and would be glad to pass
over the whole assertion with the simple comment that it seems
ToJo but to protect himself against further accusations of

‘‘inaccuracy of statement’’ he must add, it is not true.’

What I did do was to cite Dr. Spillman among three others in
substantiation of the opening sentence, ‘‘Several times recently
we have been told that the means of a character in a series of
pure lines form a ‘Quetelet’s curve.’ °? I based this on his state-
ment concerning pure lines, ‘‘They not only do not differ in
their characters as the @nothera mutants do, but their norms
present a regular series coming under ‘Quetelet’s Law’ ’’ (AMER.
Nar., Vol. 44, p. 760). Surely no injustice has been done so far.
Later in the paper I did make a statement (which still holds
true) remotely similar to the one quoted above, and said specific-
ally ‘‘A case in point is a paper by Roemer.’’ There was no
reference whatever to Dr. Spillman—expressed, suggested, in-
sinuated, intimated, implied, . . . or intended.

2. Although it is clearly without any justification, the fore-
going criticism would, it seems to me, have gained in strength by —
specificity and moderation of statement. But Dr. Spillman con- —

No. 548] SHORTER ARTICLES AND DISCUSSION 499

tinues: ‘‘I have not been able to find time to look up other similar
citations to see whether the same inaccuracy applies to them.”’
There are four ‘‘similar citations’’: one to a paper Dr. Spillman
has already read, or at least reviewed for the AMERICAN NAT-
uRALIST (Vol. 44, p. 761), one to a few lines in the German Zeit-
schrift for genetics, one to thirteen lines in the AMERICAN NAT-
URALIST (Vol. 45, p. 423) reviewing the fourth, which has again
been considered in these pages (AMER. Nart., Vol. 45, pp. 686-
700)

3. Dr. Spillman reiterates: ‘‘It is now fairly well established
that the norms of a group of related genotypes can, in some cases
at least, be arranged in a frequency curve.’’ Thus, he tells us,
genotypic differences fall under de Vries’s category of fluctu-
ating variation, while in discontinuous variation ‘‘the norms can
not be thus arranged.”’

Personally, I have not the slightest prejudice against these
conclusions, but I can not accept them without proof. Dr. Spill-
man cites none. So far as I have been able to ascertain there is
not a single series of trustworthy quantitative data in support
of these pregnant generalizations.

4. I acknowledge my fault in omitting homozygous. This was
a serious blunder on my part! By including it, one can always
reason in a circle and to prove his preconceptions assume that
the original ancestors of a line were or were not homozygous,
according to the outcome of his experiments. This is the loop-
hole through which the supple genotypist can always crawl when
the evidence on the other side gets a little too strong. Again, in
the genotypic ritual, ‘‘I wish to publicly repent.’’

5. But is not the reviewer a little over-zealous when he con-
tinues, “the definition is further inaccurate in including clonal
varieties under the definition of genotype’’? In doing this I
merely followed the example of the best specialists. My paper
ein correct in terminology, as I believe it still is in facts, when
"4 Toni to press. Several months after the paper Dr. Spillman
n criticizing appeared, changes in terminology were forced by

the dictator of the whilom orthodox”? genotypic ‘‘school’’!

J. ARTHUR

NOTES AND LITERATURE
DISTRIBUTION AND ORIGIN OF LIFE IN AMERICA!

-The aim and scope of Scharff’s very welcome book is well
stated in his own closing words (p. 435): ‘‘I have endeavored
in this work to show how the gradual evolution of our continents
and the former changes of land and water can be demonstrated
by a study of the geographical distribution of living animals
and plants. Whenever possible I have taken advantage of our
paleontological and geological knowledge in furtherance of this
object, and I venture to think that I have succeeded in unravel-
ing some intricate problems of the paleogeography of America.
Indirectly I have thus been able to indicate the manner in
which North and South America became populated and the ex-
tent to which these continents took part in supplying animals and
plants to other regions of the world.’’ Scharff has certainly
done all this and much more besides. Never before has a book
upon zoogeography appeared culling and collating the thoughts
and observations of such a host of investigators. The fact that
his evidence comes not only from practically every group of the
animal kingdom, but very often from plants as well, make the
deductions far more convincing, radical as they often are, than
they would ever be otherwise. Extensive as is the bibliography,
there are still a few unfortunate omissions, and some typograph-
ical errors mar an otherwise excellent piece off press-work. In
the second edition, which can not but appear, the use of bee
would wisely be omitted. The words ‘‘Professor,’’ ‘‘Dr.’ and
‘‘Mr.’’ are used in a rather promiscuous manner. They are not
always judiciously bestowed.

The fifteen chapters take up the study of the fauna of the
hemisphere, beginning af the north with Greenland and proceed-
ing southward.

In Chapter I the relation of the hoii of Greenland to both
Europe and America is convincingly dealt with and Scharft’s
first excellent map shows the Pliocene bridge which extended
from Great Britain through the Orkneys and Iceland to Green-
land and on across to Arctic America. In this chapter the ee

*«*Distribution and Origin of Life in America.’’ By R. F.
New York, The Macmillan Co., 1912. Pp. viii + 497, 21 maps.

500

No. 548] NOTES AND LITERATURE 501

tion ‘“Did animals survive the ice age?” is first mooted and
Scharff prepares us for what is perhaps the most welcome and
far-reaching opinion which he, we think successfully, endeavors
to prove. His belief is that we have been accustomed to en-
tirely misjudge the extent and importance of the glacial epoch
and to exaggerate excessively the part which it has played in
influencing the distribution of our present flora and fauna, In
this the, reviewer heartily agrees and recalls that in Percival
Lowell’s ‘‘ Evolution of Worlds’’ the question is attacked from an
entirely different standpoint and a similar conclusion reached.
Scharff concludes that ‘‘The prevalent geological opinions as to
the nature of the Ice Age thus dominate all biological thought
in reference to problems of distribution.’? If we emancipate
ourselves from these preconceived notions in our speculations
on the origin (for example) of the existing fresh-water mussel
fauna, we must arrive at different conclusions.

In Chapter II, then, we have discussed the general features
of the fauna of northeastern North America with special refer-
ence to this theory of a comparatively insignificant Ice Age.

The third chapter deals with the animals of the Canadian
northwest. Ptarmigans, lemmings and gophers, the bison,
wapiti deer, our tree porcupine, ete., are discussed and their
origin pointed out. There is, however, no reason for conelud-
ing, as Scharff does, that the American magpie is more similar
to the European form than the Asiatic. All these palæarctic
races of magpies are so closely related and the differences be-
tween them are so slight as not to permit of any such conclu-
sions as Scharff draws from them. The other chapters on North
America deal with the fauna of the Rocky Mountains, the ani-
mals of the eastern states, the fauna of the Continental basin
and of the southeastern states and Bermuda. In this chapter
We meet again with an opinion which, while it can not but gain
ground as time goes on, now seems radical in the extreme. Ber-
muda has always been considered a most typical ‘‘oceanic is-
land,’’ yet here we have the view advanced, and well defended,
that Bermuda is the remnant of an ancient belt of land which
joined it to a southern land mass extending across the Atlantic
Ocean. This would account for the presence of both European
and American elements in the indigenous fauna, ik which,
as Scharff shows, is composed wholly of types bearing ‘‘the im-
Press of vast antiquity.’’

Tn the following chapter (No. IX) we have more detailed oa

502 THE AMERICAN NATURALIST [Vou. XLVI

cussion of this bridge between the old and new worlds. In spite
of the fact that there are still a considerable number of natur-
alists who adhere to the old dicta regarding the permanence of
land forms and ocean basins, there can be but little doubt that
the open-minded student will be convinced by what Scharff has
to say regarding the absolute necessity of postulating extensive
changes in the forms of the continents to account for the present
distribution of animals and plants. :

Next follows a discussion of Central America and the West
Indies (Chapters X and XI). The fauna of both these regions
is carefully analyzed and the very many seeming anomalies of
discontinuous distribution are explained, often for the first time,
by postulating a series of geographical changes far too elaborate
to attempt to summarize in a short review. Suffice it only to
say that the treatment of the Lesser Antilles is disappointing
and the important part which they have played in joining
Antillea (sometimes also spelled Antillia), if perhaps only for
a very short time, with that region which is now northeastern
South America, is overlooked or receives but scant consideration.
The existence of Onycophoran types on many of the Lesser An-
tilles should have suggested further investigation.

The twelfth chapter deals wholly with the origin and relation-
ship of the fauna of the Galapagos Islands. These islands are
also considered to be the remnant of a former considerable ex-
tent of land which for a time was connected with a land mass
extending from north to south along what is now the west coast
of both Americas, but at some distance out in the Pacific Ocean.
This connection of North with South America is certainly nec-
essary to explain a host of otherwise inexplicable distributions.
Parts of the present territory of Central America are known to
have been submerged up to comparatively recent geologic times.
Baur’s pioneer theories regarding the Galapagos Islands here
receive the appreciation which they have always so richly de-
served. They are corroborated by evidence recently accumu-
lated, all of which is used by Scharff to fine advantage. After

some discussion favoring Scharff’s theory, which presumes that.

there has been an extensive land mass occupying a considerable
part of what is now the Pacific Ocean west of the present land
areas, the author passes to a discussion of the South American
fauna, which fills the rest of the book. (Chapters XI-XV;
pp. 336-435.) In this the views of von Ihering and Ameghino
carry special weight, and to this they are certainly fully entitled.

No. 548] -= NOTES AND LITERATURE 503

Every one has previously been inclined to belittle the splendid
energy and self-sacrificing zeal which have stood back of Ame-
ghino’s sensational accounts of his discoveries. To any one who
has known Ameghino and who has heard him give his own
reasons for his conclusions and describe his own treasures, this
kindly appreciation by Scharff will seem well merited. This
does not, of course, mean that we must accept all of Ameghino’s
theories, especially those regarding the origin of man.

review, in the true sense, of Scharff’s book is impossible!
Each page is replete with valuable data, well digested, and is
pregnant with suggestion. All of the critics of Scharff’s previ-
ous works will agree that they are certainly suggestive in the
extreme, even if their postulates may not be accepted. . It seems
caviling indeed to conclude this notice with a list of little faults,
yet they are very evident, some of them, and may be corrected
easily in subsequent editions. Thus on page 20 we should have
Carabus nemoralis, instead of memoralis, a species which is not
confined to Nova Scotia, but which for years has been the com-
monest large Carabid about Boston and Cambridge. On page
141, dealing with the pine-barren flora and its northward pro-
longation, we find no mention of the most important contributions
which have ever appeared regarding the relationship of the flora
of Newfoundland with the coast regions further south, those of
Merritt L. Fernald, while, besides, no mention is made of the
writings of Witmer Stone or Harshberger, both well known in
connection with their work on the pine barrens. Again on page
151 we are surprised to learn that raccoons breed well in con-
finement and also to find no mention of the species of Procyon
described by Miller from the French West Indian Island of
Guadeloupe. The knowledge of this fact would have been of
great interest in connection with the occurrence of Procyon may-
nardi on New Providence, which Scharff admits is an enigma,
and the other remarks on the dispersal of raccoons. On page 173
we read north Carolina, elsewhere correctly North Carolina.
On page 180 we find Crocodilus americanus spoken of as the only
West Indian species in the genus, the important Crocodilus
rhombifer being passed by. On page 204 no mention is made
of Boulenger’s discovery of Bombina maxima from Yunnan, a
fact which is most important, fulfilling the prediction made by
Stejneger that a discoglossoid toad would be found in this area,
the center of dispersal of the group. On page 266 we find Por-
torico, one word; on the map it is given correctly as Puerto

504 THE AMERICAN NATURALIST [Vou. XLVI

Rico. On page 281 Saurecia and Panolopus are mentioned; the
former is hardly entitled to generic rank, a fact important in this
connection, while the genus Panolopus Cope has been shown by
Garman to have been based upon a specimen artificially muti-
lated. On page 282, in speaking of Capromys no mention is
made of C. ingrahami, a peculiar species long known from one
of the Plana Cays in the southern Bahamas. On page 289
Todite should read Todide, while on page 291 it would have been
worth while mentioning the fact that de la Torre, of Havana,
has published the account of finding fossil ammonites of Jurassi¢
age in the Sierra de Vinales, western Cuba. Such trifling errors
and omissions do little to mar the book! Its general excellence
carries it.far beyond petty criticisms. While the views which
Scharff expresses will doubtless meet with opposition from many
naturalists of the ‘‘old school,’’ nevertheless they represent those
which have been gaining ground fast and which will in time be
held by all zoogeographers. As James Bryce’s ‘‘ American Com-
monwealth’’ came from England, so indeed does Scharff’s Amer-
ican Animals coming from Ireland stand as by far the most im-
portant contribution to a knowledge of the subject it discusses.
Indeed, it is likely to remain so for many a long day.
T. BARBOUR

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.
= Articles containing research work bearing on the problems of organic evolu-
tion are especially TIAR and will be given preference in publication.
z One hundred reprints of contributions are supplied to thers free of charge.
Further reprints will be supplied at cost. se
and advertisements should be sent to the Ba gota The
subscription Moon tomey ie Oe a year. Foreign postage

s fifty cents and
= postage cents additional. The Sharan. for single n ia is
; forty cents. Tr aakri ec Tone Dollars for a page.

-... THE SCIENCE PRESS
= Lancaster, Pa. Garrison, N. Y.
pe oe ea NEW YORK: Sub-Station 84

Apri F2, 1908,

, Pa., under the Act of
(of March 3, 1879.

TENTH EDITION.
THE MICROSCOPE,
an introduction

ethods and to Histol
by Sox Haney GAGE, of A aaa University. Gre

aa vo pagis on on ahd o auren a this new

‘Comstock PUBLISHING C Sa, a meee N. Y.

Archives of Psychology
RECENT NUMBERS
“19s An Empirical tye of Certain Tests for

gf appears WHIT-

LEY. Pp. 146. $1.25 cloth, $1.50
Fi tests of memory,

association i ee dae pw ‘nih indications

20. Visual Acuity with Lights of Different

=... | or Davip Epear RICE.
= ©- =P oe: 50 cents ; ath, » 75 cents.

a = te oe ‘ ‘uncolored illumination and of the de-

sensory
sensori motor experiences m gm
ie et and material, and

nce of Caffein on Mental and

H. L. Honnivewours. P ste

THE
AMERICAN NATURALIST

Vout. XLVI September, 1912 No. 549

ASYMMETRIC COLOR RESEMBLANCE IN THE
GUINEA PIG!

PROFESSOR JOSEPH H. KASTLE AND G. D. BUCKNER

As is well known, the guinea pig shows the greatest
diversity of color, and a great diversity in distribution of
: color over the body of the animal. In the course of our
experience with these animals in physiological and toxi-
ecological work we have seen pigs that were entirely black,
others that were entirely brown, others of a pure albino
variety, and more commonly than any of these pure color
strains, those showing apparently every possible varia-
tion in the arrangement and distribution of these funda-
mental colors over the body. To what extent the color of
the guinea pig and its variations as the result of the cross
breeding of several strains of different colors have been
the subject of exact scientific observation, we are unable
to say, and the subject is so far removed from those ordi-
narily engaging our attention that it would take us too
long to familiarize ourselves with this knowledge, partic-
ularly should it prove in any way extensive. Inthe course
of some of our recent investigations, however, we have
observed what seems to us to be a rather remarkable case
of asymmetric color resemblance and distribution of
color on the part of the daughter for the mother in the
guinea pig, which is perhaps worthy of note to those more
deeply interested in matters of this kind.

One of our female guinea pigs which has been under
observation and experimentation for some time, No. 68,
was aborted by means of calcium lactate? on March 29,

1 From the Laboratory of the Kentucky Agricultural Experiment Station.

? Kastle and Healy, ‘‘Caleium Salts and the Onset of Labor,’’ Jour. of
Infectious Diseases, Vol. X, 1912, 378-382.

505

506 THE AMERICAN NATURALIST [ Vou. XLVI

1912. No bad effects resulted from the abortion and the
pig soon regained her normal condition and was returned
to the piggery on April 9, 1912. Shortly after this she
again became pregnant and during the greater part of
her pregnancy she was kept in a small cage with another
female pig as one of a set of pigs employed in the study
of the calcium metabolism of the guinea pig.® At the con-
clusion of these observations she was again returned to
the piggery on June 13, 1912, and on June 24, 1912, she
was again brought back to the laboratory. During the
day she gave normal birth to three pigs, weighing respect-
ively 40, 40.5 and 43.5 grams. One of these young pigs
was dead when examined a few hours after birth, the
other two were alive. Of the latter, one was paralyzed
in its hind quarters and died, probably of inanition, a
few days after birth. The remaining pig, a female, was
normal in every respect and is alive and well to-day
(August 5, 1912), and now weighs 157 grams. It was
observed by Dr. Buckner that the color markings on the
young pig are like those on the mother, except that these
markings are on exactly opposite positions on the body.
In other words, these two pigs, mother and daughter,
show an asymmetric color resemblance. That such is the
case is evident from the photographs, Figs, 1, 2, 3 and 4,
although these fail to show this as well as the originals
for the reason that the actual colors are wanting. Fur-
thermore, this remarkable resemblance is a little obscured
and marred by the fact that the young pig has consider-
ably longer hair than the mother, probably as the result
of an Angora strain. Unfortunately, too, nothing is
known regarding the parentage of this pig on the male
side. The following is an exact description of the two
pigs, which in order to render comparisons more easy,
is printed in double columns and given for opposite por-
tions of the body, in order to bring out the asymmetric
character of the resemblance. :

MOTHER Pig Youne Pig (FEMALE)
The mother pig is 9 inches long The young pig is 6 inches long, and
from tip of tail to tip of nose and 24 inches wide ecross widest part

*Kastle, Healy and Shedd, ‘‘Caleium in its Relation to Anaphylaxis’’
(in press),

No. 549]

: 8% inches in width across widest
part of body, Weight 570 grams.
(August 5, 1912.)

Right cheek and area surrounding
right eye, light tan.
Mouth, nose and middle of forehead
white, the white area on forehead
hak has

narrowing somewhat ju ehin
the middle ears then
broadenin ut over shoulders

back.

Right ear an intimate mixture of
` black and tan, giving the impres-
sion at first glance of black.

The greater part of the right side is
almost pure light tan, darkening

shoulder to a point where the front
leg joins the body

The right front leg is light tan with
the exception of the upper part of
the foot which is white.

The inside of right ankle and foot
is nearly hairless and shows a
well-defined black spot in the skin.

Left fore foot, left fore leg and left

middle and rear portion of back.

Under part of mouth and under jaw,

a down over left fore leg
ite. Th

right side = the chest which is
very light ta

Left ear is nearly hairless and white,

ASYMMETRIC COLOR RESEMBLANCE

507

of body. Weight
(August 5, 1912.)

157 grams.

Left cheek and area surrounding
left eye, light ta

Mouth, nose and middle of forehead
hi he

over shoulders, eia aaraa
somewhat over upper part of back.

Left ear is covered especially to-

with S
amount of white hair PT
in front and under
- The greater part of the left side of
this pig from a point about the

like a white collar on the left side
of the pig near the raei and
which is broader and m

developed than in the si

The left front leg is light tan ex-
tending ay to foot, the top of
which is

The inside of the left ankle and foot
is nearly hairless and shows a well-
defined black spot in the skin.

Right fore foot, right fore i and

right side, ‘incinding right hind
leg entirely white, with the a
tion of a small area of light tan,
over the middle and rear portion
of back

Under part of mouth and under jaw,

side of the chest which is light
tan

.

Right ear is

nearly hairless and

508 THE AMERICAN NATURALIST [ Vou. XLVI

with a small area of tan extend- white, with a small area of tan
ing over the front of ear. extending over the front of ear.
Rump entirely white. Rump entirely white.
Eyes black. Eyes black.

Left cheek and area surrounding left Right cheek and area surrounding
eye is tan, extending to left ear. right eye is tan, extending back to
ear.

It will be observed — a careful examination of the
photographs, Figs. 1, 2, 3 and 4, that these pigs show
certain minor differences in color distribution over the
body. These are due in part at least to the fact that the

Fie. L

hair of the young pig is considerably longer than that of
the mother—the latter being a smooth haired pig, whereas
the former shows an Angora strain. Thus it will be seen
from Fig. 1, that the small patch of white hair over the
left eye of the young pig is larger than the corresponding
patch of white over the right eye of the mother, other-
wise the asymmetrie resemblances shown in the front
view of the two pigs is essentially perfect, barring the
somewhat longer hair of the young pig. In Fig. 2, the
white band extending around the left fore shoulder of the
young pig is decidedly wider and more evenly distributed
than the white marking over the right shoulder of the
mother. This is doubtless due in part to the fact that the
hair of the young pig is longer and consequently overlaps

No. 549] ASYMMETRIC COLOR RESEMBLANCE 509

the tan-colored hair immediately back of the white area.
It will also be observed that in the young pig there is a
narrow strip of white hair surrounding the lower margin
of the left ear, whereas in the mother such a marking is
not shown in the corresponding area of the right ear.

Fig. 2.

Here again this difference in the photograph, Fig. 2, is
accentuated by reason of the longer hair of the young
animal.

In Fig. 3 we certainly have a beautiful illustration of
the asymmetric distribution of color in these two guinea
pigs, the only points of difference being a very small area
of white hair immediately under the left ear of the
mother that is not apparent under the right ear of the
young pig; and also a carrying forward of the tan area
nearer to the mouth on the right cheek of the young pig
than on the left cheek of the mother.

‘ig. 4 also serves to show the asymmetric color resem-
blance in these pigs, somewhat imperfectly, however, on
account of the partial disarrangement of a portion of the
long tan-colored hair of the small pig. Hence we note, in
Fig. 4, a slight extension of the white area into the darker
area on the back of the small pig. As a matter of fact,
however, a careful examination of the animals showed a
perfect asymmetric resemblance so far as distribution of
color is concerned, when looked at from the back. This
could doubtless be brought out by other photographs, but

510 THE AMERICAN NATURALIST [ Vou. XLVI

it is difficult to obtain good photographs of guinea pigs

in all positions on account of their nervousness.
Despite these minor differences in color distribution

over the bodies of the young pig and its mother, there

can be no question that we have here a remarkable case of
asymmetric color resemblance between this female guinea
pig and one of her offspring. It would seem further that
this asymmetric resemblance and distribution of color
in the young pig as compared with the mother is a char-

Unfortunately, as
has already been pointed out, nothing is known as to the

No. 549] ASYMMETRIC COLOR RESEMBLANCE 511

parentage of the young pig on the male side, and no
record was kept of the color of the other two pigs of this
litter, for the reason that this peculiar resemblance be-
tween this pig and the mother had not been observed at
the time that the other two pigs of the same litter died.
Our recollection is, however, that the pig of this litter
that was born dead was pure tan, whereas the one whose
hind quarters were paralyzed and that died a few days
after birth was white with tan and black markings. So
far as we can recall, however, these markings were al-
together different in arrangement and distribution from
those of the mother.

To the chemist the matter of asymmetry as affecting
the physical and chemical properties of certain chemical
compounds, especially those containing carbon, and as
applied to the constitution of such compounds, has since
the memorable researches of Pasteur on the tartaric
acids and the later work of Le Bel and Van’t Hoff, been a
particularly fruitful field for observation and research.
It is also a matter of interest to observe in this connection
that the character of the asymmetry shown by certain
compounds of carbon greatly influences their assimi-
lability by the lower plants. To what extent asymmetric
conditions hold in the ovum and germ-cell we have no
means of determining at present, and as already indi-
cated in the foregoing, it is a subject which takes us too
far afield from matters ordinarily engaging our attention.
We have reason to believe, however, that asymmetric
color resemblance in animals such as has been described
in the foregoing, is rare and from the point of view of
the chemist, extremely interesting and suggestive, and
affords a subject which, in our opinion, would probably
repay a more extended study on the part of those inter-
ested in animal breeding and the study of inherited char-
acteristics.

In conclusion, we desire to express our thanks to Mr.
T. R. Bryant, of the Station Staff, for his kindness in
making the photographs used in the illustration of this
article.

ON DIFFERENTIAL MORTALITY WITH RESPECT
TO SEED WEIGHT OCCURRING IN FIELD
CULTURES OF PHASEOLUS VULGARIS

DR. J. ARTHUR HARRIS

- CARNEGIE INSTITUTION OF WASHINGTON

INTRODUCTORY REMARKS

In the rather voluminous literature of seed testing,
comparatively little attention has been given to the pos-
sible relationship between the characteristics of the seed
(or of the plant from which it was gathered) and its
viability. This is of course attributable to the fact that
such work has been done chiefly for immediately prac-
tical ends, the object being in most cases to determine,
by germination tests of a small sample, the suitability of
a given bulk of seed for commercial planting.

To the student of natural selection, however, the car-
dinal problem of viability is to determine whether the
capacity for development of a seed is a function of its
patent or potential (i. e., of its own measurable or of its
inherent but as yet undeveloped) characteristics. A
satisfactory solution of this very complicated problem
would, I believe, be of rather wide interest. At the out-
set, however, one must fix clearly in mind that if a selec-.
tive mortality be demonstrated, it has no necessary bear-
ing upon the question of the origin of species. Natural
selection may maintain a type already differentiated as
well as mould new forms but in order to do either the
variations upon which it acts must be heritable. But in
any case the results would be of interest to the physiol-
ogist concerned with the problems of the relationship
between form and function. Finally, exact information
on the relationship between structural characteristics
and viability—if they exist—may be of some practical
importance in agriculture and plant breeding.

512

No. 549] ON DIFFERENTIAL MORTALITY 513

The purpose of this paper is to present the first results
of a series of studies on the relationship between the
structural characteristics of the parent plant or of the
seed itself and its viability. The data here recorded
relate only to seed weight and are drawn from an exten-
sive series of field plantings of carefully selected and
individually weighed seeds of the common bean,
Phaseolus vulgaris. They are properly described as a
by-product, for the experiments were not carried out
especially nor in the most satisfactory manner to test the
existence of a selective mortality.

The conditions in field cultures of individually labelled
seeds are such that many factors besides the weight of
the seed are concerned in determining whether or not a
seed shall develop into a mature plant. Some are lost by
dashing rains separating seeds and labels, after which all
questionable cases have to be thrown away. Some are
destroyed by rodents and some by the unavoidable acci-
dents of cultivation. In natural selection terminology,
the non-selective death rate—the death rate which is no
function of the characteristics of the individual—is very
high. This tends to obseure the selective death rate, if
such exists.

For just these reasons, I have never taken account of
the characteristics of the seeds which failed to develop
into mature plants, although the desirability of testing
for the existence of a selective mortality for seed weight
has been in mind almost from the beginning of the breed-
ing experiments with garden beans in 1907. The records
of those which developed to maturity were available for
studies of heredity, influence of size of seed planted on
characteristics of the plants produced, and so on. Data
for the entire parental population from which the seeds
planted were drawn were at hand for the study of the
influences of season and environment. Under these cir-
cumstances, the only need for a record of the seeds which
failed to develop to maturity would be for testing the
hypothesis of the existence of a selective death rate. As

514 THE AMERICAN NATURALIST [ Vou. XLVI

suggested above, it seemed a priori improbable that a
selective mortality could be detected in the large non-
selective death rate of field cultures. But from an exam-
ination of the data which have accumulated during the
last several years, it appears that the a priori scepticism
which led to omitting records of the characteristics of the
seeds failing to develop was unjustified. There appears
to be, in fact, a selective mortality detectible by proper
methods, even in field cultures.

The evidence upon which this statement is based is of
the following kind. We know (a) the weight of all the
seeds weighed in any year and (b) the weights of the
sub-sample of seeds which developed into mature fertile
plants in any subsequent year. Now the sample or
samples which were planted were purely random draw-
ings from the grand population forming the entire mass
of seeds weighed for any culture. The physical con-
stants for the distribution of weights of these sub-samples
planted are, therefore, identical with those of the grand
population, plus or minus the errors of random sampling.
Ideally, to determine whether there be a differential mor-
tality, we should compare the physical constants (means,
standard deviations, and coefficients of variation) of the
seeds which produced mature plants either with the con-
stants of those of the same sub-samples which failed to
do so or with the constants for the entire sub-sample
planted. Practically, our end can be attained with rea-
sonable exactness by comparing the constants of seeds
which develop to maturity with those of the general popu-
lation from which the plantings were drawn. The only
objections to this procedure are two: (a) the results will
be vitiated if the series of seeds planted are in any Way
selected, i. e., not a purely random drawing from the
general population, (b) the probable errors of random
sampling in the drawing of the sample for planting are
added to the probable errors of sampling due to the non-
selective (purely random) death rate.

The first objection does not hold in the series discussed

No. 549] ON DIFFERENTIAL MORTALITY 515

here, for I am confident that the plantings were true
random samples of the general population of seeds
weighed. With regard to the second, we note merely that
its influence will make the detection of a selective mor-
tality more difficult. If we find indications that the
chances for development of a seed are conditioned by its
weight we must therefore consider that the influence of
weight is probably even stronger than is indicated by our
evidence. Indeed, there might be a selective death rate
scarcely detectible by the methods necessarily used in
this study, but if these methods do indicate a selective
elimination, we may have considerable confidence in its
reality.

Studies of viability are generally limited to the capa-
city for forming a growing seedling, but there is no rea-
son why tests should not be made more stringent by
extending them to the capacity of the embryo for devel-
oping into a fertile plant. This has been done in these
experiments.

PRESENTATION OF DATA

The seriations of weights of seeds for the general
populations are given in Table I; those for the sub-series
of seeds which actually developed into fertile plants in
Table II. The seriations for the grand populations are
designated by the key letters of the crops of plants which
produced them, those for the viable sub-samples by the
key letters of the crops into which they developed.'' The
weights are recorded in units of .025 gram range; class 1
being 0-025 gram, class 2, .025-.050 gram, ete. The
constants are also expressed in terms of these units, but
any one desiring to do so may easily transmute them into
fractions of grams.

The biometric constants for the grand populations are
given in Table III. Those for the sub-samples which
actually produced fertile plants appear in Table IV.

* These key letters are the same as those used in other papers published or

in preparation. Hence, further information can be obtained by tho:
who desire.

THE AMERICAN NATURALIST — [Vou. XLVI

516

t9sz |—|—|—_|—_|—_|—|—|T |t j6 [et lsz: J69 |eatizez poe soe TOF GEE 9EE 216 FFT 96 18 fot ge ts a Iti- p
ole ~ma g A je og go ite \22t +61 \Pezisoeizzesezizetisz joe jet a ie it ft |—|—I'::aoo
8286 ——— |e | |€ | OT) TL) 98 |49 |PIT/TOZ/Z08|Fzslezh/ TOF joge ezziezits tr [FT |T |e | —|—|— it }|—|—|—|'*"HDD.
yogs —|—\—|I je |9 | #1] 88/08 TOT 097 068 Z19782 £26 9TOT CLOT 616 689 £9 ESE OZZ LET TL 0G ee le ee
466 1 2/2/61) 1%| 9% | ZS) 19| $8 62 ZOT6at SIT eZ A l e 6 ig SI FI 9 Te i a ao bees age ore ers O Ð
6861 ————]——i—i—|—|—|—|—— — miim e e e Ue es Og ert ST | + ‘ASA
eee iert e l ea a | — |-—|—| |e Jort 110 92e 860 Shs Coz ce ISI | 1 | t | "Hed
D008 ee es ee ee ee Sa
cool |---| —| —|—|—|—I—|—|—|—|-1-\s. i oe ot foo te ara ven Gor ae ge mee an
COLL Emas = mT Pe BL er or eo - ee EOT TOT Z1Z 617 S91 06 s ge i~r Tio a
1688 imie il 8 a aT SS ee eA os Fer SST TPE) TE¥ 877 £78691 WO T |e
met BY YG AG Ge Nah, ps oe! Sa —|T JOT 0G |b [29 |946 |E88 [cee Z4 |8 |I |1| HAN
oes i—mar ee ost ieee ees eon ize aines (611 ~] GaN
z9gez ————] — | — \—|—-}—|—|—} —}-—] —i— —|—|—|—|—|—|$ |e jeer ATE Ra jeez BELEGO sz} T aN
el TE HG Pal BR ts S ie se ae ere ene tag oan ozo E — | E 8 28 BIT LTH |966 \ZELT/992T)Z601 80F/Z6 | ET) € |” CHN
reeL | — | — [ef em ee ee a e a — J€ [8 |F% |69 SLT 6b SZ1/120G/E961|\686 |193 |28 |S —| HHN
8429 BARE — =] —|—|—|}—|—|— = —|—| — | — |— ITI leg itor G96 628 G9ET/ZOST|6FOT T98 |48 [FT Z zp HN
aao oz |3 za |07 oe va | ez | zz | tē | oz | 6t | st | it lor er vr er ar) tr | or | 6 | 8 | i ofelr is sonog

S190. Gi

NOILV'INJOJŢ IVYANTH AOA SATAS AO LAHVITM

I TIaAVL

517

No. 549]

ON DIFFERENTIAL MORTALITY

IO fe ll a “| [9 oB ek 98 BOT|SET/ZET/S6 |99 |LE |SE [OT on ta oS 3 o TT
we oma aad tesla ame ee RAE bo ac te It jig +e jar jag ito ler ise jor ig je ie i t [—]—]—i ie aao
e [— —|— —|——|—iz |—|¥ |9 | tos los jos 94 les jeg zz jet let |—|—|t |—|—|—|—| —|—|— |" "HOD
ev |— —|—|—|— —|—|/z |t |s | or] 2t/t¢ ise lso jes |9 |t} les isg ire lat 8 fe (i |—|—|—}—|— qD
rig |— —|—|—|—|—|1 leje |z |6 | else i» o [eg pa Jeo e fee ire jot fe ie jf n Ti a99
66 _— —|— —|— 1 |1 |z |z |¢ |s | 61 (te jos los jos joz jog les [ze let 8 le læ j9 je |—|—|—\—j—| HDD
eso |—|—|—|—a—|—|1 £ g |6 |zz|eslor oz jos ler jeż |s9 lor ise jeg jez js jo je jt |—|t |- =i] HDD
6oL | 1 | ¥| & |ST|9T| ST! eb| ¥9| ¥9| £9) 18/96 zs [zo [ze + loe zt FL le je t [ft fr |—|— eee DD
288 =s a ede ees eet ews Se Gs ig log jes jogtiezjor) z |t|- aasa
er =i Peel (pel Gee (aaa ees ae Gosliee ae inci eet eae ee ee ; ot lo9 \Tetlggtier (etje |t| HHSA
əs — —|— -| —— -| Baars ees nae et ) lte j2z |pgtisotise |or| ot} 1 | 1 poet Osu
seh — IJ |—|—|—|—Iz loz ler [zorisetisz oz) or) ejg | asa
GLY — — —|—|—)/—|—,—|—|—|-|-|-— —|-—|-|-|-|-|-) f (Fe Po |TETIOITIOS erie | 2)|—j\ HS4
a a a a ae bee =i ee eee JI IT/POT|THZIOAT| ZL) 1s} 9 |—| ssa
Be | —|—|— ||| = |— \—| = ll il —|t_ |e [ot jee [ve fto ler jte jetje |z |—|—|—\-"---aasa
e iiel o ee ee ere ee E HHSA
oss |— — —|—'— — | —|—|—|—|2 |z |s jor|ertl|tz isz joo loz jeorize izo izg jo + je |—jt itti- osn
ze |—|— | —|— |—|—|—}— i—|@ |t |* £ 8 |zt\6 (FI jet jeg jo9 lsg (sp e [= = =l ete asa
8 ie '—i/t lt |t lttlg ‘trier let jez lto 99 lez jem izg jor it ft tt |—|—j|—!—l HSA
o9 |—|— —|— —|— — —|— z |e |9 let! or|ezlet jpe ler loz jzttlpetisitize jst ls |—|—|—|—|—|—| sa
| a a a ee ee —|—|—|—Ie le lz» lostirotootzr |et|t |—| 1 °° HHAN
ese |—|—)—|— | —|—|—|-—|-—|—-|-—|-|-|-|-|- —|—|/—|—le lo laz |as \zetipetisy stl? | 2 |— | aaan
BIG ee fe ee ee ee e p |28 |tSt\ept| es og) e |—j aaN
og |—|— —— — — -i—i — llnl 7 e toreoziset 92/96; 8 | T ee HAN
post |—|— | —| — —|— -l-l -| -|-—| - —' — -|- — — —J |9 [ez jos \zct\sszjeteiogz| 08 | #1 | £ |g] adHN
et ee ee —lt |e |p |g lte |tet\soglgzeisoezi¢s | ot| —|—| —|°* * -HHHN
srt |—|— — — —|— — —|—)|— | — | -|-— -|- | —|—i¢ le lsz +6 otzoseltor 18296 |egzje | t —i HHN
wt tam tap ben ces ee sd ee ee —|—le |g lsz se |trzlgeziose|ozzies | st|¢ |—| t |: AHN
STRIOL | E8 | ZE te 08 | 62 8% | 28/98 /9s| FS S| Vz} 1S 0Z| 6I| ST] 2} 9T| ST] HL) ET | ST TT or eļs]|2 o|e|+|sg| seueg

SINVId WIM ONIONGONG SAMAS dO LHXIAM
II 44V L

518

TABLE III

BIOMETRIC CONSTANTS FOR GENERAL POPULATION OF SEEDS WEIGHED

THE AMERICAN NATURALIST

[Vou. XLVI

Mean wo het pene

Standard Deviation

Coefficient of Variation

; ES
eerie and Probable Error esa and Probable Error

MES ec oe | 9.474 =.013 1.490+.009 | 15.727 +.101
Iaa Ea Ss E 9.774 =.011 1.421+.008 | 14.537+.082
13 A eG E 8.529 = 012 1.458+.009 | 17.099+.103

Bee aoe 7.496 +.0 1.310.013 | 17.470 +.182
jhe 5 ee A eae 8.487 +.019 | 1.377 +.014 16.218 +.163
NOR e 8.852+.015 | 1.555 +.011 | 14.089 +.121
We cl ere nee | 2.605+.025 | 18.218+.184
ten... 04 +.04 2.413 =.034 | 16.987+.244
ee 10.690 =.041 | 1.921+.029 | 972 +.279
et 8.229+.018 | 1.434+.013 | 17.427+.156
DAE 8.516+.016 | 1.092 +.011 12.826 +.135
PD- Se 56+.016 | 1.034 +.011 14.858 +.161

Sea gg hae Ren een 22.272+.081 | 3.796+.057 17.044 +
OG. Go ae | 17.601 +.024 | 3.264 = .017 18.542 +
GCR eoo | 18.919+.034 | 2.674 =.024 14.131 +. 177
GGD. os 14.972.036 | 2.498 + .026 16.681 +.193

ee ae coe 15.667 + .04 | 3.137 + .028 20.023 +.186

TABLE IV

BIOMETRIC CONSTANTS FOR SEEDS WHICH PRODUCED FERTILE PLANTS

Series Mean and Probable Standard Deviation Coefficient of Variation
Error and Probable Error and Probable Error

NED oo 9.458 = .027 1.507 = .019 15.932 +.207
NEG be 9.445 + .026 1.508 + .019 15.962 + .203
NHS ooo 9.275 + .025 1.332 + .018 14.356 + .196
NAD 2 5 459 = .0 1.409 = .019 16.657 + .235
NDA ges 7.624 + .033 1.283 + .024 16.833 +.319
NDD n 7.733 =.039 1.318 + .028 17.049 +.
NÐDD -n 756 = 1.304 + .029 14.892 +.339
NOBH o 8.929 = .035 1.244 + .025 13.930 +.285
ti]. > POO Sete 14.143 = .070 2.713 + .050 19.180 + .364
AS E Soe Rea TET 14.330 = .094 650 =. 18.495 +.480
Wee ete 14.394 =.101 632 +. 18.283 +.510
pt) SSR Se doer rs 14.496 = .076 2.583 = .054 17.819 +.381
USHR oes 13.942 = .087 1.927 + .061 13.820 +.449
Usp 6 ee 10.852 = .078 TTT =. 16.371 +.521
iao Ee A 8.204 = .033 1.434 = .023 17.478 +

PH o o. 8.326 + .045 1.438 = .032 17.276 + .389
PSD a oe 8.224 + 047 1.430 = .033 17.383 +.413
PO oe ee 8.312 + .039 1.397 + .028 809 =.
Pane 2 8.483 + .035 1.075 = .025 668 =.
DE -o 7.194 = .036 1.045 =.025 14.520 +.359
GO es 22.324 = .091 3.727 = .065 16.695 = .297
GOR oe 17.659 + .094 3.359 + .066 19.020 +.389
GOM on. 17.305 = .093 3.079 = .066 17.792 +.392
GGD.. = og 17.545 =.095 3.188 = .067 18.168 =.395
COne R E 17.448 = .096 3.023 = .068 17.327 + 402
GONE... o 18.038 = .083 2.445 = .059 13.555 +.331
DIE oe 15.515 =.088 2.410 + .062 15.536 + .410

Beate ee as 15.524 + .059 +, 19.121 +.280_

hase 4)

No. 549] ON DIFFERENTIAL MORTALITY 519

in all cases. Only the deviations of the sub-samples from
the general population, i. e., the differences obtained by
Sheppard’s correction was appked to the second moment
subtracting constants for the general population from
those for the sub-samples producing fertile plants, as
given in Table V, require our attention.

Consider first the differences in mean seed weight.

They are equally divided between positive and nega-
tive, the average weight of the seeds which produced fer-
tile plants being in 14 cases higher and in 14 cases lower
than that of the general population of seeds from which
the plantings were made. The average value of the posi-

TABLE V

aena OF ee FOR SEEDS PRODUCING gms PLANTS WITH
OSE FOR THE GENERAL POPULATIONS WEIGH

s Mean and Probable | s andard Deviation | Coefficient of Variation
Series Compared Error | a and Probable Error | and Probable Error
NBD-NH....... —.016 =.030 | +.017+.021 | + .205+.230
NHH-NG......: —.030 = .030 +.018 +.021 ob 2a
: NHHH-NHH#.... — .500 = .028 — .089 + .020 | 181 =.212
NHDD-NHD.... — .070 +.030 — .049 + .021 — 442 +.257
NDH-ND....... +.128 + .038 .026 + .027 — +
3 8 A +.237 = .043 +.009 = .031 — 421+.412
NDDD-NDD + .269 + .045 — .073 + .032 — 1.326 +.376
NDHH-NDH +.078 =.038 —.312 +.027 — .159+.310
(SOUR TA —.154=+.025 +.108 + .056 962 + .408
UBH-US8.: o. +.033 =.101 +.046 +.071 277 + .514
Ms Pie eco +.098 = .107 +.027 +.076 + 542
cy A +.200 = .084 —.021 +.0.
USHI USH eine .262 + .099 — 486 + .070 3.167 + .511
USDD-USD..... +.163 =.088 —.145 + .062 — 1.601 =.591
FSS Pn.. EG — .025 +.037 .000 + .026 + .051 +.331
PSH-fS oe _ + .048 +.004 + .034 151 +.419
FSD-FS. 5 ee; 004 + .050 —.004 = .035 — .044 + 442
PSC FE o +.084 = .043 —.037 = .030 — .618 +.374
FSHH-FSH...... — .033 = .039 —.018 = .027 158 + .325
FSD Doon +.238 + .039 +.011 =.028 Bn
CO +.052 +.122 -. .086 ~
GH-GG........ +.057 = .097 +.095 = .068 + .478 +.402
GGH:-GG....... — .297 = .096 —.185 + .068 ~
GD-GG........ —.057 = .098 —.076 + .069 374
GGD:-GG....... 154 = .099 — .240 + .070 —1.215 + .414
GGHH-GGH.. 2+.089 228 + — 576 +.375
GGDD-GGD + .542 +.095 —.087 + .067 — 1.145 + .453
ce eas +0 —.168 + 36

tive differences is +.163 units; of the negative differences
—.188 units; of the entire series of 28 comparisons

520 THE AMERICAN NATURALIST [ Vou. XLVI

—.012 units. Considering the differences individually
in comparison with their probable errors? we note that if
we regard a difference at least 2.50 times its probable
error as statistically significant, 10 of the differences may

T
-===

se mm m OO

-----0
-------0

kh-------0

——
a

L

e.
a
Peed

DIAGRAM I
Differences in Means.

be looked upon as trustworthy. Of these 5 are positive
and 5 are negative. If we require Diff./E arr. — 4.00, we
find only 7 trustworthy cases, 4 positive and 3 negative m
sign.

Diagram I shows graphically the amount and the sign

*It is somewhat difficult to decide just what ratio of the difference to its
probable error should be used to indicate trustworthiness. As already
emphasized, we have in reality two instead of one random sampling to take
into consideration in the comparisons which we are drawing. The ratios
Dif'./Eaifs. are available to the reader who may assign to them any signifi-
acting he sees fit. The general trend of the series furnishes much stronge?
evidence for a selective mortality than does the apparent statistical trust-
worthiness of any individual comparison.

No. 549] ON DIFFERENTIAL MORTALITY 521

of the differences between the general and the viable
samples. Here the different comparisons are shown from
left to right in the order in which they are given from
top to bottom in Table V. The distances of the solid dots
below the zero bar indicate the amount of the negative
deviations, the circles above the positive deviations.

From all of these considerations, we conclude that
seeds which produce fertile plants are on the average
neither lighter nor heavier than random samples of the
population.

While the results for the means furnish no evidence
for a selective mortality, the standard deviations are very
suggestive of its existence. In 19 cases, the S.D. of the
seeds which produce fertile plants is lower than that of
the series from which they were drawn, while in 9 cases
it is larger. The deviation from the equality to be ex-
pected if the differences were due purely to an infinite
number of random sampling is therefore

5 +. 67449 v.5 X.5 X 28=—5+1.79,

which is perhaps statistically significant. The chances
against the result being due to the errors of sampling are
roughly the same as those against 19 heads and 9 tails, or
vice versa, in coin tossing.

Not only the inequality of the divisions of the signs
into positive and negative, but the magnitude of the
deviations themselves evidence for a selective mortality
which reduces variability without sensibly affecting the
average. The 9 positive deviations average -++.037; the
19 negative — .122; the whole 28, —.071. The far greater
magnitude of the negative deviations is made clear by the
generally greater lengths of the bars below the zero line
in diagram 2. Considering the individual differences in
their relation to their probable errors, we note that in 7
cases the difference is over 2.5 times its probable error.
All of these are negative in sign. -

The coefficient of variation, that is, 100¢/m, should show
most clearly whether both larger and smaller seeds fail

522 THE AMERICAN NATURALIST [ Vou. XLVI

to develop, thus bringing about a reduction in variability
which is independent of any change in the mean. The
results given in the final column of the comparison table
V, evidence even more strongly than the standard devia-
tions in favor of such a selective elimination. They are

-—-- == www ee

ey yap

a

Di
Differences in Standard Deviations.

lower for the seeds which develop into mature plants in
21 out of the 28 cases. Thisisa deviation from equality of
T + 67449 V.5X.5X28—7 + 1.79.
The averages are:
For positive deviations, + .325
For negative deviations, —.712
For all deviations, — .453
Diagram 3 makes the reason for these differences clear
to the eye. With regard to their probable errors, the sx

No. 549] ON DIFFERENTIAL MORTALITY 523

cases which are perhaps statistically significant, are all
negative in sign.

SUMMARY AND Discussion

Taken altogether the data seem to me to indicate a real
differential mortality in seeds of Phaseolus vulgaris in

o TI eet a Be I
z |

Differences in Coefficients of Variation.

field cultures. This selective death rate is of such a
nature that the mean of the viable seeds remains prac-
tically the same as that of the original populations while
their variability is reduced. In short, both large and
small seeds are less capable of developing into fertile

524 THE AMERICAN NATURALIST [ Vou. XLVI

plants than are those which do not deviate so widely
above or below the type.

While the evidences in support of these conclusions are
fairly strong, it must not be forgotten that we are dealing
with a problem of great delicacy upon materials grown
under conditions such that the accidental (purely non-
selective) death rate must be quite high, and with records
not especially collected for our present purpose. Posi-
tiveness of assertion must therefore be reserved until
more critical evidence collected ad hoc is available. The
reader will note, however, that the conclusions here
recorded are based on many thousands of observations.
Although experiments are already under way, the gather-
ing of the data necessary for a more thorough investiga-
tion of the problem will be a long task, and it seems only
right that the results obtained incidentally should be
placed on record for the benefit of those who may have
opportunities for like observations.

To me personally, the results were surprising. First,
I had doubted whether a selective mortality could be
detected by the methods used. Second, I had supposed
that if a differential viability were found, it would be
limited to a weeding out of the lighter seeds. This latter
presupposition was based on the fact that a positive cor-
relation had already been demonstrated? between the

weight of the seed planted and the number of pods pro- -

duced, and a priori it seemed reasonable to suppose that
viability and capacity for forming plants with large num-
bers of pods would bear the same relationship to the
weight of the seed. Both questions deserve much more
detailed and refined study.

Two questions concerning these results will be fore-
most in the mind of the biologist: (a) What is the signifi-
cance of the selective elimination for evolution? (b)
What are the underlying causes of the differential mor-
tality? ae

* Harris, J. Arthur, ‘‘On the Relationship between the Weight of the

Seed Planted and the Characteristics of the Plant Produced,’’ Biometrika,
Vol. 9, No, 1, 1912.

No. 549] ON DIFFERENTIAL MORTALITY 525

The answer to (a) depends, as has been pointed out
above, upon the inheritance or non-inheritance of the
variations in seed weight. This question can not be dis-
cussed as yet, but from the data in hand, it seems to me
likely that this selective elimination has little or no evo-
lutionary significance. Certainly it is, as far as our evi-
dence goes, not a cause of progressive change but only a
factor tending to preserve an established type. It would
then be an illustration of the ‘‘periodic selection’’ of
Pearson.

The answer to (b) must be sought in the physiological,
and of course ultimately, in the chemical and physical
properties of the seeds of different weights. Experi-
ments directed to its solution are under way.

COLD SPRING HARBOR, L. I.,
July 15, 1902

A CASE OF POLYMORPLISM IN ASPLANCHNA,
SIMULATING MUTATION. II

PROFESSOR J. H. POWERS,

UNIVERSITY OF NEBRASKA

I may next state some further observations which I
was able to verify again and again in regard to the trans-
ition of one form of the species to another. As I have
said, forms intermediate between the saccate and the
humped, between the humped and the campanulate, and
even between the saccate and the campanulate, occur.
This statement applies to external body form, and to
some-extent to the nephridia and other internal organs,
with the exception, however, so far as I have yet observed,
of the size of the contractile bladder. This latter seems
to have its large size only in the small saceate type, and
I have observed no indications of gradual transitions to
the form possessed by the larger rotifers. More sig-
nificant, however, is the case of the trophi. I have ex-
amined these in a great many individuals that in body
form were more or less intermediate between the dif-
ferent types; but in nearly every instance the variations
of this organ seem to be abrupt and discontinuous; the
trophi are either of one type or the other. The only
instances found that in any sense transgress this state-
ment were the trophi of a few saccates produced as the
result of the slow degeneration of the larger form, in the
culture which I have before mentioned. These animals
showed trophi that had plainly lost a number of the more
delicate characters which I had otherwise found universal
through such a wide range of material, and, simultane-
ously with this, they had become in a degree transitional
between the two types; the angle of the inner tooth and
a slight crossing of the tips plainly related them to the `
cannibalistic type, while in size and general form they

526

No. 549] A CASE OF POLYMORPHISM 527

belonged to the other. I will record here also two in-
stances in which I have found the cannibalistic trophi
in the humped rotifer. The two specimens were almost
the last humped individuals found in a culture on the
verge of extinction, through cannibalism; they were prob-
ably the progeny of campanulates. Both were large
specimens of the humped type, one showing a rather heavy
corona and rather small humps, and being, therefore, in
some sense, of a transitional character. It bore campan-
ulate trophi 2574 long—rather undersized—in which,
however, the large lamellate teeth of this type were even
unusually developed. The other bore typical campanu-
late trophi 270% long, These animals contained large
unborn humped rotifers of normal type, with normal
trophi of 154» and 170», respectively. In these two in-
stances, therefore, the transition from the campanulate
to the smaller type occurred one generation sooner in
general body form than it did with the trophi, thus em-
phasizing the partial separateness and non-correlation
between the variations in these differently formed struc-
tures.

Next as to duration of transition periods and the num-
ber of transitional individuals. If conditions are favor-
able the periods are very brief and the number of transi-
tional types so few that they are readily overlooked un-
less careful search is made at just the right time. The
entire population of a teeming Asplanchna pond readily
changes from the saccate to the humped type in one week.
As before said, the saccates give birth directly to forms
with well-developed humps, and these humped young may
be at birth as large or even larger than the parent type.
One more generation of growth and reproduction may
then give large-sized, fully typical humped individuals.
Along with these abrupt transitions there usually occur,
however, a lesser number that are a little more gradual.
Individuals occur like saccates in all respects save that
they possess the inconspicuous dorsal hump; others are
small with the lateral and ventral (posterior) humps

528 THE AMERICAN NATURALIST [ Vou. XLVI

scarcely showing at birth, but developing to a moderate `
extent rapidly afterward.

With less favorable conditions the transition is pro-
longed and the number of intermediate individuals is
greatly increased. I have observed no instance, how-
ever, in which the species remained for longer than two
weeks in a chaotic condition. Either the transition is
soon effected or the numbers rapidly decrease and the
species disappears

Nearly similar statements may be made with regard to
the transition from the humped to the campanulate type.
As already stated, its advent usually occurs by the ap-
pearance of a very few individuals with the utmost
abruptness. Aside from the fact that the ontogeny is
here somewhat more extended—they, the young, being
considerably smaller than the adults, with much less ex-
panded corona—there is apparently little, if any, sense of
transition. I think it probable that the humped indi-
viduals which actually give rise to the very first campan-
ulates are individuals of somewhat extra size and vigor.
Such individuals have been found at different times as
well as in the case of the two mentioned in my first inves-
tigation, but their actual production of the young canni-
bals has not been observed. In any case their deviation
from the ordinary humped type is not great and the usual
transition has all the abruptness that the most pro-
nounced mutationist could anticipate. Moreover, as long
as the species is thriving and reproduction copious, the
two forms remain separated from each other as sharply
as do the most distinct species. This is the most fre-
quent condition by far in which one finds them.

When, however, conditions become less favorable,
which fact usually means that the food supply of the
- humped form is failing, a change intervenes. The
humped individuals usually remain quite as they were,
without reduction in size or loss of other characteristics,
save a much slower rate of reproduction. But this re-
duces numbers, and especially the number of young.

No. 549] A CASE OF POLYMORPHISM 529

The cannibals can and do ingest their full-sized con-
geners, but they are by no means successful in every at-
tack. One may observe them, with empty stomachs,
making scores of furious but futile attempts at capturing
their adult neighbors. It is largely the young hump-
bearers which, though nearly full grown, fall ready vic-
tims to the all-embracing coronex of the cannibals. Thus
it follows that any reduction of the food supply of the
lesser type immediately impoverishes the larger one as
well. The consequences of this are curiously dissimilar
in different cases, although always one of two results in-
tervenes. The cannibals may become even more canni-
balistic, destroying the entire humped population of all
ages, and their own young as well, until the culture is
finally obliterated by the death, from old age, of a few
veterans which are without further food supply. This
has happened again and again in my large culture
dishes.” In a few cases it has happened that a culture,
when at the point of extinction, would again revive by the
multiplication of the humped form. This is due to the
fact that the last starving cannibals, reproducing, as
they always do, both their own type and the humped type
as well, fail to eat up perhaps a single member of their
humped progeny, which then survives to start a new cycle
under less strenuous surroundings. I have carefully fol-
lowed this decline and survival as thus stated.

In about half of the cases, however, a very different
effect is registered upon the campanulate form by the
lessening food supply and the falling numbers of the
other type. It undergoes a considerable degeneration,
which may perhaps reduce it to the form from which it
arose, although I have not been able to fully demonstrate
this. But forms more or less intermediate are produced
by the starving cannibals. The trophi remain typical,
but the enormous coronas, as well as the breadth of the
entire animal, are much reduced. In the single instance
already mentioned in which degeneration gradually re-
duced the humped type to a small size, and finally to the

"I have recently observed one very similar instance in nature.

530 THE AMERICAN NATURALIST [ Vou. XLVI

most diminutive saccate form, I was surprised to find
that the campanulates present in the culture degenerated,
pari passu, with the other type. They became much
smaller, lost their flaring coronas, and nearly every sign
of their outward specialization. Yet, generation after
generation, they maintained their cannibalistic habits,
their heavy musculature, and above all the campanulate
type of trophi, the only change in these latter organs
being a reduction in size.

In the main, then, transitional periods are brief ; transi-
tional forms few. Unfavorable conditions prolong some-
what the existence of both. But the species always is
soon eliminated or sets up a new equilibrium under the
new conditions.

A few words further may be added at this point upon
the matter of fluctuating variation shown by the different
forms of this species. Without recording such varia-
tions mathematically, I have endeavored to ascertain as
fully as possible the answers to three questions: First,
how great is the amount of such variation? Second, is
fluctuating variation especially correlated with one or
other of the types of heterogenesis above described?
And third, what causes are operative in producing it?

As to the amount of fluctuating variation, certain facts
have already been mentioned that come under this head-
ing. I will but add here the general statement that each
of the three types is, in itself, highly variable—quite suffi-
ciently so to be regarded as a decidedly variable species
were it really an independent form. |

As to the second point, the question of the correlation
between fluctuating variation and the mutation-like
transitions, this has also been partially discussed. under
the heading of transitional types. But it is necessary to
add the unqualified statement that no evidence has been
discovered for such correlation. Variation is one thing;
heterogenesis another. The two phenomena contrast,
rather than are related. Thus the transition from the
saccate to the humped rotifer is often made when the

No. 549] A CASE OF POLYMORPHISM - 531

saccate type is in its most typical condition, at least so
far as form is concerned. Of all the minor fluctuating
forms found among the saccates the one that most
suggests the humped type is that which I have character-
ized above as urn-shaped. The bulging sides of such a
form might readily be thought to be the forerunners of-
at least the lateral humps; but during the entire study I
have been unable to observe the humped form originating
from this urn-like variety.

Furthermore, the saltations from type to type do not
necessarily occur, and I think do not usually occur, when
the amount of fluctuating variability is greatest. This is
especially true of the formation of the campanulates from
the hump-bearers. This transition occurs when the
latter are at a culmination of development and vigor, and
in this condition the species is, until saltation occurs,
relatively uniform.

Under the third question, as to causes of fluctuating
variability, I will record at present but two points, one
general and one special.

The greatest amount of general fluctuation seems
always to occur under relatively unfavorable conditions.
Favorable conditions, on the other hand, tend to produce
full development with relative uniformity among the
individuals of any given type.

One special instance of variation interested me so
much that I followed it whenever found, in the effort to
get at its exact cause. This is the variation in the length
of the three conspicuous humps which characterize the
commoner form. The amount of this variation is,
although I have not measured it, very great. The humps
may be but angular projections upon the body’s outline,
or they may elongate until they might be appropriately
described as finger-shaped. The type represented in the
illustration, Fig. 1, may be taken as typical. This type is
repeated in countless numbers, with but moderate varia-
tion, so long as conditions are normal, which means
chiefly, so long as the food supply is uniform and ade-

532 THE AMERICAN NATURALIST [ Vou. XLVI

Fie. 2.

papieeonne amphora, OF THE HuMPED FORM (FORM B), Teor
PIC YOUNG VELOPING Pasonssoonsericaziy WITHIN THE Bopy OF THE
PARENT. Magnification ee 8 diame sit
Fic. 2. Asplanchna amphora, campanulate form (form C), showing hetero-
typic young developing ‘parthenogenetic within the body of the parent. This
figure is dra from ecim n because it exhibited well the hetero
typic teprotüstiok in| ther ei idee k Ta not represent the true form of the

nd contracted the cor na, obliterating thus pa bell-like form ring
the anterior portion highly convex instead of concave as in life. The convex
end of Fig. 1 is, on the contra ary, quite ia pos oT Bigs being much
more paged killed in perfect form. Ma gnification about 8 d eters.

-$ ed, ophi of humped and campanulate SSS respectively.
The ydh in size is Teeter shown by these figures. Had average speci-
mens been chosen, Fig. 4 would have been twice the length of Fig. 3. Im
figures the “a y jaws” are omitted, they being so weakly
to be quite destroyed in process of extraction. I

e :
pieces, the outriggers, are omitted for the same reason.
210 diameters.

No. 549] A CASE OF POLYMORPHISM 533

quate. The individuals with smaller humps are always
transitional or degenerative in origin.

But what could be the cause of the hypertrophy of
the humps to fully double their usual prominence, and
this in individuals that always gave evidence of starva-
tion? Such individuals occurred in certain cultures in
considerable numbers, and constituted a very extreme
type; the animals were always much more transparent
‘than any others, the body wall being thin and the internal
organs usually pale, shrunken, and undeveloped, the
stomach empty, and embryos lacking. The general body
form was extremely slender, with corona but two thirds
average width, while the ventral, or rather posterior,
hump was not only long, but developed a secondary pro-
longation, as it were, from the end of the original one.
Such animals swim, all but habitually, with the lateral
humps retracted, and in this condition are so slender as
hardly to suggest the genus Asplanchna, the form being
apparently more nearly that of Hydatina. But with the
thrusting out of the lateral humps a singular transforma-
tion occurs; these structures are so long that their ex-
panse equals or exceeds the animal’s length, and so
slender that the animal’s forward motion bends them
backward. In my notes I designated these extreme
animals as the ‘‘cross-bow type.’’

As already mentioned, they occurred in considerable
numbers in several of my mass cultures. I also obtained
them several times under controlled conditions in isola-
tion experiments, but I found no clew to the cause of
their production until I discovered that they were fre-
quently produced by the gaint campanulates, and espe-
cially by the campanulates that were Moina-feeders. It
seemed very striking that this variant, which carried the
development of the humped form to its utmost extreme,
should be thus produced by the robust companulates in
which there are no humps, and in which, indeed, all the
characteristics are at the farthest possible remove from
the type in question. Yet these cross-bow hump-bearers

534 THE AMERICAN NATURALIST [ Vou, XLVI

formed a regular part of the progeny of the massive
crustacean-feeders. All but invisible, they swam rest-
lessly about seeking for available food, which was not
present, until many of them fell victims to the greedy
members of the parental stock.

This combination of overfed parent and foodless
progeny offered the suggestion of the cause I was seek-
ing, which was then readily confirmed by experiment.
Maximum nutritive conditions before birth and the entire
absence of available food for at least 24 hours after birth
produces the slender transparent type with the hyper-
trophied humps. Under these conditions the body wall,
and its projections, which are highly developed even at
birth, continue to develop for a considerable time after-
wards, undoubtedly withdrawing nutrition from the in-
ternal organs—stomach, digestive glands, ovary, ete.—
these thereby undergoing a partial atrophy.

A certain interest attaches to this explanation, because
it not only furnishes the rationale of an extreme type of
fluctuating variation in this rotifer, but because the facts
closely parallel the incidents in the development of the
male of the same species. The males at birth lack, of
course, the chief internal organs of the female, and
can not draw upon them as sources of nutrition, but they
do draw upon the rudimentary digestive tract until,
before death, it has frequently quite disappeared. More-
over, the males undergo a progressive development of
the body wall and to some extent of the humps during
the two to four days of their active life. The male thus
becomes more differentiated in the active portion of its
organization, absorbing meanwhile what little inactive
tissue there is to absorb. The same thing happens to the
young foodless female, save that there is more tissue to
absorb, and the process is not carried so far.

Sufficient investigation would doubtless unravel each
of the other minor fluctuations which the three forms of
the species undergo, and most of them will all but cer-
tainly resolve themselves into factors of nutrition. Few

No. 549] A CASE OF POLYMORPHISM 535

species of animals are capable of so numerous and varied
nutritive transitions and respond to them in so funda-
mental and varied manners as does this Asplanchna.

Before closing this paper it becomes a disagreeable
necessity to attempt some more definite systematic place-
ment of the forms of Asplanchna here discussed. The
task is a difficult one for several reasons. In the first
place, it is obvious that the facts here recorded tend to
disturb our very conception of what constitutes a species
in this genus. If we accept the interpretation of poly-
morphism as, on the whole, a little more applicable to the
facts of heterogenesis here cited than would be the in-
terpretation of mutation, the question is obviously raised
whether several of the other Asplanchna types hitherto
described as distinct species may not likewise be closely
related genetic forms, connected, as are those here de-
scribed, either with each other or possibly with these
very types. Thus a relationship is readily thinkable
between A. ebbesbornii and A. intermedia or A. sieboldi,
or possibly of one or the other with A. brightwelli;
although the disparity of the males in this last type
renders relationship less probable.

Moreover, in the literature of the subject the claim has
been made at least once that such a relationship exists
between European forms, as the following quotations
from Wesenberg-Lund’ in which he cites Daday to this
effect, will show:

Uber die Fortpflanzungsverhiltnisse der Asplanchnen kann ich Fol-
gendes mittheilen—v. Daday (“Ein Fall von Heterogenesis bei den
Riiderthieren.” Mathem. und Naturw. Berichte aus Ungarn, 7. Bd.,
1888-1889, p. 140) hat fiir Asplanchna sieboldi einige ganz merkwürdige
und bisher exceptionelle Fortpflanzungsverhiiltnisse geschildert. Seiner
Meinung nach findet man hier zwei verschieden geformte Weibchen,
theils solche, die den gewöhnlichen schlauchförmigen Asplanchna-Typus
haben, theils solche, die den eigenthümlich geformten Männchen dieser
Art gleichen; diese sind durch 4 conische Erhöhungen characterisiert,
und sind diese Erhöhungen derartig vertheilt, “dass je eine auf die
Mittellinie des Bauches und der Rückseite, eine auf die rechte und eine

* Wesenberg-Lund, C., ‘‘Ueber danische Rotiferen und über die Fort-
pflanzungsverhiiltnisse er Rotiferen,’’ Zool. Anz., 1898, Bd. 21, pp. 200-211.

536 THE AMERICAN NATURALIST [ Vou. XLVI

auf die linke Seite fällt, wodurch, von vorn betrachtet, die Form eines
gleichschenkeligen Kreuzes sichtbar wird” (Daday, p. 153). Jedes
dieser zwei verschieden gebauten Weibchen vermag sowohl Weibchen
ihrer eigenen Gestalt parthenogenetisch hervorzubringen, als auch Weib-
chen der anderen Art; ferner auch Männchen und, nach der mit diesen
erfolgten Begattung Dauereier (pp. 206-207).

Daday’s work on the rotifers has in general been fre-
quently criticized and all but discredited, and this ob-
servation on his part of a reciprocal relationship between
A. sieboldi and a saccate Asplanchna fares no better at
the hand of Wesenberg-Lund. He replies that he has
himself reared A. sieboldi in an aquarium for a month,
that he has studied them thoroughly and has found no
such reproductive phenomena. He continues, that Daday
has simply been mistaken in his observations, having
failed to distinguish the humped A. sieboldi from the
saccate rotifer, because the humps of the former species
are often retracted, giving it for the moment a saccate
form.

This criticism of Daday’s reported observations may
of course be correct, but it seems as naive as it is severe.
It is of course true that the humped rotifers retract the
lateral humps; but the position of these protuberances
always remains marked by folds of the body wall, while
the ventral hump is not retracted at all. Daday must
indeed have been a poor observer to be thus deceived, the
more so, in that the moment these animals are placed
under the pressure of a cover glass or even in a very
shallow drop of water on a slide, the pressure of the cover
glass or their own weight forces them to expand the
humps and instantly reveal their type of structure. In
the light of my own study it seems far more probable that
Daday was correct in his reported observations than that
Wesenberg-Lund is correct in his criticism of them.
The fact that Wesenberg-Lund reared the humped
Asplanchna for a month without the occurrence of hetero-
genesis is of no especial significance. The writer has
reared the humped form discussed in the present paper

for longer periods than this with the same result.

No. 549] A CASE OF POLYMORPHISM 537

Heterogenesis is confined to special periods or caused by
special conditions as above set forth.

However, not only is the question of species confused
by the presence of heterogenesis in the genus, but another
difficulty which I had not anticipated manifests itself;
viz., the discovery that the descriptive work which has
already been done upon the genus has not, even the best
of it, been sufficiently accurate to be trustworthy. I
make this statement with the utmost reluctance, and only
after I have spent weeks of effort to bring my observa-
tions upon single points into accord with the statements
of Rousselet, who is not only the highest authority on the
group in question, but who has, as already stated, made
the last and most detailed pronouncement upon the
species of the genus. I have failed, however, in my
efforts. Nor are the discrepancies such as may, with
probability, be explained by the assumption of differ-
ences in the material which we have examined. ‘Thus, to
take an example from Rousselet’s description of the jaws.
He says:

At the tip there is really but a single point . .. ; on crushing the
jaws a “ah a ridge seen in side view is oe over and simulates
a second t

SRRA in his supplement, has also spoken a similar
effect, viz.:

When the ramus is subjected to pressure from above the deep plate
is bent by the glass (to which it stands at right angles), and its free
lower corner is twisted, so as to look sometimes like a second tooth, just
below the extreme apex, sometimes like a small plate.

Now in spite of these statements I seemed to see the
thin lamellate plate-like second tooth near the apex of
every jaw in case the conditions for its vision were at
all adequate, and in the giant campanulate this structure,
so delicate in the ordinary type, becomes greatly de-
veloped, forming a large cutting tooth, which, when the
jaws are closed, meets with its fellow of the opposite side,
in the middle line. That this thin triangular tooth in the
ordinary form was a false appearance produced by the

538 THE AMERICAN NATURALIST [ Vou. XLVI

bending over of the corner of a ridge by the pressure of a
cover glass seemed improbable, and I immediately put it
to test in various ways. For one, I extracted a large
number of the trophi by means of potassic hydrate in
deep watch-glasses. Here there was obviously no pres-
sure, yet the structures in question were quite visible,
even before the trophi had been transferred to a slide, or
touched by any instrument. Moreover, these thin
lamellate teeth are never quite symmetrical on the two
rami, and this delicate discrepancy is always on the same
side of the animal, as I ascertained later in stained and
mounted preparations. I carried my study of the trophi
farther by mounting many which I had extracted in deep
hollow-ground slides, including with them a small bubble
of air. By giving the slide a quick tilt the air bubble
could be made to strike and overturn the trophi in dif-
ferent ways. By then replacing the slide quickly under
the microscope, views could be had of the trophi before
they had settled to the ordinary horizontal position. A
half hour of such attempts readily furnished views of
every part of the trophi, seen from almost every possible
angle. Portions so thin as to be invisible in one view
become visible in another; optical sections at all points
make possible the arrival at the correct form. I regret
that in my drawing I have been able to show so little of
the delicate complexity of these structures; but Rous-
selet’s view of their structure—that ‘‘the chitinous ma
terial is bent at right angles throughout the length of
the rami, forming an inverted L in cross section’ ’—is
certainly very far from correct. The structure varies at
different points; ridges thicken and fade out in complex
and sinuous fashion, quite as we should find them in the
complex chitinous jaw of an insect, or, for that matter,
in the jaw-bone of a mammal.

The tips of the jaws are interesting, and I find no
description of the jaws of any species of Asplanchna, by
any author, which coincides with my observations.
Nevertheless, this may be due to the inadequate study of

No. 549] A CASE OF POLYMORPHISM 539

these difficult structures by systematists who can devote
but little time to a given point. I find the two rami are
never alike at the very tips. I am not speaking now of
the delicate lamellate teeth already mentioned which are
a little distance removed from the tips, but of the very
extremities. Of these latter, one is bifid, or ends in two
delicate tips; even these again are never quite sym-
metrical, but the one which is toward the animal or
posterior is a little smaller and shorter. Furthermore,
the split in the tip of this jaw is not a simple cleft such
as one might produce by splitting the end of a stick with
a knife, but is a triangular groove, the base or open side
of which is toward the inside or concave aspect of the
ramus, the apex toward the outside. As aforesaid, this
cleft divides the tip of the ramus, but it is also continued
on the inner aspect of it considerably farther than it
extends on the outer, becoming thus shallower and
shallower as it extends farther from the divided tip.
The opposite ramus is not bifid, but tapers to a point, and
the tapering is of such a nature that the jaw near the
tip is more or less triangular in cross section, so as to fit,
not only into the cleft between the tips of its fellow
ramus, but farther into the triangular groove on its inner
side as well. Thus these delicate chitinous jaws, when
closed, lock together in double manner.

The study of hundreds of examples of the trophi of
the humped rotifer as it occurred in the material first
_ examined left upon the mind of the writer a very distinct
impression of the minute delicacy of detail and very great
uniformity which prevails in these structures. Varia-
bility seemed almost wholly confined to the matter of
size,

Turning briefly to the trophi of the ecampanulate type,
I will say that they differ regularly from those just
described, not only in the features shown in the figure,
such as general size, breadth of rami, more acute angle
of the lower inner tooth, ete., but in other marked features
besides. The inner tooth, smaller in proportion as well

540 THE AMERICAN NATURALIST [ Vou. XLVI

as set at a different angle from the corresponding struc-
ture in the humped type, is here fused with the ramus
instead of being merely bent over inward from its outer
margin. But the tips differ most; the secondary lamel-
late teeth, as before mentioned, become very large,
though variable, structures. -They always meet in the
middle line when the jaws are closed; they have wavy or
corrugated surfaces, and thin down to a sharp cutting
edge. The tips of the rami are modified most of all.
They are slender and greatly extended in length, meeting
and passing at an acute angle. Neither tip is bifid, and
the asymmetry between the two rami is much less
marked. The jaws do not interlock, when closed, in the
sense in which they do in the humped type; instead, the
tips invariably cross, like the mandibles of a crossbill, the
farther closing being prevented by the meeting in the
middle of the lamellate teeth. Occasionally I have
noticed a campanulate whose jaws had sheared past in
the wrong way; the lamellate teeth then did not meet
to prevent farther closing, and the animal had apparently
lost control of the organs, as the two halves remained
crossed well down to near their bases. ;

Returning to Rousselet’s description of the trophi of
A. amphora, which must, of course, be compared only
to the trophi of the humped type, I will say that despite
the discrepancies in detail which I deem due to inaccuracy
of observation, it remains true, nevertheless, that his gen-
eral figure of the trophi of this species coincides essen- ,
tially with the general appearance of the trophi as I find
them in the humped and saccate types, and this con- `
stitutes a fair reason for assigning, provisionally, the
material which I have studied to the species Asplanchna
amphora.

However, I can not leave this matter of the trophi with-
out instancing a surprising observation which I have
made the past summer on the trophi of the related
species, Asplanchna brightwelli—an observation which :
again complicates, in an entirely new way, the question r
of species in the genus Asplanchna. a

No. 549] A CASE OF POLYMORPHISM 541

As soon as I had discovered the fact that the humped
Asplanchna which I was studying was represented by a
saccate type which in many characters approached closely
to Asplanchna brightwelli, I began an extended search
for this latter species, in order to study closely the ques-
tion of relationship or non-relationship. At different
times, throughout a period of one year, I succeeded in
finding A. brightwelli in five different ponds in my own
vicinity. The species tenants ponds of a different char-
acter from those in which the larger Asplanchna
flourishes, and is associated with a somewhat different
micro-fauna. In but one instance have I found the two
species developing together, and in this case the larger
Asplanchna did not pass beyond the saccate condition, in
which it was also present but sparingly. The resem-
blance of its occasional representatives to the more
numerous and likewise saccate individuals of A. bright-
welli was so great as to almost prevent its detection.
Only my constant work with the larger species could have
sharpened my attention to the point of noticing any
especial lack of homogeneity in the material. Yet the
saccates of the larger species were regularly a little
longer and about one fourth broader than the adult A.
brightwelli. They differed also in a number of minor
constant characters. But none of these have hitherto
found place in any specifie descriptions, with the excep-
tion of the difference in the trophi. The trophi of the
larger saceates agreed with the description and figure
which I have given in everything save that they were a
little undersized. The trophi of A. brightwelli agreed in
their general outline with the figure given for this species
by Rousselet; they possessed the more delicate, perfectly
oval contour, and invariably lacked the large inner tooth,
just as Rousselet asserts that he has always found them
to do. Into the question of the form of their tips and
the presence or absence of the all but invisible lamellate
teeth I do not go. Such study as I have given them led
me to think that they were constructed in these respects

542 THE AMERICAN NATURALIST [Vow.XLVl

essentially, but not exactly, as were those of the type
which I had already studied. I was much pleased to thus
substantiate, at least in a general way, on this American
material, Rousselet’s judgment on the distinction between
the trophi of these two species. I will add that a score
of culture experiments started with single individuals
of the two types fully confirmed their distinctness. De-
spite their very close resemblance, I reared from one set
of the delicate saccates the humped amphora type with
which I was so familiar; while parallel cultures, with
identical conditions as to food and temperature, pro
duced no modification in the A. brightwelli other than a
slight increase in size.

I therefore reached the conclusion that, delicate as are
the differences which separate the saccate form of 4.
amphora from the invariably saceate A. brightwelli, they
were none the less sharply demarcated. Ignoring other
features, it seemed perfectly safe to trust the one char-
acter of the absence or presence of the larger inner tooth
on the trophi.

Imagine my surprise, then, when, upon visiting an
entirely different locality—Custer County, South Dakota
—I discovered an Asplanchna in countless numbers
which completely upset this distinction and introduced
me to a seemingly new type of variation within the genus.

It was in the charming little mountain lake (or rather
reservoir, for the original site contained a mere pool
which has now been increased to a depth of 80 feet by an
artificial dam) called Sylvan Lake, that I came upon the
rotifer in question. The lake was indeed swarming with
rotifers of different species, which constituted the
majority of its plankton. Monarch of them all, and
profiting greatly by its superior size and ingesting power,
was a superb Asplanchna. Aside from a slight excess
in size every outward character indicated A. brightwelli.
Moreover, I had found the species in the very nick of
time, for both males and resting eggs were copiously
present. Among very large numbers of these which I

No. 549] A CASE OF POLYMORPHISM 543

immediately examined not one differed outwardly from
the brightwelli type. But examination of the trophi
yielded the astonishing result that in every instance they
bore a strong inner tooth in the exact position in which
this is found in A. amphora. I examined large numbers
of them and found that in all the features which I could
study with the facilities I had in the field, there was no
obvious variation whatever.. It should be mentioned that
in outline these trophi presented the close approach to a
perfect oval which is characteristic of the brightwelli
type. The strong inner tooth alone gave them decidedly
the aspect of the jaws of Asplanchna amphora.

This discovery is plainly again confusing, as to specific
distinctions between the types. Fortunately, however,
it serves at least to clear up certain contradictions in the
literature of the subject. Rousselet, in the article above
mentioned, dealt especially with this point. He figures
the jaws of A. brightwelli, to use his own expression, as
he has ‘‘invariably found them’’—i. e., with an oval out-
line and without the inner teeth. He concludes that the
earlier writers on the genus—Dalrymple, Brightwell,
Hudson—had certainly confused two different species of
Asplanchna, describing the trophi of A. amphora as be-
longing to A. brightwelli.

My examination of the Sylvan Lake material shows
that no such error need be ascribed to them. A. bright-
welli simply exists in two distinct races (genotypes?) ;
Rousselet has invariably found but one of these, just as I
myself have done in my own vicinity; while Brightwell
very probably found and described the other, which I
have found so abundant in the South Dakota lake.

The finding of A. brightwelli with two distinct types
of trophi may seem but a trivial matter, but taken in its
full connection it is not without interest. A. brightwelli
seems, in general, an all but constant species. Yet, judg-
ing by morphological test, it should be closely related to
A. amphora, a species which experiment shows to be
phenomenally variable.

544 THE AMERICAN NATURALIST [ Vou. XLVI

What, then, is the relationship—the physiological and
genetic relationship—between these two types? Jen-
nings in his recent work on the ‘‘Characteristics of the
Diverse Races of Paramecium’’ has prophesied that the
more exact study of the life history of rotifers will
demonstrate that much of their apparent variability is
really due to the presence, within specific limits, of
numerous fixed: races.

Now the study, as here outlined, of the variation of A.
amphora, brings to light a condition-which in no wise
substantiates this prophecy. No fixed races are present;
but strongly demarcated yet temporary types, on the one
hand, and fluctuating variations, on the other, which are
all or nearly all the result of nutritive stimuli. Is it
possible that, in spite of this, the closely related A.
brightwelli will present the fixed races which Jennings
suggests?

I have aleady indicated that a series of about twenty
culture experiments with the type of A. brightwelli first
found by me yielded no significant modification. At the
present writing I am again following this species in
copious natural development and again conducting a
few mass cultures without finding anything but farther
proofs of constancy.

I am also succeeding in rearing very large numbers
of A. brightwelli of the type whose trophi present the
inner tooth. . The resting eggs, which were brought from
Sylvan Lake the preceding August, were kept over winter
in a small amount of the lake water and hatched out in
March by adding tap water and raising the temperature
by placing the dish in the sunlight. The culture medium
has slowly been quite changed to the somewhat alkaline
and saline water of the writer’s locality. The cultures
have also been heavily fed upon organisms to which they
are certainly unaccustomed in nature. Some cannibal-
ism has been induced. But the trophi remain obstinately
true to their own type, and the general morphological
changes have been confined to a considerable increase in

No. 549] A CASE OF POLYMORPHISM 545

size of a few individuals, with perhaps.a somewhat dis-
proportionate expansion in breadth of corona. Yet the
results of a preliminary six weeks’ culture of this type of
_ A. brightwelli are essentially similar to those which fol-
lowed my attempts to modify the first type: they are
negative.

Much more extended and varied experiments must be
made before reaching final conclusions upon the con-
staney of these two races of A. brightwelli. Yet they
certainly promise to bear out in the main Jennings’s pre-
diction of relatively fixed races within the species.

Yet the foregoing does not entirely complete the pic-
ture of variation as I have found it in A. brightwelli.
While the rearing of thousands of individuals and the
examination of a very large amount of material in nature
give the appearance of two stable genetic types, yet in the
rarest instances mutation-like changes of the most
marked character probably do occur, just as they do in
A. amphora.

While at Sylvan Lake it occurred to me that the very
favorable conditions under which A. brightwelli was there
developing, including the preying upon a number of dif-
ferent organisms, were as well adapted as possible to
bring about mutational changes such as I had found to
occur in A. amphora. Day after day I examined large
amounts of material with wholly negative results. But
a favorable morning at the very close of my stay at the
lake enabled me to collect several liters of plankton as
thick as cream in consistency, with perhaps four fifths of
its bulk living Asplanchna. Pouring this in the thinnest
possible layers, into broad dishes, and placing these above
a black surface, I proceeded, by means of a powerful
reading lens, to search for any individual Asplanchna
showing marked deviation from type. Minor deviation
could not, of course, be thus detected. To my great sur-
prise, my search was finally rewarded by the finding of
three individual rotifers, and three only, of quite aston-
ishing proportions. They were certainly Asplanchna;

546 THE AMERICAN NATURALIST [ Von. XLVI

that they were derived from the brightwelli I of course
have no proof, but in the light of my study of A. amphora
it seems probable. They were campanulate forms, dif-

fering from the slender saccate A. brightwelli even more |

than the campanulate A. amphora differs from the
smaller types of its species. Seen in dorsal view, when
freely swimming in a drop of water without cover glass,
they presented almost the form of an equilateral triangle
with one rounded corner; this was the posterior end; the
entire opposite side being taken up by the loose flapping
corona. I regret that, in my haste, I was unable to study
these forms precisely, and much less to prove their re-
lationship. But I hope that the isolated observation may
perhaps induce others to seek among crowded stocks of
Asplanchna of different species for rare and much mod-
ified forms. If, as I believe will be the case, they are
found to occur occasionally in A. brightwelli and perhaps
other species, it will throw an added light upon the changes
which so readily take place in A. amphora. The rarity
of their occurrence will render clearer the relationship of
the phenomena to the recognized instances of mutation.
Before closing the discussion of facts relative to the
specific determination, statements must be made with re-
gard to the males and to the resting eggs. Similar males
are produced by all three of the forms which the am-
phora-like Asplanchna assumes. The humped and cam-
panulate types produce them copiously; the saccate type
but rarely and at periods when it is about to pass over
into the humped form. These males are always of the
well-known type bearing two lateral humps. They quite
agree with Rousselet’s determination, except that he evi-
dently assumes the size to be uniform, whereas I find it to
be extremely variable, the limits being as three to one.
The largest males, often present in abundance, reach fully
the size of the humped females; i. e., a length of 1500».
The cause for the wide divergence in size is the varying
degree of development at birth. This affects them as it

affects the young females, except that the young males, —

No. 549] A CASE OF POLYMORPHISM 547

being unable to feed and thus continue their development,
are obliged to remain at approximately the same diver-
gent sizes at which they are born.

In regard to the resting eggs, they are, of course, as are
the males, produced by all three types, and but rarely by
the smaller saccate form. The number produced by one
individual varies greatly with the degree of nutrition.
But one to three are matured if the females are poorly
fed after fertilization; whereas as many as six are fre-
quently present at one time in the body when nutrition is
high, and very rarely as many as nine may be seen. The
large campanulates usually show a high number, but it
does not exceed the maximum produced by the humped
form. The color of the egg, which Rousselet uses as a
specific character, is variable in this species. In in-
dividuals fed on Paramecia the eggs are quite white;
in individuals reared on Brachionus they are light yellow
to orange; while in Moina-feeders they are dirty white
to brown. Again, the volume of yolk, i. e., the filling or
not filling of the egg cavity, which Rousselet also regards
as important, I find to be highly variable in the eggs
of both this species and of A. brightwelli. It is par-
tially a matter of the age of the egg; but eggs are fre-
quently deposited in the most different conditions with
regard to this character. There remains the size of the
egg and the appearance of the egg coats, both of which are
highly characteristic and relatively uniform. The size
of the egg is much less, relative to the size of the animal,
than is the case in A. brightwelli; but the actual size is
larger and exceeds the dimensions given by Rousselet
for A. amphora, viz., from about 200» to 225», as con-
trasted with his figure, 170». This is surprising, in that,
in the case of A. brightwelli, my measurements of the
resting egg—170» to 190»—is less than the figure—205p
—given by Rousselet.

The egg envelopes, which I have studied in an almost
indefinite amount of material, grown under very diverse
conditions, are the most uniform and at the same time the

548 THE AMERICAN NATURALIST [ Vou. XLVI

most peculiar feature which I have found in the species.
They plainly do not agree with Rousselet’s characteriza-
tion of the egg of A. amphora: ‘‘The outer shell en-
velope consists of numerous much smaller globular trans-
parent cells’’ (smaller than the cells in egg coat of A.
brightwelli) ‘‘through which a finely dotted inner mem-
brane can be seen.’’ I find that at a certain intermediate
stage of development a dotted inner membrane can be
seen, the dots being the ends of either tubes or
rods making up a thick inner coat; the rapidly devel-
oping outer shell, however, soon obscures these dots and
the coat assumes at first a wrinkled, then a heavily cor-
rugated, surface. The corrugations are so disposed that
many of them converge at two opposite poles of the egg.
I deem it quite impossible that this characteristic and
beautiful structure should have been overlooked by any
one studying this species in detail and with the full char-
acters which it possesses in the writer’s vicinity. It
therefore seems very probable that the type of A. am-
phora studied by Rousselet was not identical with that
studied by the writer, and it may accordingly prove
necessary to eventually separate the form I have
studied from the original type of the species, ascribing
it varietal rank, based on at least this one character of
the egg coats. The systematic predicament in which
this would place these beautiful rotifers would indeed be
pathetic or intolerable or humorous, according to our
attitude toward things systematic. We should have two
varieties, separated from each other by a single fixed
character only, and one of these varieties would comprise
within itself, besides a host of minor variations, three dis-
tinct types, each of which differs from its fellow, not only
more than do the varieties differ from each other, but
more than the whole species, at its nearest point of ap-
proach, differs from its closest congeners. |

There is not the least known reason why actual facts
of genetic relationship should not be as complicated as
this, and if they are so we must deal with them systemat- —

No. 549] A CASE OF POLYMORPHISM 549

ically in some fashion. It is evident, however, that the
species in question and other allied forms should be more
intensively studied by workers in other localities before
we venture upon the final solution of so intricate a ques-
tion. For the present all that needs be said is that the
material studied by the writer and designated by the
phrases, the saccate, the humped, and the campanulate
forms, belongs to the species Asplanchna amphora, as at
present constituted; and it seems no less certain that this
material is sharply segregated from A. brightwelli, de-
spite the exceeding closeness of this latter species to the
above mentioned saccate form of A. amphora.

A brief résumé of the chief characters of Asplanchna
amphora, as here studied, will be of use to rapid workers.
It is as follows:

Species trimorphie, each of the three forms showing fluctuating vari-
ation and occasionally intergradations.

Form A, saceate type, produced from resting egg and multiplying by
rapid parthenogenesis, through several generations; corona about equal-
ing diameter of body or less, nearly cireular in outline, agreeing with
the cylindrical body, which rests on side when water is withdrawn;
body without humps or with small dorsal hump only; flame cells vary-
ing in number from approximately 20 to 40; contractile vesicle large;
trophi as in next form, oe smaller—about 95 p to 135 p long.
Length of entire sar 500 p to 1,200 p
Form B, humped type, PREBE E known as Asplanchna
amphora, PEARS from form A by rapid saltation and reproducing
chiefly its own type; body conical, strongly flattened dorso-ventrally,
with one posterior, one dorsal, and two lateral humps of varying size
and habit of carriage; corona oval, agreeing with the flattened body,
which causes animal to rest on dorsal or ventral surface when water is
withdrawn; flame cells 40 to nearly 60; contractile vesicle small; trophi
strong, typically from 150 p to 170 p in length, though varying from
130 » to 200 p, enclosing when closed an area which is not oval but
widest in its distal third, with prominent tooth projecting inward
seemingly from the inner though really folded over from the outer
margin of each ramus, delicate lamellate teeth near the tips and the two
rami interlocking when closed by means of one bifid and one pointed
tip; accessory jaws, as also in forms A and C, very pid developed.
Length of entire animal approximately 1,000 » to 1,800

Form C, campanulate type, originating usually from gums Basa
result of cannibalism, and reproducing both its own form and form B;

550 THE AMERICAN NATURALIST [ Vou. XLVI

body very broadly saceate to broadly campanulate in form, with very
heavy walls and musculature, strongly flattened dorso-ventrally, never
with humps; corona oval and very broad, its breadth frequently equal-
ing the length of the animal; anterior end of animal, within corona,
concave instead of convex; flame cells approximately 80, to 115 p;
contractile vesicle small. Animal resting when water is withdrawn on
dorsal or ventral surface ; trophi very large, typically from 300 p to 340 p
in length, enclosing a narrowly oval area; inner teeth relatively less
prominent than in preceding types, set at an acute angle with the ramus
and more firmly fused with it than in the preceding types; lamellate
teeth near tips of rami much developed and meeting, with cutting edges,
in middle line; tips of rami not interlocking but shearing past each other
when closed. Length of entire animal approximately 1,800 u to 2,500 p.

In conclusion, it may be pointed out that the type of
variation shown by the rotifer here discussed seems
somewhat peculiar, in that it lies seemingly on the line
between germinal variation and variation which is com-
monly supposed to be somatic. To use recent phraseol-
ogy, it is difficult to say whether the types which this spe-
cies of Asplanchna produces should be called genotypes
or phenotypes.!° They are like genotypes in that when
once produced they manifest a marked tendency toward
stability, each type reproducing itself through a number
or even a multitude of generations after the special con-
ditions which favored their origin have ceased to be
present. They are to some extent like phenotypes in that
this stability is less than that of true species, yielding,
though rarely, to degenerating or other modifying con-
ditions.

* As the proof of this article passes through my hands, one of the above
terms—‘ phenotype’ ’—is already a matter of ancient history; while

to the study of other rotifers and protozoa—are already known to the
writer which harmonize even less than the facts of the present paper vin
the rigid conceptions which some set forth with show of finality.

No. 549] A CASE OF POLYMORPHISM 551

It is worth noting—though this is in part but restating
the last thought in different language—that the varia-
tions here described differ from the majority of those re-
cently recorded for minor invertebrates ;'' for example,
the modifications in Daphnia, Bosmina, and Asplanchna,
so carefully observed by Wesenberg-Lund. These latter
variations are in the main variations in external form
only, and seemed to be pure reactions to external con-
ditions, taking place, for example, when the surrounding
medium has reached a certain temperature, and again
lapsing very soon after the temperature has dropped.
Such variations fall naturally under the rubrics of sea-
sonal polymorphism, temporal variation, or eyclomorpho-
sis. The variations which we have studied in Asplanchna
refuse to be thus classified.

It is true that the stability of these variants is mark-
edly different, being greatest for the humped type and
least for the minor saccate form, but a stability that
tends strongly to resist external influences is none the
less obvious for each. And this seems to the writer to
render it highly probable that each of these variations is
of germinal origin. If this is the case it is the more
striking that this germinal variation 1s itself a variable
and elastic quantity, originally initiated by nutritive
causes,

In one sense only may it be said that these mutational
variations do occur in a rhythmic or cyclical fashion, in
that, namely, each form may produce a fertilized or rest-
ing egg that tends to return to the common starting point.
The variations are therefore obviously not transmitted
through the resting egg as they are through partheno-
genetic ova. It is, however, by no means certain that
there is complete community of kind in all the young
hatched from resting eggs. Such observations as have
already been made seem to show that the progeny of the
resting eggs of this species are by no means uniform,

1 Some of the variations recorded in the recent work of Woltereck ap-
proach more closely to those here recorded for Asplanchna,

552 THE AMERICAN NATURALIST [ Vou. XLVI

physiologically or morphologically. Some stocks seem
larger than others from the start, and apparently gave
rise more readily to the second and third types. It will
require much careful experiment to ascertain the cause of .
these diversities, and whether a tendency toward the
transmission of variations actually lies in the resting egg.
If such proves to be the case, light will be immediately
thrown upon the farther problem, namely, whether the
saltations here described are intimately related to a true
species-making process. All in all, it seems that they
probably are thus related, especially as the forms pro-
duced parallel so closely other types of the genus which
are now universally regarded as definite and circum-
scribed species. |
But are these other types of the genus definite and cir- ‘
cumscribed species, or are they (some of them at least) 7
but semi-independent types, occasionally brought into
existence by unusual nutritive conditions and then main-
taining for a time only their partial or complete auton-
omy? Unfortunately these remaining forms of the
genus are not accessible in the writer’s vicinity. But a
they would seem well worthy of careful study, both ob- |
servational and experimental, where they may be found,
and it seems to the writer that such study, sufficiently
prolonged, will bring to light a species-making process
in rotifers which is somewhat different from any as, yet :
demonstrated in the animal kingdom. 3 i
It is just possible that these saltational phenomena may
be purely local, or at least greatly exaggerated in the
genus Asplanchna. The food reactions of this genus are
undoubtedly extreme, and the development of their par-
thenogenetic ova in close proximity to this spasmodic
and very variable nutritive supply may possibly make
this genus exceptional. But no fundamental organic phe- _
nomenon is wholly isolated and unlike the phenomena of
other species. If nutrition can modify the germ cells in
the genus Asplanchna and thus bring into existence new
types, nutrition surely must be a factor on a wider scale.

SHORTER ARTICLES AND DISCUSSION

AN UNUSUAL SYMBIOTIC RELATION BETWEEN A
WATER BUG AND A CRAYFISH

WRITERS on animal ecology and popular entomology have
made us familiar with a remarkable habit of various species of
water bugs belonging to the genera Zaitha and Serphus. In
these forms it has been established that the female seizes a male
and by superior strength and apparently against his will, cov-
ers his back with her eggs. These adhere together and to his
tegmina to form a dense mass as thick as his own body. Con-
verted thus into ‘‘an animated baby carriage” as Howard puts
it, the male carries the whole brood about with him until they
hatch, providing them with protection and, possibly, improved
aeration. It is worthy of note that the habit has been observed
in widely separated parts of the world, viz., both coasts of
America, Europe, and Japan.

The writer has observed a somewhat analogous adaptation in
another group of aquatic Hemiptera, the Corixide—a habit that
appears to be so extraordinary that he has refrained from de-
scribing it until it should have been found to be other than a
local instance. Quite by accident he discovered that a similar
observation had been made previously by S. A. Forbes" so that
it seems rather improbable that the circumstances should be
accidental.

During the summer of 1910 it was observed that numbers
of the crayfish taken from a pond near Columbia, Mo., were
covered with the eggs of some insect. When hatched out in
aquaria these were found to be a species of waterboatman.? The
crayfish were those of a common species of the neighborhood,
Cambarus immunis Hagen and both young and old were in-
vested with the eggs. In well covered specimens the telson, the
legs, the sides of the abdomen, and nearly the whole of the
cephalothorax, including the eye stalks, and basal parts of the

1 Forbes, Bull. IU. Mus. of Nat. Hist., I, pp. 4-5, 1876; ibid., AMER. NAT.,
XII, p. 820, 1878.

2 The species has been described elsewhere (Canad. Entom., XXXI, pp.
113-121, 1912) as Ramphocoriza balanodis n. gen. et sp. and a full account
of its metamorphoses given.

553

554 THE AMERICAN NATURALIST [ Vou. XLVI

antenne, were clothed with a felt-like cloak of tiny eggs each a
little less than a millimeter long and imbedded in a small cup
which is affixed to the carapace. The cup has been described
for other species of Corixide which attach their eggs to stems
of water plants and is not to be considered as a special adapta-
tion in the present instance. It was found, however, that the
carapace was slightly impressed for the reception of each egg
cup, as if the affixing of the egg had either softened the chitin
somewhat, or had taken place before the hardening subsequent
to ecdysis had been completed.

In a ‘List of Illinois Crustacea,” under Cambarus immunis,
Forbes (l. c.) states that:

About one fourth or one half the specimens taken from stagnant
ponds in midsummer [in Central Illinois] are more or less completely
covered above by the eggs of a species of Corixa, probably C. alternata
Say, since this is much the commoner of the two species found in
such situations, the other being as yet undescribed. :

As the present writer has taken Ramphocoriza in Illinois it is
highly probable that it was the ‘‘yet undeseribed’’ species men-
tioned by Forbes. The same species of water bug is also found
in Texas, so that it is more than likely that the distribution of
the insect is coincident with that of the crayfish Cambarus im-
munis. Forbes states also that a ‘“‘careful search of the weeds
and other submerged objects in the ponds discovered no other
place of deposit of these eggs.’’ The writer also can testify to
the same point. The waterbug in question is abundant where
found, but its distribution is not general and it is not improb-
able that it is conditioned by the presence of the crustacean
species with which it has undertaken this unusual partnership.
All the Corixide are strong flyers and ‘‘swarm’’ at maturity,
so that with the general similarity of habitat which exists
throughout the Mississippi Valley there is no other reason why
Ramphocorica should not be equally as well distributed as some
other species of Corixids (e. g., Arctocorisa interrupta Say)
found there. nae

The insect when mature, measures but 5-54 mm. in length
and a very large number of females must simultaneously par-
ticipate in the egg laying so to cover an individual crayfish.

o count was attempted of the eggs on any one crayfish but the
number must often run well up into the hundreds. ;

The investiture of eggs commingled with debris certainly —

No. 549] SHORTER ARTICLES AND DISCUSSION 555

556 THE AMERICAN NATURALIST [ Vou. XLVI

renders the crayfish less conspicuous and it probably profits by
the arrangement in much the same was as do various shore-
crabs which are decorated with sponges, alge or ccelenterates.
Whether the water bug improves its chances against racial ex-
termination by the adoption of such a pugnacious protector it
may be too much to assume, but at any rate whatever the util-
itarian value of the habit it must be of the same nature as that
which obtains in the widely distributed genus, Zaitha. An ob-
servation of the manner of egg laying on the crayfish would be
of much interest.

JAMES F. ABBOTT
WASHINGTON UNIVERSITY

DOUBLE EGGS!

UNDER some such caption as the above there have appeared
from time to time in zoological literature various accounts of
anomalous eggs, chiefly of the common hen. These have nat-
urally elicited more or less popular interest, and various expla-
nations have been proposed concerning them. While it is no
part of the present purpose to review the history of these phe-
nomena it may not be amiss to merely call attention to a few of
the more striking titles under which they have been described.
For example, Barnes (63, ’85) has described cases under the title
“Ovum in Ovo”; and Schumacher (96), “Ein Ei im Ei”;
Parker (06), ‘Double Hens’ Eggs”; and quite recently Pat-
terson (711), “A Double Hen’s Egg,” are typical of numerous
titles appearing in the literature. The chief purpose of the
notes which follow is to call attention to an earlier paper by the
writer (’99) and to describe subsequent facts which have come
to his knowledge. The only reason for specially referring to
the earlier paper (’99) is that it seems to have been wholly
overlooked by later observers of these phenomena, and this is the
more strange in that both Parker (706) and Patterson (711), to
whom the journal (Zool. Bull. ) was quite familiar and access-
ible, make no mention of it.

In Fig. 1, which is reproduced from the article just cited, are
Shown the essential features of the first case which came to my
direct knowledge some time prior to the date of the paper in
question. As will be noted this presents a very clear illustra-

* Contributions from the Zoological Laboratory, Syracuse University.

No. 549] SHORTER ARTICLES AND DISCUSSION 557

tion of that class of egg anomalies known as ‘‘ovum in ovo,”
and its simplest interpretation appears to be that originally
given to it by Schumacher (’96), namely, that it is the result of
a return of the egg up the course of the oviduct by an anti-

Fig. L

peristalsis of that organ, and then later a descent during which
the egg would receive a second deposition of albumen, shell
membrane and finally a second shell, giving it just the consti-
tution shown in the figure, and described in my paper (p. 228).

In Fig. 2 is shown a case which differs in essential respects
from the preceding. The egg came to my knowledge through
the kindness of my colleague, Dr. C. G. Rogers, in whose father’s.

poultry yard it was produced. This egg, as will be observed, was
double in a rather unusual way. Here we have as shown from
the outside an egg of rather larger size than usual, but other-
wise apparently perfectly normal. When broken to be used in
the kitchen the anomalous internal condition was revealed. The
sketch will make clear in just what this anomaly consisted,

558 THE AMERICAN NATURALIST [ Vou. XLVI

namely, the inclusion of a miniature egg within the larger and
in about the position and relation shown in the figure.

A double egg of similar character has been recently described
by Patterson (Am. Nar., Jan., ’11), though differing in that the
anomaly comprised two fairly large eggs, as shown in his sketch
(Fig. 4), while in my own specimen the inner egg was quite
minute though otherwise normal. Some further discussion of
these cases will appear in a later section of the paper.

Fic. 3. Abnormal hen’s egg x io.

In Fig. 3 is shown a third anomaly differing from either of
the preceeding in a very marked way. The photograph of the
specimen, about one half natural size, gives a better impression
of the specimen than any verbal account could do. The most
striking feature is that of shape, which is rather gourd-like, and
was sent to me by the father of Dr. Rogers with the rather
facetious suggestion that the contiguity of the poultry lot to the
garden, over whose fence hung a squash vine, might afford a
clue to an explanation! The egg was laid aside for a time
awaiting photography, and when later I opened it for a critical
study it was found to have lost so largely by evaporation that
an exact account of all its details could not be made. This may
be stated, however, that in the larger end of the egg was an ap-
parently normal yolk and normal albumen. The smaller end
seemed to have had only albumen, though it was yellowish, as if
there might have been yolk matter distributed through it.
this one can not be certain, and I must leave the matter as doubt-
ful. However, I am disposed to submit the general statement
given above, namely, that the egg was comprised of about nor-

No. 549] SHORTER ARTICLES AND DISCUSSION 559

mal parts in the larger end, and the smaller probably having
only albumen, its yellowish tint having resulted perhaps from
the evaporating process which had taken place.

In the matter of explanation or interpretation of these facts I
have little to add to what has been presented in the earlier paper
or by other observers. Of the literature at my command the
paper of Parker (op. cit.) seems to me to present upon the
whole the best discussion. And I may add in this connection
that Parker’s paper is further valuable in its rather full bibliog-
raphy of the subject. As already mentioned in connection with
the account of Fig. 1, the true interpretation seems almost cer-
tainly that there cited. One has but to apprehend the essential
physiological operations involved in the process of the so-called
antiperistalsis to perceive just how there would result the strue-
tures present in the egg. If it should be queried why such depo-
sition might not have taken place on the ascent of the egg by
antiperistalsis as well as during the later descent, it may suffice
to admit that perhaps it did occur. However, in case the re-
turn of the egg up the oviduct took place soon after its original
descent the glandular structures would be in a state of exhaus-
tion and hence capable of only slight discharge; but in either
case, save only the action of the shell gland whose only effect
would be to add to the thickness of the original shell, the effect
would prove the same, namely, a second layer of albumen, a
second shell membrane and finally a second shell just as was the
case. Parker’s contention as to the fact of antiperistalsis seems
to me conclusive. The facts of normal eggs in the body cavity
of hens, cases of which I have known, seem impossible of ex-
planation by any other view.

The case involved in Fig. 2 is rather more complex, though not
so difficult of correlation with known processes as might seem.
First, let us direct attention to the minute inner egg. Such
miniature eggs are fairly familiar to any one who has much to
do with poultry culture or care. They are oftenest found with
the first ovulation of young hens, and the writer has known of
them from boyhood as pullets’ eggs. They probably represent
an early or premature ovulation at the beginning of sexual ac-
tivity. The discharge of such minute yolk would involve only
comparatively slight stimulus of the uterine glands and hence
a meager discharge of albumen, ete., hence the minute size. Ex-
cept in matter of size such eggs are usually normal and call for

560 THE AMERICAN NATURALIST [ Vou. XLVI

small account in themselves. Now in the second place, let us
consider what might happen at any time with the discharge of
such premature eggs from the ovary. If followed soon by the
discharge of a mature egg from the ovary and its normal descent
it might well overtake the smaller specimen at some portion of
the oviduct and easily include it within the larger mass of al-
bumen. This, it seems to me, is probably just what happens in
the majority of such eases, possibly in all. I do not overlook the
still more anomalous case cited by Herrick (’99), in which the
smaller included egg is in the yolk instead of the surrounding
albumen. Of this Herrick offers no definite explanation ; indeed,
there may be some doubt as to exact facts in this case, the inclu-
sion having been found in a cooked egg and details being un-
certain.

Concerning the specimen of Fig. 3 there is little to be said.
Its bizarre shape is remarkable, but here again the element of
doubt as to the definite composition of the contents of the
smaller end—handle of the squash—render unprofitable any at-
tempt to discuss or speculate as to its real significance. Whether
there may have been some rupture of the original yolk and the
segregation of a portion in one end with the extruded part in
the other may be a possible explanation; or whether some mal-
formation of the oviduct may have been a disposing cause must
remain open questions. Various egg shapes are familiar to
those handling large numbers of eggs. I have myself seen many
such, though none resembling the one here under consideration.
That conditions of confinement, close inbreeding, or other fea-
tures of habit or environment may have an influence in such
matters are altogether possible. Association with unusual
shapes, colors, ete., at certain times may affect domestic animals
variously; e. g., witness the very interesting story of Jacob’s
spotted cattle (!), still the contiguity of garden and poultry
yard referred to above can hardly be considered as a vera causa
in this instance! Cuas. W. HARGITT

SYRACUSE UNIVERSITY

LITERATURE CITED

Hargitt, Chas. W. Some Interesting Egg Monstrosities. Zool. Bull., Vol.
TI, 1899.

Herrick, F. H. 1899a. A Case of Egg within Egg. Science, Vol. IX, P-
364; ibid., b. Ovum in Ovo. Am. Nat., Vol. 33, p. 409.

Parker, G. E Double Hens’ Eggs. Am ie: Vol. 40, p. 13, 1906.

Patterson, J. T. 1911. Aw. Nart., Vol. 45, p. 54.

NOTES AND LITERATURE

AMERICAN PERMIAN VERTEBRATES!

Tus work might have been entitled Some American Permian
Vertebrates. It is not a general treatise on the vertebrates
found in the Permian of America, but one on a few amphibians
and a number of reptiles to which the author has recently been
giving his attention. The book is, however, not less valuable
because of its limitations.

For a number of years Dr. Williston has been making col-
lections from the Permian deposits of Texas. He has been
studying these collections, as well as the materials secured by
Cope and now in the American Museum of Natural History in
New York, and the collections, now in Yale University, brought
together by Marsh. Dr. Williston has found some remarkably
well preserved remains and these have been most skillfully pre-
pared by his assistant, Mr. Paul Miller; and in this book we
have some of the results of their labor.

Thanks to Williston, Broili, and Case, our knowledge of the
interesting animals of the Permian has been greatly increased.
We seem to be justified in believing that during the Permian the
principal orders of reptiles took their origin, or at most had not
yet diverged far from the parent stem. It is therefore of the
highest importance that every scrap of materials be studied that
is likely to throw light on these reptiles and their relationships.

As it seems necessary for.a reviewer to discover some errors
and deficiencies, some fly in the ointment, let this duty be first
accomplished.

The text is well printed and the text-figures well made and
effective. Most of the plates are excellent, especially these made
after drawings. Those reproduced from photographs, as Plates
XXVI-XXVIII, are useful mainly in showing that the author
had a sufficient basis for his line drawings. These Permian
fossils are very refractory subjects for photography, being vari-
ously mottled and stained. There are, however, methods for

1:1 American Permian Vertebrates,’’ by Samuel W. Williston, professor
of paleontology in the University of Chicago. The University of Chicago
Press, Chicago, Ill. Pp. 145; 38 plates and 32 text-figures. Price $2.50 net,

2.68 postpaid.

$2.68
561

562 THE AMERICAN NATURALIST [ Vou. XLVI

hiding such stains and giving the objects a uniform color, so
that light and shade produced by the varying surfaces need not
be interfered with; and it might be well to test these methods on
such fossils.

he reader, at least this one, can not always determine the
exact size of the animals described; for example, that of
Seymouria baylorensis. On page 140 we are told that the figures
of the plates are of the natural size, unless otherwise stated,
wherefore we might conclude that the figure on Plate X XVI is of
the size of nature. However, on pages 51 and 52 the figures of
the same skull are explained as being one half the natural size,
and they are somewhat more than two thirds the size of the skull
of Plate XXVI. As the author seems not to state the size of
the animal we are left in doubt.

The present writer would suggest that the important Plate y
ought to have had its figures lettered so as to indicate what names
the author intended to apply to the various elements. By
digging in the text with sufficient assiduity the unfamiliar stu-
dent may, after struggling perhaps with such expressions as
‘‘the real, so-called coracoid”? (p. 57) and ‘‘the so-called true
coracoid” (pp. 97, 100), determine to what parts the various
terms are to be applied.

Inasmuch as Dr. Williston argues that the exact content of the
terms Theromorpha and Pelycosauria and the exact relation-
ships of the groups can not yet be determined, it would appear
better to have retained Pelycosauria for the order which he calls
Theromorpha, especially since Case has employed Pelycosauria
in his monograph on the group. It is still more difficult to
follow Dr. Williston in displacing the well-founded family name —
Clepsydropide in favor of Sphenacodontide ; when, according tO

is own researches, the genus Sphenacodon, with great prob-
ability, does not belong in the same family as Clepsydrops.

Having uttered these mild complaints, it is a pleasure to
recognize the value of the services rendered to science by Dr.
Williston in his descriptions of Limnoscelis paludis, Seymour@
baylorensis, Varanosaurus brevirostris and Casea broilii. These
descriptions are based on materials so complete and so abundant
that practically the whole osteology of each is known. The re-
mains form a marked contrast with those on which Cope was —
compelled to found most of his work on the Permian reptiles —
and amphibians.

No. 549] NOTES AND LITERATURE 563

The genera Limnoscelis and Seymouria belong to the Coty-
losauria; Varanosaurus and Casea to the Theromorpha. The
types of Limnoscelis paludis are in Yale University, and were
collected many years ago in New Mexico for Marsh. One speci-
men is a skeleton lacking only the skull, the front feet and a part
of one hind foot; the other lacks only parts of the hinder feet.
And all these parts are in their natural positions! What more
can the paleontologist desire? Doubtless he will regret that the
animal had not fallen into some pool of asphalt that had the
property of preserving the flesh and internal organs. The prin-
es skeleton described by Williston had a length of about 7
fee

a genus Seymouria was originally iaa by Broili on
two skulls obtained in Texas. Williston secured in 1910 a speci-
men of another species of the genus and this specimen had
missing only a part of the tail; and he expects yet to secure even
this. The bones are all in the closest natural articulation and
are neither distorted nor compressed. This reptile was about
2 feet long.

Varanosaurus was described by Broili on a skull and part of a
skeleton. Williston has secured of another species 25 skeletons,
of which 6 or 8 have been recovered in greater or less perfection
from the matrix. He figures a mounted skeleton and states that
it measures just 44 inches in length. The head is long, narrow
and pointed in front.

Casea broilii was a reptile about 3 feet in length. Its head is
small, short, broad and deep. Williston presents a figure of a
restoration composed of three individuals; but he thinks that in
his collection there remain other skeletons. Among the pecu-
liarities of the reptile are a large parietal foramen and a large
infratemporal vacuity.

r. Williston presents at length the structural features that
belong to the two orders Cotylosauria and Theromorpha. These
are very instructive; but when we compare the two sets of
characters we find that nearly all of them are either common to
the two orders or of no great value. The Cotylosauria, how-
ever, possess no temporal vacuities, while the Theromorpha have
one on each side. The former are said to have the lachrymal
prolonged to the anterior nares; the latter not so. However,
the figure, 25, of Varanosaurus represents this bone as reaching
the nost

564 THE AMERICAN NATURALIST [ Vou. XLVI

Williston evidently regards the presence of a temporal vacuity
as sufficient to justify the separation of Varanosaurus and Casea
from the Cotylosauria; and he may be right. His position could
not be questioned if it could-be shown that the presence or the
absence of this feature indicated the divergence of two phyla;
that the one group gave origin to descendants that retained the
temporal roof intact, while the other started a line that developed
one or two vacuities on each side. However, that proposition
can hardly be proved as yet.

In Varanosaurus the temporal roof is mostly lacking and there
is no lower temporal arch, differing in the latter respect greatly
from Casea. Dr. Williston is led to discuss the value of the
vacuities and arches in the classification of the reptiles. He
recognizes three chief types, perhaps three chief phyla: (1) the
Cotylosauria, with unbroken temporal roof; (2) the type in
which there are two vacuities and two arches; (3) the single-
arched type, in which there is a single vacuity bounded below
by the jugal and quadratojugal. He thinks that there may be
a fourth type, that in which a vacuity is bounded below by the
postorbital and the squamosal. He is, however, unable to see the
distinction between the two types with a single vacuity, and is
inclined to believe that all single-arched reptiles have arisen from
asingle type. The present writer is unable to understand clearly
the position taken. . l

Inasmuch as the temporal roof is primitively, as in the Coty-
losauria, complete and composed of two series of bones, it is ae
vacuities which developed in them that are the important matters
to consider. It seems to the writer that a single vacuity may
have originated in five different ways:

1l. By the development of the upper vacuity alone.

2. By the development of the lower one alone.

3. By the appearance and extension of a vacuity in the
postorbito-squamosal arch.

4. By the gradual reduction of the postorbito-squamosal bar,
allowing the upper and the lower vacuities to unite.

5. By the reduction of the lower arch, leaving only the upper
vacuity.

The matter may be further complicated by changes in ee

temporal roof such as are found in some of the turtles: (1) Its

lower border may be eaten away, resulting finally in a condition
such as appears to exist in Varanosaurus ; (2) the hinder border |

No. 549] NOTES AND LITERATURE 565

and upper part of the roof may by degrees disappear until there
is left only a narrow lower arch; and even this may waste away.
Among the turtles the modifications in the temporal roof, numer-
ous and extreme as they are, are not regarded as of great im-
portance. It may be different, however, among the other
reptiles. If so, then, as it appears to the writer, there might be
five phyla of reptiles possessing in the temporal roof a single
vacuity.

It is to be hoped that Dr. Williston’s researches will lead
to a solution of the difficult problem involved in the higher
classification of the reptiles.

O: P: Hay

U. S. NATIONAL MUSEUM

FEDERLEY’S BREEDING EXPERIMENTS WITH THE
MOTH PYGÆRA

INTERESTING results have recently been obtained by Federley*
by breeding moths of the Notodontid genus Pygæra. Three
common European species furnished the material—P. curtula,
P. pigra and P. anachoreta.

The hybrids were not all equally easy to obtain. Numerous
matings involving P. anastomosis were made, but no offspring
were obtained. Anachoreta males show little inclination to
pair with curtula females, but when such pairing occurs nearly
all the eggs start developing, yet only a few reach the adult
stage. On the other hand, the reciprocal mating (curtula male
to anachoreta female) is easily accomplished, but produces only
about 30 per cent. fertilized eggs. Of these most of the males
and some of the females reach the adult stage. Thus it appears
that ‘‘Paarungsaffinität”’ (tendency to mate), ‘‘sexuelle Affini-
tät” (tendency toward fertilization) and ‘‘physiologische Affin-
ität (tendency to produce fertile offspring) are independent.

One great difficulty met with was that the adult F, hybrids
were very sterile. Only a single F, moth was raised, and only
a few from the various back crosses (F, by P,).

One of the characteristics of the species anachoreta is the
presence of a white spot on the first abdominal segment of
the caterpillar. In one of Federley’s races of pure anachoreta
there appeared, in the same brood, two caterpillars lacking the

*Arch. Rass.- u. Gesellsch.-Biol., 8, 281, 1911. Reviewed also by M.
Daiber, Zts. ind. Abstamm.- u. Vererb.-Lehre., 6, 90, 1911.

566 THE AMERICAN NATURALIST [ Vou. XLVI

spot. The mutation proved to be an ordinary Mendelian reées-
sive. This spot is absent normally in curtula, and in crosses
between anachoreta and curtula it does not appear in F,, or at
least is never of the full size. Its behavior in this hybrid is
somewhat complicated, and more data will probably be re-
quired in order to explain it. But Federley’s assumption of
imperfect dominance of the same gene which behaves as a com-
plete dominant in the anachoreta mutant seems hardly. justifi-
able. The fact that the character behaves in the mutant as
though due to a single factor does not mean that it must. al-
ways so behave. It may depend upon the simultaneous pres-
ence or absence of several genes. If in the ‘‘spotless’’ mutant
one of the required genes has dropped out, then the addition of
that one to the complex will give the spot, and a case of Men-
delian monohybridism will result. But curtula may be ‘‘spot-
less’’ because it lacks some other part of the required combina-
tion, in which case the behavior might be quite different from
that in the case of the mutant.
When curtula and pigra were crossed, some of the F, imagos
emerged after a pupal stage of about two weeks, while the. rest
hibernated as pupe. The moths resulting from the two lots
were quite different, the first (summer generation) being more
similar to curtula, the second (spring generation) more like
pigra. That this difference is not due to the effects of temper-
ature is indicated by an F, moth reared from eggs laid by an
individual of the summer generation. This moth hibernated
in the pupal stage, yet resembled the summer generation. Fur-
thermore, low temperature experiments on these curtula-pigra
hybrids and upon curtula-anachoreta hybrids gave entirely neg-
ative results. Several facts bearing on this problem are given.
Seasonal dimorphism is never a well-marked phenomenon in-
Pygera, and does not seem to occur at all in the three species —
dealt with by Federley. From pigra he was unable to rear a
summer generation. In the case of curtula, the Finnish races —

father. From the reciprocal cross only males were reared. These
also resembled anachoreta. However, Standfuss reared both —

No. 549] NOTES AND LITERATURE ` 567.

sexes but does not mention any dimorphism. Federley seems un-
decided as to whether this is a case of ‘‘spurious allelomor-
phism’’ (i. e., sex-linkage) or a reversal of dominance due to
a difference in sex (similar to the case of horns in sheep). But
if, as he is inclined to suppose, Standfuss really got no dimor-
phism in his reciprocal cross, then this can not be a case of re-
versal of dominance, since if it were, reciprocal crosses would
give the same results. It seems more probable that there is
here a case of true sex-linked inheritance, the female being
heterozygous for sex, as in Abraxas. Just what character is
caused by the sex-linked gene is difficult to discover from Fed-
erley’s account, but since this gene must be carried by anacho-
reta, let it be represented by A. The following formule, “which
I would suggest, in which MM denotes a male, Mm a female,*
will then explain Federley’s results:
curtula g — aM aM
anachoreta g AM am
aM Pr per similar to anachoreta.
M am — Ẹ similar to curtula.
anachoreta ĝ— AM AM
curtula r aM am
AM oM — —d similar to anachoreta.
AM am — Q similar to anachoreta.

The following back crosses were made:
curtula 9 — aM am

AM aM —o similar to anachoreta.

aM aM — ¢ similar to curtula.

AM am — Q similar to anachoreta.

aM am — 9 similar to curtula.
This last mating produced only three males, which were very
like the F, males. The next cross, and the expectation on the
hypothesis of sex-linkage, is:

anachoreta 9— AM am

AM AM) _ dé similar to anachoreta.
aM AM
AM am — Q similar to anachoreta.
aM am — Q similar to curtula.
I have given my reasons for adopting this sex formula for birds and
estan in another paper (Jour. Exp. Zool., 12, 499, 1912).

568 THE AMERICAN NATURALIST [Vou XU

All the males from this cross were again similar to anachoreta,
and there was apparently a fair number of them raised. All
the females belonged to the anachoreta type, but they are said _
to have been few in number.
Thus, although the classes are not all filled, because of the
small numbers obtained, the results of the back crosses are in
agreement with the hypothesis that we have here a case of sex-
linkage of the Abraras type.
One interesting point is that in the cross of curtula male by
anachoreta female, from which ‘‘hundreds’’ of females were
raised, there occurred a single female resembling the males.
This furnishes another case of partial sex-linkage, in addition
to the one reported by Bateson and Punnett? and the others
which I have analyzed in another paper.*
In practically all of Federley’s cases the offspring of back
crosses strongly resembled the hybrid parents, but he explains
this as probably due largely to the great mortality of the
caterpillars. In only a few cases were more than three or four.
offspring reared from such crosses. In two such back crosses :
there appeared caterpillars which had entirely new colors, pre-
sumably due to recombination, but unfortunately none of these —
survived until the imaginal stage.
A. H. STURTEVANT —

* Jour, Genet., 1, 293, 1911.
* Jour. Exp. Zool, 12, 499, 1912.

VOL. XLVI, NO. 550 OCTOBER, 1912

THE
.MERICA
NATURALIST

A MONTHLY JOURNAL 7
Devoted to the Advancement of the Biological Sciences with
Special Reference to the Factors of Evolution =

en
ER
ee
ES cea
pike

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.
Articles containing research work bearing on the problems of organic evolu-

tion ma AaS welcome, and will be given sroteeeaee in publication.
rea reprints of contributions are supplied to authors free of charge.

Fo ants will be supplied at cost

Subscriptions and a dvertisements should be sent to the lp ye The
subscription price is raii Pov ae ear. Foreign pantser

ay s fifty
e cents additional. The c = for single copies is
forty cents. The advertising rates are Four Dollars for a

THE SCIENCE PRESS

Lancaster, Pa. Garrison, N. Y.
S _ NEW YORK: Sub-Station 84 :
E ter, April 2, 1908, at the Post Office at Lancaster, Pa., under the Act ot
Congress of March 3, 1879.

5 Anor . FOR SA TENTH EDITION.
: IC, ICELAND d GREENLAND
| BIRDS" SKINS, THE MICROSCOPE,

an introduction to Microscopic Methods and to Histologn,
Weil Prepared Low Prices by Son Haney GAGE, 2E, of Cornell f r e in bhis new
: Particulars of ry tally evel oicion. aoe se oo
a G DINESEN, Bird Col CO., Ithaes, N Y
= Husavik, North Icelana, Via Leidie, England COMSTOCK PF UBLISHING CO.» :

oohives of Psychology

RECENT NUMBERS ©
19. An Empirical Study of € or Gortain Tests s e

T ; and Intens;
: Pp. 39. Scents oloti pason ai gi
= BrE on ga

si s

THE
AMERICAN NATURALIST

VoL. XLVI October, 1912 No. 550

GAMETIC COUPLING AS A CAUSE OF
CORRELATIONS

G. N. COLLINS
Bureau or Puant Inpustry, U. S. DEPARTMENT OF AGRICULTURE

\ Waen two distinct characters consistently occur to-
gether in the offspring of a hybrid, the phenomenon is
termed in Mendelian parlance ‘‘gametic coupling.’’ If
the characters sometimes occur independently, but appear
together more frequently than they should by chance, the
term ‘‘partial gametic coupling’’ is applied. If the char-
acters occur together less frequently than is to be ex-
pected, the condition is termed ‘‘repulsion,”’ while if they
never occur together in the same individual it is called
‘‘spurious allelomorphism.”’ .

In mathematical language, characters that show par-
tial gametie coupling are said to be ‘‘positively corre-
lated,’ those that show repulsion, ‘‘ negatively corre-
lated.” If they always occur together in the same indi-
vidual, the correlation is said to be ‘‘perfect’’ or ‘‘com-
plete,’’ while if they never occur together the negative
correlation is perfect.

When characters that show a positive correlation or
partial gametic coupling are derived from the same
parent, they are termed ‘‘coherent’’ by Cook (1909, p.
16). Many of the examples of partial gametic coupling
formerly reported in Mendelian literature were probably
of this nature, but with the attention foeused on the idea
that all cl ters were independent units, the possibility
of such coherence was not considered, and it is sometimes

569

570 THE AMERICAN NATURALIST [ Vou. XLVI

not to be determined from the data presented whether
the correlated characters were derived from the same or
different parents. This phenomenon of coherence, which
appears to most field naturalists and practical breeders
as a well-known fact, strikes the followers of Mendel as a
novel idea, to judge from the following (Bateson & Pun-
nett, 1911, p. 6):

The fact, however, that the mode in which factors are combined in
the original parents ean influence the distribution of the factors
among gametes of F, introduces a new conception into genetic phys-
iology.

The difficulty of ‘‘breaking up” combinations of char-
acters has so long been a stumbling block to breeders that
to them the conception can hardly be considered new.
The suggestion that the number of individuals in which
two characters are combined bears a definite relation to
the number in which they occur singly is without doubt a
direct outgrowth of Mendelian investigations and meth-
ods of thought. Perhaps this latest application of
mechanical conceptions to biology may stimulate research
as did Mendel’s original discovery. On the other hand, —
there are many who think that the application of Mende-
lian formule has already been pursued to an absurd
point by the factoring and subfactoring of characters and
the assumption of intensifying and inhibiting determ1-
nants. Those not already abreast of Mendelian literature ~
are not likely to be impressed with the further refinement
that aims to devise formule for expressing gametie
relations between two of these already complicated Men-
delian systems. a

In considering the relation between two Mendelian
characters four possible combinations are involv

hus, if A and B be taken to represent the appearance 0!
the two characters, and a and b their non-appearance,
these four combinations would be expressed as follows:
AB, Ab, aB, and ab. The theory of gametic coupling
assumes that an attraction or other unknown relation
exists between the determinants of the two characters,

No. 550] GAMETIC COUPLING 571

which results in their occurring together in gametes more
frequently than they occur separately.
`` There appears to have been a preconceived idea that
the ratio between the number of gametes in which the
determinants for two characters occurred together and
the number in which they were separated would be either
as 3:1, or 7:1, or 15:1, ratios that are common in Mende-
lian inheritance of single characters. The reason for this
expectation is not apparent.

The persistence with which it is sought to utilize the
numbers 4, 8, 16, etc., representing powers of 2, in the
interpretation of correlations reminds one of the argu-
ments used at the dawn of science to bring all natural
phenomena into some relation with the numbers 4 and a
Galileo’s suggestion that there were more than seven
planets was answered by Lizzi:

There are seven windows in the head, two nostrils, two eyes, two
ears, and a mouth, so in the heavens are there two favorable stars, two
unpropitious, two luminaries, and Mereury alone undecided and in-
different. From which and many other similar phenomena of nature,
such as the seven metals, ete., which it were tedious to enumerate,
we gather that the number of planets is necessarily seven. (Snyder,
1907, p. 203.)

The levity of the comparison disappears when one
seeks a reason for assuming that the number of gametes
in which *the characters occur together will be to the
number of gametes in which they occur apart as 7:1 or
15:1 or 31:1, etc., and not some intermediate proportion.
~ In the following pages an attempt is made to review
briefly the experiments: which have been advanced as
proof that the various degrees of correlation fall into
this definite series.

The first attempt to refer the correlation of characters
to definite differences in the gametes was made by Bate-
son, Saunders, and Punnett (1906, p. 9) in explaining
observed correlations between purple flower color and
long pollen grains in hybrid sweet peas.

. A close approach to the Observed F, numbers would be given

572 THE AMERICAN NATURALIST [ Vou. XLVI

by a system in which each 16 gametes were composed, thus: TAB +
laB + 14b + Tab, where A is long pollen and a round pollen.

Purple Red | White

Long | Round | Long | Round | Long | Round
Observed ...... tee 18 | 7 381 | 1,199 | 394
Calculated ..... 1,448.5 | 122.7 | 122.7 | 401.5 | 1,220.5 | 407.4

The numbers under ‘‘White’’ may be disregarded in
this connection since the distribution of long and round
pollen grains in this group shows a close approach to the
normal 3:1 ratio.

Some of the numbers in the ‘‘observed’’ series are
above and some below the corresponding numbers of the
‘“ caleulated’’ series. That the calculated series approx-
imates the observed series is obvious, but there is no way
of determining the degree of this approximation. No
method has been proposed for making definite compari-
sons between such series of numbers. Without some
standard of comparison it is difficult to see that anything
is gained by resorting to gametic formule to represent
the degrees of association between characters.

A customary and direct method of comparing the de-
grees of relationsbip that exists between any two charac-
oie is to compute the coefficient of correlation or Yule’s

‘‘coefficient of association.’’

In the following discussion Yule’s ‘‘coefficient of asso-
ciation”? (1900) is used. By this method the complete —
independence of two character pairs is represented by 0,
complete association by 1. Intermediate degrees of re-
lationship are expressed by the intermediate decimals. —
If the four classes of individuals are represented g a; b,
c and d, the coefficient of association is

(a xX d)—(bXc)
(axd t Oxo)
Since this coefficient can be computed directly from the
observed numbers the predication of gametic formule af
a means of expressing degrees of association of charac: :
ters becomes unnecessary.

No. 550] . GAMETIC COUPLING 573

With relationships expressed as coefficients of asso-
ciation, probable errors can be calculated. Thus it be-
comes possible to determine whether the approximations
between observed ratios and those calculated from the
different gametic formule are closer than would result
from chance.'

The difficulty of securing a 7:1 distribution by dichot-
omous cell division was appreciated by Bateson and his
collaborators in the example first cited. The possibility
that the coupling was in an 8:1 ratio was also suggested
and kept in mind for a time until a grouping that approx-
imated that resulting from a gametic coupling of a 15:1
ratio was obtained. Regarding the choice between the
7:1 and 8:1 ratios to explain the numbers observed in
the first experiment, the authors remarked (Bateson,
Saunders and Punnett, 1908, p. 3):

. we are still unable to decide finally between them.

But this indecision was apparently overcome in a subse-
quent paragraph on the same page. After citing two
examples that were referred to a 15:1 ratio, the theory
that the couplings follow a definite series was launched
in the following statement:

The undoubted existence of these two grades of gametie coupling in
the Sweet Pea suggests that each may find its place in a scheme of
increasing intensity of gametie coupling, such as is shown in the
accompanying table, where the two allelomorphie pairs are represented
by Aa and Bb:

1 Heron has pointed out (1911, p. 109) that the results secured by Yule’s
formula for the ‘coefficient of association’? do not approximate the true
coefficient of correlation, except where the two divisions are near the mean
of the entire population. This condition is not met, of course, by characters
that oceur in one fourth and three fourths of all the individuals. But since
the present discussion is confined to examples that approximate the form
3n? — (2n — 1) :2n —1:2n— 1:n?— (2n— 1), where 2n = the number in
the gametie series, they afford a regular sequence and, though the actual
values are somewhat arbitrary, they should not be misleading as a means of
comparing the degrees of relationship represented by the different gametic
formule. Thus if the classes observed in an experiment show the same
coefficient of association as the classes calculated from some particular
gametic ratio, the two series will also give the same coefficient of correla-
tion, though the actual values may differ.

574 THE AMERICAN NATURALIST [ Vou. XLVI

pameni oo Zygotes yrs mag | peee
AB ab Aand B ADA. B oT A nor
1: 7. cy i 4 9 3 3 hy z> |
ariel: = B 41 : 7 To s 9 =
Tii- T16 177 eb ip : 49 == S
ei si: D-3 737 See S 3 Si 22 024
m—1:1:1:n—l=2n 3n?—(2n—1) : ma. 2n—1 : n?—(2n—1)=—4n?

The first term in the series is a simple case of dihybridism in nied
no coupling exists. The second term we have not yet encountered.
But we have an ample series of experimental data which satisfy the
third term; and the experimental evidence for the existence of the
fourth term rests upon two independent cases.

To facilitate the use of the coefficient of association as
a means of determining what gametic ratio most closely
approximates observed ratios, Table I has been pre-
pared. This table shows a series of gametic ratios in
which the less frequent combinations are taken as 1, and
the resulting zygotic series, with the number of indi-
viduals that would occur in each of the four classes, fol- —
lowed by the coefficient of association, calculated from
the numbers in the zygotic series. The ratios in heavy
type are those representing the powers of two to which
correlations have been referred under the theory of —
gametic coupling.

TABLE I
RATIOS OF GAMETIC COUPLING

Coefficient of
: nes Series Zygotic Berkes Association
123.449 0:3 :8:4 0
oiii oe Eo 558
auS i154 +3 41:7:7: “ 766
“a a 66:9:9:1 858
O45 145 97 oe 905
Gt st G 134 : 13 : 13 : 36 .932
7 1 i:? 177 : 15 : 15 : 49 -949
S-1 11.8 226 : 17:17: 64 961
9- 1:1:9 281 : 19 : 19 : 81 969
12s 10 342 : 21 : 21 : 100 975
coe re ee 409 : 23 : 23 : 121 979
B:1i:1i- p 482 : 25 : 25 : 144 982
13:1:1: 13 561 : 27 : 27 : 169 .985
M:i: 1: i 646 : 29 : 29 : 196 .987
w:1:1i: B Tal * S1 : 931 : 95 .
16 :1:1: 16 834 : 33 : 33 : 256 990
pS eee ry : 68 : 68 : 961 :
S:1:1:6 12,161 : 127 : 127 : 3,969 -9993
197: i:i- : 255 : 255 : 16,129 9998

No.550] GAMETIC COUPLING 575

Returning to the consideration of the original example
of gametic coupling in the sweet peas, the coefficient of
association, .958 + .004, is seen to be intermediate be-
tween that resulting from a 7:1 and an 8:1 ratio.

Notwithstanding the fact that the figures correspond
somewhat more closely with an 8:1 than they do with a
7:1 ratio, an evident preference for numbers that are
powers of 2 is shown when the possibility of an 8:1 ratio
is discussed. Instead of saying that 18 gametes are con-
cerned, the number is spoken of as 16+ 2, and the fact
that a ratio of 15:2 would give almost exactly the ob-
served association is not even considered.

The next example of gametic coupling to be reported
was in the same series of crosses (Bateson, Saunders
and Punnett, 1908, p. 11), where the progeny of four
individuals showed the following grouping: 296:19:27:
85. This is referred by the authors to the theoretical 7:
1 ratio, though it shows almost the same association as
an 8:1 ratio, that is, .960. The probable error, .008,
would indicate, however, that with this number of indi-
viduals such a deviation might easily be due to chance.
The numbers are, therefore, too small to afford evidence
affecting the choice between a 7:1 and an 8:1 formula.

From the progeny of one of these four individuals
consisting of 111 plants and showing an association of
.914, 10 individuals showing both dominant characters
were selected and propagated. The progeny of the 10
plants taken together gave the following grouping: 493:
25:25:138. This grouping is considered only in con-
nection with the 7:1 and 15:1 distribution, though the
association .982+ .004 would indicate that a gametic
series of 12:1 would most closely fit the numbers. The
deviation from the 7:1 ratio is 9 times the probable
error, and from the 15:1 ratio, about twice the probable
error.

With respect to the characters considered separately,
the classes secured from each of the 10 individuals
showed a remarkable conformity to the expected 3:1

576 THE AMERICAN NATURALIST [ Vou. XLVI

ratio. In the degree of association between the char-
acters, however, no such uniformity was exhibited.
Leaving out of consideration two families represented
by only a few individuals, the coefficient of association
varies from .910 to .990. In terms of gametic coupling
this shows a range from 6:1 to 16:1, and since the series
as a whole accords with a 12:1 ratio it is not apparent
why only 7:1 and 15:1 ratios are considered. :

Individuals were again selected from two of the fam-
ilies which showed the highest correlation, and grown the
following season. The progeny from the first family
behaved irregularly and the presence of some disturb-
ing process was suspected, though the deviations with
respect to the individual character pairs were less than
the probable error.

The second family gave individuals with the follow-
ing grouping (Bateson, Saunders and Punnett, p. 12):
983:26:24:170 (association .987 + .0026). Of this it
is said:

It is obvious that the numbers in this group of families accord very
closely with the figures expected on a 15 : 1 : 1 : 15 basis, and the
view that this is the system actually followed receives confirmation
from the distribution of the pollen and color characters in F, families
from the Bush X Cupid crosses where the following figures were ob-
tained: 131:6:5:42. (Association .989 = .005.)

Here again when it is said that the figures ‘‘ accord
very closely,’’ it can only be meant that they accord
closely with one of the formule in the hypothetical series
as compared with other members of the series. The
numbers are, of course, inadequate to afford evidence as
to whether the observed figures accord more closely to
those resulting from a 15:1 combination than they do,
for example, to a 16:1 or a 14:1, yet the results are
taken to support the original assumption that the group-
Ings are in powers of two.

The examples of gametic coupling thus far reported —
are summarized by Bateson and Punnett as follows
(1911, p. 5):

No. 550] GAMETIC COUPLING 577

. No ease yet known.

Sweet Pea. Blue factor and long pollen.
Primula sinensis. Magenta color and short style.
Sweet Pea. Fertile anthers and dark axils.

No ease yet known.

. Pisum. Development of tendrils and round seed.
. Sweet Pea. Blue factor and erect standard.

Oo Ge
m Go + Or sy =) GO
Tea Sep E

n
bo

As we have seen, the first example is closer to an 8:1
ratio. The second example is apparently based on an
experiment comprising 47 individuals (Gregory, 1911,
p. 12). The classes were 33:3:1:10, and the author
states,

.. the partial coupling observed is almost entirely certainly of the
for T 1:1

The reason for this assurance is not apparent since the
grouping is really nearer to that resulting from a 15:1
ratio. The probable error, .015, is so large, however,
that it would be impossible to determine the grade of
coupling closer than to say that it probably falls some-
where between 8:1 and 31:1.

The third example referred to the 15:1 ratio appears
to be based on 885 individuals showing an association of
.993 + .0017, indicating a coupling of about 20:1, the
deviation from the 15:1 association being 2.9 times the
probable error.

With respect to the examples where the association is
closer, none of the reported experiments have been con-
ducted on a sufficient scale to determine whether the
gametic ratios that are multiples of two are approxi-
mated more closely than other ratios. One of the most
exact approximations thus far reported is that of Vil-
morin and Bateson (1911, p. 10) where the numbers 319:
4:3:123 (association .9994+.0003) were obtained.
This is certainly a close approximation to the figures
that would result from a gametic coupling of 63: 1:1: 63,
which for this number of individuals would be 333.27:
3.48:3.48:108.77. Yet the observed numbers can be still

more closely approximated by assuming that the coup-

578 THE AMERICAN NATURALIST [ Vou. XLVI

ling was in the proportion of 75:1, which would give the
figures 334.57 : 2.98 : 2.98 : 109.47. The observed numbers
are somewhat closer to the 63:1 than to the 127:1, the
proportion next above in the proposed series. But why
avoid the intermediates?

It should be kept in mind that the series was built up-
in the first place, not because the observed numbers
agreed with some member of this series more closely
than was to be expected by the laws of chance, but ap-
parently for a priori reasons, because the numbers in
this series were in accord with the Mendelian ratios for
the appearance of single characters which represent = |
powers of 2. oe

When the correlation is as high as in the above ex- —
ample it would require, not hundreds of individuals, but J
tens of thousands to prove that the observed numbers a9
were in accord with those resulting from any particular
gametic ratio. In the experiment referred to the 61:1
ratio, there are two classes represented by 4 and 2 in-
dividuals, respectively. A change of two or three in-
dividuals in each of these classes would cause the array
to correspond as closely with a 31:1 or 127:1 ratio as
it now does with the 63:1.

In the very nature of things any observed association
must fall nearer to some one of the calculated ratios
than to any other, yet in the discussion of these experi-
ments this seems not to have been appreciated. The
data would have supported in the same way any other
choice of preferred ratios. |

COHERENT CHARACTERS In Hysrips or CurnesE MAIZE

Beginning in 1908, experiments have been conducted
with a variety of maize secured from China. The endo-
sperm of this Chinese variety is of a peculiar waxy tex-
ture, a character thus far not reported in any other
variety.

There are two strains of the Chinese variety, Onè
white, the other with colored aleurone. In a series of

No. 550] GAMETIC COUPLING 579

hybrids between Chinese and American varieties made
by Mr. J. H. Kempton and the author, the waxy char-
acter was found to be definitely recessive in the first gen-
eration. In the second generation the character reap-
pears apparently unchanged in slightly less than 25 per
cent. of the individual seeds.

Crosses between white Chinese and colored American,
and colored Chinese and white American showed a pro-
nounced coherence between the texture of the endosperm
and the color of the aleurone layer.

Results of the first season’s work in this field were
reported at the International Conference of Genetics in
Paris, 1911. At that time no attempt was made to de-
termine the gametic ratios necessary to account for the
observed correlations.

In these crosses the characters segregate with defi-
nitely alternative expression, but the classes seldom
show simple Mendelian ratios. In only one cross were
two of the four classes even approximately equal.

In the results of the next season’s work there were
five additional ears sufficiently near the 3:1 ratio in both
characters to be considered from the standpoint of gam-
etic coupling. The data derived from these five ears,
and from the one of the previous season, are shown in
Table IT.

TABLE IT

7 re armena manaa panan Shs SAART ig ve v e

Ear | No. |Colored|Colored| White | White| Coefficient of | Deviation from|_ Deviation 1

No. gai Ros Waxy Horny kasd Association Tenn k iiia
152 | 183| 112| 20 | 22 | .761 =.049 w t ow
301 | 579| 372| 62 | 63 | 82 | .773«.020 .007 0.24
302 | 536 343 | 52 | 53 | 88| .833+.024 067 2.79
627, 409| 57 | 62 | 99 | .839+.021 073 3.48
325 | 650, 434| 55 | 61 |100 | .856+.019 .090 4.74
380 | 161| 104| 17 | 18 22 | .764+.058 .002 0.03
“Total 2736) 1774 | 263 | 279 |420| 8215011 | 055 | 500

The ratios obtained in these six cases might be hailed
as a prediction fulfilled, since in every case the numbers
approximated those that would result from a 3:1:1:3
ratio, which has remained a gap in the theoretical series.

580 THE AMERICAN NATURALIST [Vou.XLVE

In three of the six ears the deviation from the cor-
relation resulting from a 3:1 ratio is less than the prob-
able error. The deviation for the total is less than
three times the probable error, and might readily be
only a chance deviation. In two of the ears, however
(Nos. 303 and 325), the deviations are rather large to be
ascribed to chance.

In view of the fact that previously reported experi-
ments fail to show an equally close approximation to
other members of the proposed series, there is no ade-
quate reason for assuming that the present approxima-
tion to the numbers of a 3:1 ratio belongs to a series of ©
formule represented by the powers of 2. That no such
regularity exists in the interrelation of different char-
acter pairs is more definitely demonstrated by the ex-
periments described below. These results indicate that
the association of characters may be determined after
the conjugation of the gametes. .

COHERENCE oF CHARACTERS NOT ALWAYS THE RESULT OF

Gametic DIFFERENCES a
If correlations are in all cases due to gametic differ-
ences, there should be no correlations exhibited in a
cross with a Mendelian formula Aabb XaabB. In our-
experiments with corn this would be represented by a
cross between colored-waxy and white-horny, where the
first is heterozygous in aleurone color and the second |
heterozygous in endosperm texture. The colored aleu-

of the gametes should bear the white character, and
those of the male parent the waxy character. The =

No. 550] GAMETIC COUPLING 581

nificance of crosses of this nature was not realized at
the time pollinations were being made, and but 5 ears of
this kind were secured. In two of these ears there is a
significant correlation. The classes exhibited in the five
ears are shown in Table III.

TABLE III

Ear No. Gant Pha White | White Colored Colored | Coefficient of
No. Seed | White Waxy Waxy | Horny Waxy Horny Association

472 | 505 | 30.1 | 464 | 91 | 61 143 210 | .873+.057
471 | 368 | 48. | 480 | 109 | 68 70 121 | 47 *.056
314 | 283 | 44.5 | 52.7| 65 | 61 84 73 | —.038+.081
272 | 395 | 61.2 | 47.1 | 122 | 120 64 89 | .171+.068
256 | 141 | 468 | 489 | 27 | 39 42 33 | —.295 =.105

The two ears in which the correlation appears sig-
nificant (Nos. 471 and 472) have a common ancestry,
different from that of the other ears. The history of
these ears is as follows: In 1908 a plant of a white Mex-
ican variety was pollinated by a Hopi variety with col-
ored aleurone, producing a pure white ear, Mh19. In
1909 a plant from Mh19 was pollinated by white Chinese,
the variety possessing the waxy endosperm. The result-
ing ear Dh14 had white and colored seed inthe proportion
of 2 white to 1 colored, with no trace of the waxy endo-
sperm. In 1910 a plant from a colored seed of Dh14 was
self-pollinated, producing an ear Dh142L2 with the fol-

lowing classes: white-waxy 99, white-horny 89, colored
waxy 51, colored horny 348. This represents an associa-
tion between colored and horny of .767, almost exactly
that expected on a coupling ratio of 3:1, but it will be

582 THE AMERICAN NATURALIST [Vou XLVI

noticed that the two middle classes are not equal. In
1911 two plants grown from white horny seeds of
Dh142L2 were pollinated by a plant from a colored waxy
seed of the same ear, producing the two ears, Nos. 471
and 472. The pedigree of these two ears is graphically
shown in Diagram 1.
The plants which bore ear 472 bore also a second ear
which was self-pollinated. Contrary to expectation,
this ear showed slight traces of aleurone color. That
there was a tendency to produce aleurone color in this
extracted recessive is also indicated by the low percen-
tage of the total white seeds in ear 472, which is 30 in-
stead of 50.
With ear 471 there is every indication that one of the
parents was homozygous with respect to the recessive
color character and the other with respect to the texture
of the endosperm. A self-pollinated second ear from
the plant that produced ear 471 had 24 horny seeds and
5 waxy, all of them white. A self-pollinated ear was
also secured from the plant which was the male parent
of both 471 and 472. This ear had 31 white seeds and
128 colored, all of them waxy. The total percentage of
white seeds in ear No. 471 was 48, a close approximation
to the expected 50. ,
The nature of this experiment seems to preclude the
application of the theory of gametie coupling which —
would explain correlations by assuming attractions Or —
repulsions between the character determinants in the
gametes. While attractions which might cause some
combinations to be represented by larger numbers of
gametes would disturb the Mendelian ratios, they could
not produce the effect of correlation or coherence of
characters. o
Thus the female parent might produce more gametes
bearing white and waxy than it did colored and waxy, m
there might be a greater fatality among the colored
waxy gametes. Any such disparity in the classes of
the gametes would result in more than 50 per cent. of

No. 550] GAMETIC COUPLING 583.

the zygotes showing the recessive characters, but even
so, there would be no correlation.

On the presence and absence hypothesis, a positive
correlation between two dominant characters like aleu-
rone color and horny texture must be considered a dif-
ferent phenomenon from a positive correlation between
color and waxy texture (Bateson, 1909, pp. 151, 158).
The first case has to be looked upon as an example of at-
traction, the second of repulsion. This appears an un-
necessary complication, as Emerson (1911, p. 79) has
pointed out. The significant fact would seem to be that
both cases may be considered as examples of coherence.

Appreciation of the fact that many of the cases of
gametic coupling are examples of coherence makes any
assumed regularity in the degree of correlation still
more absurd. But in spite of cases of coherence that
result in reversal of correlation, the idea that the char-
acters are represented by material particles that remain
unchanged by association in the zygote continues to be
held. To preserve the conception of pure unit charac-
ters, theories of positional relations of unit characters-
are now being proposed. Simple Mendelian ratios can
also be explained by positional relations of character de-
terminants as well as by theories of alternative trans-
mission (Swingle, 1898; Cook, 1907, p. 353), and this view
has the advantage that it can accommodate facts which
indicate a complete transmission of characters.

A theory to explain how positional relations of de-
terminants in the chromosomes may be responsible for
the phenomenon of coherence has been proposed by
Morgan (1911). The suggestion is that coupled or pos-
itively correlated characters may lie close together in
the chromosome and that in the conjugate generation
pairs of chromosomes are twisted around each other.
If, as claimed, the chromosome pairs split in a single
plane, characters which lie close together in the original
chromosomes would seldom be separated, while char-
acters remote from each other in the chromosome would
stand an even chance of being separated.

584 THE AMERICAN NATURALIST [ Vor. XLVI

If the position of the various determinants in the
chromosomes is definite, only slight variations in the de-
gree of association among hybrids between the same
parental strains would be expected, but there would be
no reason to expect that the different degrees of correla-
tion should fall into a definite series.

SELECTIVE PoLLINATION

From the standpoint of complete segregation or alter-
native inheritance of characters there remains the pos-
sibility of correlations being the result of selective pol-
lination. It is conceivable that the ovules which bear
one of the segregating characters are more readily fer-
tilized by pollen bearing another character with which
it was associated in the parent. As applied to our own
experiments, ovules which are potentially waxy might
be more readily fertilized by pollen which is potentially
white.

If correlations are the result of selective pollination,
no correlation should be shown as the result of crossing
an individual that is heterozygous with respect to both
of the character pairs with an individual showing both
of the recessive characters, for in such a cross the gam-
etes of one parent would be all of one kind. In our own
experiments, crosses between plants from hybrid seed
showing horny endosperm and colored aleurone (both
dominant characters) with plants from white waxy seeds
(both recessive characters) should throw light on this
point.
If the correlations result from numerical inequalities -
in the classes of the gametes, the seed classes resulting
from such a cross as that described above should be the ©
Same as the gametic classes produced by the heteroZy—
gous parent. Thus, if the seed classes in a self-pollin-
ated ear correspond to those expected from a 3:1 ratio, :
the same plant when crossed with a plant showing both ;
recessive characters should show classes in the ratio, —
3 white-waxy, 1 white-horny, 1 colored-waxy, and 3 col

No. 550] GAMETIC COUPLING 585

ored-horny. On the other hand, if the correlations re-
sult from selective pollination, no correlation should ap-
pear in such a cross since the pollen is all of one kind.
Ears that represent crosses of this kind are described in
Table IV. It will be seen that in all of the seven ears
there is a significant correlation.

TABLE IV
Colored | Colored White Whit Coefficient of
Ear No. No. Beeds H nt Waxy Horny Waxy Association
144 600 125 40 232 203 464 + .054
250 245 19 74 104 48 —.788 + .04
251 377 52 29 138 158 B45 + .077
390 458 115 35 94 214 764+.
391 231 65 10 49 107 868 = .032
392 316 84 16 68 148 840 = .051
401 518 3 29 263 223 —.839 +.061

It is interesting to note that in two ears, Nos. 250 and
401, while the correlation is relatively high, it is reversed;
i. e., the positive correlation is between colored aleurone
and waxy endosperm instead of between colored aleurone
and horny endosperm as was the case in the original F,
ears. Whether the correlation is positive or negative
does not affect. its use as evidence in eliminating the pos-
sibility of selective pollination, so long as the correlation
is significant. From the standpoint of the presence and
absence hypothesis, the coupling or attraction has given
place to a repulsion.

The results shown in Tables III and IV indicate that
the association of different characters may be deter-
mined at different stages in the ontogeny of the indi-
vidual, much as the appearance or non-appearance of a
simple character may occur at different times in the life
history. Thus in ears 471 and 472, Table III, the corre-
lated characters must have become associated after the
formation of the gametes, while the absence of positive
correlation in ears 314 and 317 of the same table shows
that in these ears where associations in the gametes were
excluded no correlations were subsequently produced.
The results in Table IV further show that excluding the

586 THE AMERICAN NATURALIST [Vou. XLVI

possibility of selective pollination did not prevent the
appearance of significant correlations. It appears no
more reasonable to assume that the number of individuals
that are to show a particular combination of characters ©
is always determined at the time the gametes are formed,
than it would be to argue that the number of nodes des-
tined to bear a juvenile type of foliage are definitely de-
termined at this time.

The theory of gametic coupling advanced by Bateson
and Punnett leads them (1911, p. 6) to entertain the idea
that, while segregation is definite and complete, the ap-
parently significant cytological processes of maturation
may have nothing to do with the phenomenon. These
authors would have the association of characters deter-
mined before the maturation divisions.

Now that we know of a series involving as many as 256 terms
(127 +1+1-+127) it is most difficult to conceive that such a se
tem can be produced in the maturation-divisions of the ovarian tissue
of such a plant as a sweet pea. We may well be tempted to look
much earlier in the developmental processes for the establishment of
these differentiations, and it is not impossible that they may be estab-
lished as early as the embryonic constitution of the sub-epidermal
layer itself.

The suggestion that segregation occurs early 1m be
ontogeny of the individual was apparently occasioned by
the belief in the definiteness of the mathematical rela-
tions between different character pairs. As We have
seen, there is little evidence for this belief. The correla-
tion shown in ear 471, Table III, affords a definite indica-
tion that such associations of characters may be form
after the production of the gametes.

Even though the definite and material segregation of
characters be maintained, all grades of correlation could

still be determined during the two nuclear divisions that 7

follow synapsis. The process might be looked at aS
follows:

When two character pairs are involved in a hybrid, the ss
tetrads resulting from the different mother cells WOU

No. 550] GAMETIC COUPLING 587

all be of three types with respect to the two characters
involved: (I) AB, AB, ab, ab; (IL) AB, Ab, ab, aB; (III)
Ab, Ab, aB, aB. Each of these fulfills the condition
deemed necessary from a Mendelian standpoint, that two
daughter cells of each tetrad receive the dominant and
two the recessive determinants. Where A and B are
correlated, it may be assumed that tetrads represented
by (III) are not formed. If the two remaining kinds
were produced in equal numbers, a gametic ratio of
3:1:1:3 would result. By the formation of two mother
cells of type I for every one of type II, a ratio of 5:1:1:5
would result. The formation of these two types of mother
cells in different proportion would provide for all de-
grees of correlation.

Additional examples of gametic coupling, some of
which, at least, are in the nature of coherences, are re-
ported by Gregory (1911, B, p. 128). Asin our experiments
with hybrids of Chinese maize, examples of gametic
coupling were encountered of a lower order than 3:1:1:3,
the lowest ratio in the theoretical series. Even this dis-
covery did not result in his questioning the validity of
the hypothesis. A further refinement was devised to
accommodate the results. Advantage was taken of the
possibility that the gametic coupling may exist in only
one sex:

For the time being it may be pointed out that a very close approx-
imation to the observed numbers is given by the assumption that a
coupling of the form 7 : 1 : 1 :7 is present in gametes of one sex only,
gametes of the opposite sex being produced in equal numbers of all
four kinds.

In such a case, a gametic coupling of the 7:1 in one sex,
with all classes equally represented in the other sex,
would result in an association of .542, slightly lower than
that resulting from the ordinary 3:1 ratio. By a similar
fractioning of the 3:1 ratio, the series can be provided
with a still lower member, with an association of .390.

In all recently reported experiments the reality of the

588 THE AMERICAN NATURALIST [ Vou. XLVI

proposed series of ratios seems to be taken for granted
and results are viewed only from this standpoint.

It is hoped that the previous discussion may operate to
check this practise of referring examples of coherence
to preconceived gametic ratios without considering the
possibility that an apparent agreement may be only a
chance approximation.

So long as the preconceived series is taken for granted
and intermediates are not considered, the results of all
experiments will seem to give additional evidence in sup-
port of the series.

We have seen that the results of previously reported
experiments do not correspond to the ratio in the ex-
pected series more closely than they do to others that are
intermediate, and furthermore that in some cases, at
least, the associations are not due to relations of the
determinants inside the gametes at all, but occur after
the stage of karyapsis, or nuclear fusion, has been
reached (Cook and Swingle, 1905).

SUMMARY
The theory of gametic coupling assumes that correla-
tions between two Mendelian character pairs are caused
by attractions or, repulsions between character-units K
determinants, previous to the formation of the germ cells. _
These attractions or repulsions are supposed to increase
the number of gametes bearing certain combinations of
determinants. o
The further assumption that the various degrees of :
association observed between different character-palts
will fall into a regular series represented by powers of 2,
as in simple Mendelian hybrids, appears to have been
accepted without adequate analysis of the data on which
it was based. a
An examination of the early examples shows that it
was only by neglecting the possibility of intermediate
ratios, and thus begging the question, that the observed

No. 550] GAMETIC COUPLING 589

numbers could be said to agree with those of the proposed
series.

‘The lack of any standard or method for making quanti-
tive comparisons between observed and expected series
has made it impossible to determine the degree of the
supposed approximations. Yule’s coefficient of associa-
tion is proposed as a criterion of comparison, and to
make possible the determination of probable erors.

In several cases correlations have been found to be
reversible, depending on the way the characters were
combined in the parents. This fact has further compli-
cated the theory of gametic coupling, making it necessary
to assume that characters which at one time attract each
other, at others exhibit repulsion.

In hybrids between Chinese and American varieties
of maize coherence has been found between the texture of
the endosperm and the color of the aleurone layer. Ina
few cases, the degree of the correlation approached very
closely to that expected from a gametic coupling ina 3:1
ratio (Table II).

Correlations were found in crosses of the Mendelian
form Aabb X aabB (Table III). Such correlations are
held to indicate that in some cases at least, the correla-
tion between the characters must be determined after the
formation of the gametes.

On the other hand, correlations resulting from crosses
of the form AaBb X aabb eliminate the possibility of
selective pollination as a general cause of correlations
(Table IV).

The general conclusion is reached that associations
between characters, like the appearance of single charac-
ters, may rise at different stages in the ontogeny of the
individual.

WASHINGTON, D. C.,
March, 1912

590 THE AMERICAN NATURALIST [ Von. XLVI

LITERATURE CITED
Bateson, W.
Mendel’s Principles of Heredity. Cambridge.
Bateson, W., and Punnett, R.
1911. On the Inter- AE EA of Genetic Factors. Proc. Roy. Soc., B,
Vol. 84, ar C pp. 3-8.
Bateson, W., Saunder and Punnett, R. C.
1 Sasser N peas in a cea ere of Heredity. Report
III, Evolution Committee of Roy. Soc., pp. 1-53.
1908. Experimental Studies in the Eilear of ere Report
IV, Evolution Committee of Roy. Soc. 1—60.
Cook, O. F.
1907. Aspects of Kinetic Evolution. Proc. Wash. Acad. Sci., Vol.
é 7-403.
1909. Suppressed and Intensified Characters in Cotton Hybrids. U. 8.
Dept. of Agriculture, Bureau of Plant Industry, Bulletin 147.
Cook, O. F., and Swingle, W. T.
1905. Evolution of Cellular Structures. U. S. Department of Agri-
ulture, Bureau of Plant Industry, Bulletin 81.
Emerson, R. A.
1911. Genetic Correlation and Spurious Allelomorphism in Mai
wenty-fourth Annual Report of the Nebraska avin
Experiment Station, pp. 58-90
Gregory, R. P.
1911. A. Experiments with Primula — Jour. Genetics, Vol.
y No. 2, pp. 73-132, Mar
B. On Ganette Coupling and ee in Primula Sinensis.
Proc. Roy. Soc., B, Vol. 84, No. 568, pp. 12-15.
Heron, D.
1911. The Danger of Certain Formule Suggested as Substitutes for
the Correlation Coefficient. Biometrika, Vol. 8, pp. 109-12
Morgan, T. H.
1911. Random Segregation versus rie in Mendelian Inheritance.
Science, N. S., Vol. XXXIV, p. 384
Snyder, C.
7. The World Machine.
Swingle, W. T.
1898. Some Theories of Heredity and of the Origin of Spá cies Con-
sidered in Relation to -e Phenomena of Hybridization. Bot.
Gazette, Vol. 25, pp. 113.
Vilmorin, P. de, and Bateson
1911. A Cis of Gametic Coupling in Pisum. Proc. Roy. Soc. B,
ol, 84, No. 568, pp. 9-11
Yule, G. U.
1900. On the Association of Attributes in Statistics. Phil. Trans.
Roy. Soc., A, Vol. 194, pp. 257-319

MICE: THEIR BREEDING AND REARING FOR
SCIENTIFIC PURPOSES!

DR. J. FRANK DANIEL

UNIVERSITY OF CALIFORNIA

I. INTRODUCTION

NotTWITHSTANDING many shortcomings mice have con-
tributed much to the advancement of science and the serv-
ice of mankind. To realize this we have but to recall
that it was by crossing the white with the gray mouse
that Mendel’s Law of Inheritance was first found to
apply to the animal kingdom (1); that from a study on
mice some of the earliest concepts of immunity were ob-
tained (2); and that from experiments now in progress
on them an insight is being gained into the nature of
cancer (3). These and similar experiments indicate
something of the scope to which these animals have been
put.

The ease with which white mice can be handled poe
them, in many ways, preferable for experi n to
vihor and larger rodents. But, owing to a dasni
notion that they are difficult to rear under laboratory
conditions, their usefulness has been greatly curtailed.

The method usually employed in the breeding of mice
has been what we may term extensive. By this I mean
that many animals are kept from which to obtain off-
spring. I have set myself the task of breeding mice in-
tensively, that is, of keeping relatively few, but of keep-
ing these under dantition: which will insure their pro-
ductivity. It is the purpose of this paper to describe the
way in which this was done.

Il. Tue Iytenstve Breeprne or MICE
A. Detrimental Factors
l. Marked Fluctuations in Temperature.—Probably
no single factor is more likely to be overlooked than that
Mice, to produce to the best advantage, require an
- equable temperature. While they can withstand extremes
“From the Zoological Laboratory, University of California. |
591

592 THE AMERICAN NATURALIST [ Vou. XLVI

of heat or cold, such extremes are not conducive to their
productivity. At 35° C. I have found breeding to be
greatly impeded, and at a temperature as low as 2° C.
the young born are subject to a number of mortal ills
which practically prevent their reaching maturity. But
a constant temperature of either of the above extremes
is not so detrimental as is great fluctuation in tempera-
ture. A mouse taken from favorable conditions and sub-
jected to daily fluctuations of from 30° C. to 2° C. soon
becomes a different animal physiologically. The fur
which was sleek and glossy roughens, the exposed veins
in the ears and tail darken, and the animal is readily re-
duced to a condition which, if prolonged, not infrequently
terminates in death. If after having reached this con-
dition, however, the animal be promptly restored to
equable conditions of temperature, its fur becomes sleek,
signs of anemia disappear and the mouse regains its
normal health and vigor often with surprising rapidity.

2. Parasitism.—lf mice, even under the most favor-
able conditions of temperature, become badly parasitized
breeding ceases and unless they are ridded of the para-
sites the adult mice as well the young fall victims to this
pest. To test the effects of parasitism, I have taken
mice from fresh stock, kept under excellent conditions,
and have placed them in infested nests, with the result
that in a few days the mice became sluggish, and many
sooner or later died.

When the two factors—parasitism and wide fluctua-
tions in temperature—are combined, the animals, espe-
cially the young, die in great numbers.

B. Factors Essential to Intensive Breeding
1. Construction and Equipment of Cases.—In general,
where a number of mice are to be kept together, the wire
and wood cases described by Yerkes (4) has been much
used. But for intensive breeding I have found it better
to keep few mice in a case, and to keep these under better
conditions of sanitation than is possible in the above ease.
I know of no better plan to insure sanitation than to
construct cases which will offer little surface upon which

No. 550] BREEDING MICE 593
dirt may collect, and which at the same time will make

evident that which has accumulated. Such a case should
be made with a perforated bottom and should be pro-

(A j E E. g CA g

L

sH

1

; ()
E

'
yf

O Q
| O

p O

SF

FRONT VIEW or CASE. W.R., water receptacle; M.R., milk recep-
dicts: ri PEE to nest; F, opening to food.

Na a a a

i
9

oe

6-3 Wes View co Caan. FF, food funnel; M.R., milk receptacle; N. B.,
nest box: Gb, galvanized iron back (near its bend) ; F, entrance to food cup BO.
vided with sides of glass. Briefly described,’ the case
that I have constructed consists of a framework of light

. Wood, a back of galvanized iron and sides, front and par-

*For detail see description accompanying drawings 5h ses nae

ji

594 THE AMERICAN NATURALIST [ Vou. XLVI

tition of 10 X 12 glass. The top is made of screen wire;
the bottom of + inch wire mesh (hardware cloth).

WR MR

ee) OE) CO. Ooa

Screen Wire

Lii W \
L

Fic. 3. ToP oF CASE, WITH A PART REMOVED TO SHOW How THE GLASS
PARTITIONS AND ENDS (X) ARE FIXED INTO THE FRAME. WR., water recep-
tacles; T, a bent piece of tin covering the upper end of the glass to increase
the height of the partition and at the same time to cover the sharp edge of
the s.

AA E a SERS
<? ?,
ee a

aS

EuGl

>
SSK x

SRE SS

IAF i Sc SS een:

Oe x S
SE
RR
BSS SS
BOERS

res

2?

0,
Xx
oO

are
<>

Fic. 4. Borrom Virw or THB Case. The wire mesh is partly drawn in tò
Show its relation to the end (End. gl.) and front glass (Fr, gl.), and to the
galvanized iron back (Gb). Fr., frame.

From a sanitary point of view too much emphasis can
hardly be placed upon the construction of the bottom.
By using 4 inch wire mesh through which waste material
will readily fall, I have succeeded in providing a case
which in a large measure is self cleaning. The cases thus
constructed are placed on a long trough-like table, cov-

No. 550] BREEDING MICE 595

ered with galvanized iron, which drains into a sink, By
flushing off the top at intervals, the waste is easily dis-
posed of.

While the use of glass in case building may imply an
increase in the cost of construction, yet in the above plan
it is believed that this cost has been reduced to a mini-
mum, and that a case has been provided which in addi-
tion to its quality for observation assures unusual con-
ditions of sanitation.

A point of great importance in the construction of
such a case is that the galvanized iron back and the side
and front glasses rest on the mesh-bottom so that waste
material in falling through does not strike the frame-
work of the case. It will be seen from a view of the bot-
tom (Fig. 4) that the back of the case is bent inward
an inch from its base so as to come well out on the mesh
bottom, and that the sides and front are placed well over
the line of the framework of the bottom.

2. Nest Boxes and Nesting Materials.—I have followed
with satisfactory results the largely used plan of hav-
ing a winter and a summer nest box. The winter nest
box is made by cutting a chalk box to three-fourths size.
This is filled two-thirds full of nesting material. The
summer nest is a small sized box similarly furnished.

Various materials have been tested for nesting, to
many of which objections can be made. Cotton although
warm, retains odor and at the same time offers a more
serious objection in that the young often become en-
tangled in it, and are thus permanently injured or even
killed. Excelsior I have found to make a good nest if
lined with some sort of soft material, as, for example,
crude floss. One of the most satisfactory materials
which I have tried for nesting is the shredded paper
used in the packing of china. If this can not be procured
at the china store it may be prepared by cutting up any
kind of soft paper. ;

3. Food Receptacles and Food.—1 have tried various
kinds of food receptacles, most of which have given little
or no satisfaction. The difficulty of keeping the food
reasonably clean in open food cans is so great that some

596 THE AMERICAN NATURALIST [ Vou. XLVI

sort of device for its protection is essential. The plan
that I have found most successful has been te place bird
cups in the nest box so that the mouse can procure the
food through an opening in the front of the box. The
cups are further provided with an apperture in the top
through which, without opening the case, food may be
added. This method of feeding, by preventing the mouse
from running over the food, holds in abeyance the spread
of disease.

The most satisfactory food that I have used for mice is
wheat, added to which is a small amount of stale bread.
These, together with milk—which is given by a modifica-
tion of the method below described for water—constitute
the daily and constant diet. Occasionally sunflower seed
and a few leaves of lettuce may well be given for a change.

4. Water Receptacles—tThe inverted bottle which is
now in general use for supplying water has done much to
eradicate ills resulting from a bad water supply. By this
device a large quantity of water can be provided which is
a great advantage in general cultures. In many kinds of
experiments, however, it is desirable to keep water in
greater purity than is possible even by this method. To
do this I have found a device which is strikingly simple
and at the same time singularly effective in that it keeps
the water in contact only with glass. My plan consists in
closing the end of a specimen-tube (or test-tube) in a
frame so that the opening is just large enough to retain
the water drop when the tube is filled and inverted in the
case.” These inverted tubes are inserted through holes
made in the top of the case and are prevented from fall-
ing through by means of small rubber bands placed
around the closed ends of the tubes. The tube may then
be removed and refilled without opening the case, the
refilling being done from a siphon bottle.

The addition of these devices for food and water to a
mesh-bottom case I have found to aid much in the inten-

*The tube for milk, instead o
in the test-tube, has the en
this being the ease with w
per can be cleaned.

f having one end closed permanently as
d closed with a rubber stopper, the advantage of
hich such an open tube upon removal of the stop-

No. 550] BREEDING MICE 597

sive breeding of mice. But to rear mice successfully
further requires a practical knowledge of their breeding
habits.

III. Breeprne Hasrrs or Mice

Copulation in mice is a well-defined act which usually
follows only upon the persistent efforts of the male. It is
marked by a period of union which lasts for several
seconds (ten to twenty-five) and is followed by an interval
of more or less complete rest. :

Practical questions for the breeder are: 1. When will
copulation oeccur—that is, when is the period of heat? 2
What is the duration of this period? 3. How often does
the period of heat recur? To the last of these questions
my experience can offer no answer; and to the second my
observations add little. In one case, however, after a
double copulation had been observed in the evening copu-
lation again took place on the following morning. The
fact that the beginning of heat is shown in some as early
as five hours after parturition and is delayed in others as
long as thirty-six hours thereafter makes it difficult to
determine with exactness the duration of this period.

As to the first question—When may copulation be
secured?—two periods can be determined with consider-
able accuracy. One of these closely succeeds parturition,
the other follows upon a period of rest.

In the first case, if the female has given birth to young,
copulation will usually take place, if she is put with the
male, within from five to twenty-four hours.* In my own
experience I have found that the greater number of births
take place in the early morning, and that copulation will,
in such a case, occur from seven to eight o’clock in the
evening. The female is not invariably in heat at this
time, however, as has been shown by a considerable
number of cases which I have observed.

The second period in which copulation may be expected
is after a mother has gone through an interval of rest,

* This is seen to correspond roughly to the time of ovulation in mice, as
shown by Long (5), the period given being from 144 to 28} hours after
parturition.

598 THE AMERICAN NATURALIST [ Vou. XLVI

either after having suckled her young or after having lost
them. She will then ordinarily copulate within a few
days after having been put with the male. The following
representative table of ten eee cases emphasizes
this point.

TABLE I
Example | Put with Male | Young Born Interval Elapsing, Days.
1 April 2 April 25 23
2 April 5 April 26 21
3 April 12 May 22
a April 19 May 11 22
5 April 29 May 21 22
6 May 14 June 6 23
7 May 18 June 11 24
8 May 26 June 24 29
9 May 27 | June 18 22
10 June 11 | July 3 22

Out of the above ten cases in which the females had not
suckled young for some time, and then were put with
males, nine cases of copulation evidently resulted early,
since the young were born within an interval only a little
greater than the normal period of gestation following a
rest (that is, 20 days) (6).

The eighth case, although unusual, has been further
accentuated by more recent data which show that the
female may remain with the male for long periods with-
out becoming pregnant. This is not conclusive evidence,
however, that an unsuccessful copulation may not have
taken place within that time. In fact, I have found that
in a surprising number of cases copulation does not result
in fertility. As an example of this may be cited 25 con-
secutive cases which I observed, 10, or 36 per cent., of
which resulted in infertilty.

In interpreting this sterility I was at first inclined to
believe that it was due entirely to the females, but since
then I have found males in many cases unproductive-
Some of these were useless for breeding because they —
were practically unresponsive, rarely if ever copulating 5
others, although among the most active males, proved un-

fertile. The following example may be given to show
such a case.

29

No. 550] BREEDING MICE 599

TABLE Ila

RECORD OF ¢ No. 6

No. of Q Palos ts | s | epa] Eo 14 | 10

12/14 12/23 12/27 12/30 1/6 | 4/3 | 4/19
Fert. Infer. Fert. |Infer. Infer. Infer. Infer.

12/3 | 12/7 | 12/8
A E A |Fert. |Infer.|Fert.

In the above table is given the record of one of the most
active males that I have yet had. The record, however,
shows that, although mated with vigorous females, only
40 per cent. of the copulations resulted in offspring. The
table further shows that after a rest from January to
April infertility is still shown. At this later period the
male had become inactive, so that it was difficult to secure
a copulation.

A relatively high index of fertility is shown in the
record of a male designated as No. 5.

TABLE IIb

RECORD OF ¢ No. 5

No. of Q sjaje eas Pee 8.

Date of copulation. |

| | | | |
| 12/3 | 12/7 | 12/28) 1/3 | 1/17 | 1/21 | 3/8 | 3/13 | 3/13 | 4/6
Result | Fert.| Fert.| Fert. Infer. Infer.| Fert.| Fert. Fert.| Fert.| Fert.

Number 5 when mated with females equally as active as
those with which No. 6 was mated gave 80 per cent. of
fertility. Within this series were also two other copula-
tions by the same mouse, but since these were with female
No. 4, which always proved fertile, they were eliminated
and only those counted which were entirely comparable
with those of No. 6.

It may be said that both males and females are found
which have a low index of fertility. Intensive breeding
requires that these be eliminated and that those be
Selected the copulations of which result in a high per-
centage of fertility.

When fertility does result from a copulation, the en-
suing period is of singular interest to the investigator.
This I have discussed in a former paper® in which I have

* Seriously ill after the birth of her young.
* Loe. cit.

600 THE AMERICAN NATURALIST [ Vou. XLVI

shown that the period of gestation depends upon the state
of the female. If the pregnant female is not suckling
young, parturition with but rare exceptions takes place on
the twentieth day after copulation. If she is suckling on
the other hand, the period varies with the number of
young suckled. Thus, for example, if the mother be
suckling five during gestation she may be expected to go
about twenty-five days; if ten thirty days. :

Parturition, which terminates the period of gestation is
normally of brief duration, even in case a large litter is
born. But this is not invariable, for I have observed
eases in which labor was prolonged, and some in which
unaided birth was impossible.

We are inclined to believe that in mankind the difficulty
of giving birth, which not infrequently results in the death
of the mother, is due to the artificialities of civilized life.
But here we find the same stern fact emphasized in a type
remotely removed from any such influence.

IV. REARING or THE YOUNG

The most hazardous time in the life of a mouse is the
first few days of its existence. Born helpless and naked
it is dependent upon the mother not alone for nourish-
ment but for warmth as well. Some mothers at this time
are most solicitous for their young, building elaborate
nests for them and giving such care to the young as to
tide them over this early period. Others there are that
not only withhold the requisite care, but which at this
time prove the most serious menace. Some of these
bundle their young away in the nesting to die; while
others in bad conditions openly destroy them. This sin-
gular and, so far as I am aware, unexplained phenomenon
of destroying the offspring, is carried to a high pitch in
the case of mice. Under unfavorable conditions of tem-
perature, nesting and the like, I have seen three or four
litters destroyed in succession. Miller (7) in a study of
the brown rat shows that this destructiveness in the rat

1s carried to even a greater extent than in the case of the
mouse,

No. 550] BREEDING MICE 601

If the young escape the perils of the first few days they
usually grow rapidly and, at the end of a few months,
reach maturity. There is a disease, however, which at the

` end of the second week may attack the young, leaving

them emaciated or, when more severe, killing them in
great numbers.

Between birth and maturity four well-defined stages
occur. To know these is often of practical service to the
breeder for the determination of age, sex and the like.

The first stage is that in which the newly-born young
have a peculiarly red and transparent skin through which
is seen the stomach white with milk. Following this at
the end of the sixth or seventh day a second stage is evi-
dent in which the body is covered with flaky scales of
dandruff—the forerunners of a coat of silky fur. A third
important stage, which I have designated as the early
Stage for distinguishing sex’ is usually shown on the
ninth or tenth day, at which time the mamme in the young
females appear. These can be observed for an interval
up to the thirteenth or fourteenth day, at which time the
fur usually obscures them.

Determination of sex after the body is covered with fur,
for example at the time of weaning, is often difficult. Be-
cause of this I have found it advantageous at the end of
the third period to mark the young females by clipping a
tuft of fur at the root of the tail, so that later, when they
are to be mated, no difficulty is found in distinguishing
with certainty males from females.

The fourth period, on the fourteenth day, is character-
ized by the advent of sight. This like all other periods
Shows slight variation. While in a few cases I have found
the eyes to open as early as the thirteenth day, in others,
equally normal in other respects, I have found them to be
delayed until the fifteenth and even the sixteenth day.
The regularity with which this period occurs, however,
1S a sufficiently exact criterion to make it an index of age.

From the fourteenth day to the twenty-first, the date

“This does not mean that sex can not be determined earlier than this
Period. As a matter of fact, sex can be determined at birth, this, however,
18 difficult and less certain than to determine it at the third period.

602 THE AMERICAN NATURALIST [ Vou. XLVI

at which the young should be weaned, no definite change,
except increase in size, is shown.

With the period of sexual maturity we may count the
cycle of development complete. This does not imply, how-
ever, that growth ceases at this time. The time at which
mice reach sexual maturity varies greatly. While I have
had some mice to pair at six weeks, this is rather unusual.
It has been my experience that under ordinary cireum-
stances both the males and females reach sexual maturity
in the second or third month. From this time on for the
next few months the mice are in the prime of the repro-
ductive period, beyond which, at the end of ten to twelve
months of age, activity diminishes and, for purposes of
breeding, the mouse is of little further service.

V. PRACTICAL SUMMARY

The ease with which white mice can be handled makes:
them in many cases preferable for experimentation to
other and larger rodents, but their usefulness has been
greatly curtailed because of a wide-spread notion that:
they are difficult to rear under the conditions of the
laboratory.

The purpose of this paper is to give an intensive
method by which I have been able to rear then in abun-
dance. By ‘‘intensive’’ I mean that relatively few mice
are kept from which to breed, and that these are kept
under conditions which insure productivity.

Especially detrimental to intensive breeding are
parasitism and marked fluctuations in temperature. Hot.
air heat, the temperature ranging from 20° to 25° C., has
proved most satisfactory. Heat from an oil stove whem
continued for a considerable length of time proved un-
satisfactory.

Mice which are badly parasitized are useless for breed-
ing purposes. Various kinds of sprays and powders are
used to rid them of the parasites. Some of these, how-
ever, I have used with disastrous results. Parasitism
may be best prevented by using a case that is readily
cleaned. My cases are washed weekly with hot water to

ae

No. 550] BREEDING MICE 603

which is added ‘‘gold-dust’’ and a small amount of petro-
leum; the nest-boxes, into which the mice are put at the
time of cleaning, are partly closed and then removed from
the case.

Essential to intensive breeding is an adequately
equipped case. A most essential requisite is a wire mesh
bottom through which waste material readily falls. If
such a case be placed on a table with a trough-top lined
with galvanized iron, waste matter can easily be drained
into a sink. For nests the case is provided with three-
fourth sized chalk boxes filled two-thirds full of shredded
paper. Food (wheat) is procured by the mice through
an opening in the front of the nest box from bird cups
placed inside of the box. These are filled from the out-
side without opening the case. Water may be kept in
excellent condition in test-tubes which are closed suffi-
ciently to retain the water drop when the tube is inverted
in the case; these are filled from a siphon bottle.

A knowledge of breeding habits is of great importance
in intensive breeeding. Copulation in mice is a well-
defined act, which lasts from ten to twenty-five seconds
and is followed by a more or less complete rest. It nor-
mally takes place on the day that a litter is born. In
females that have gone through a period of rest it will
usually occur a few days after the female has been put
with the male. Copulation may or may not result in fer-
tility. By a selection of males and females with a high
index of fertility the number of offspring may be greatly
Increased.

The period which a non-suckling mother carries her
young is a few hours short of twenty days. A mother
suckling young, on the other hand, carries her litter
twenty plus the number that she is suckling. Thus a
mother suckling five will go practically (20 +5) twenty-
five days, while one suckling ten may be expected to run
(20+10) thirty days. |

In rearing the young it is well to remember that the
Sreatest mortality results within the first two or three
days of life. At this time the young must be kept warm.
Under bad conditions the mother may destroy them.

7”

604. THE AMERICAN NATURALIST [Vou. XLVI

From birth to maturity mice pass through several well-
defined stages which to the breeder are of importance for
the determination of age, sex and the like. These are 1
an early stage in which the skin is peculiarly red so that
through it may be seen the stomach white with milk; 2
a second stage at six to seven days in which the body is
covered with flakes of dandruff; 3 a stage at nine or ten
days in which the mamme appear in the females. This I
have designated as the stage for the early determination
of sex; 4 on the fourteenth day a stage at which the
eyes open. :

At twenty-one days the young should be weaned. From
this time on slight change 1 is shown except increase in size
until sexual maturity is reached. This usually occurs in
the second or third month. From this time up to the end
of ten months or a year of age the mouse is in the height
of the breeding period; beyond this time, for purposes of
breeding, the mouse is usually of little further service.

BIBLIOGRAPHY

1. Cuénot, L. 1902. La loi de Mendel et L’hérédité de la pigmentation
chez les souris. Arch. de Zool. Exp. et Gén., 3°-série, Tome X (Notes
et Revue, p. xxviii).

2. Ehrlich, P. 1891. Experimentelle Untersuchungen über Immunität.
Deutsche med. Wochenschrift, p. 976. i

3. Bashford, E. F. 1909. Cancer in Man and Animals. Lancet, clxxxvii,
p. 691

4. Yerkes, oberi M. 1907. The dancing mouse; a study in animal be-
havior. The Macmillan Company. .
5. Long, J. A. and Mark, E. L. 1911. The pein of the Egg of the

Mouse. Pub. on Inst. (Washington, D. C.), No. 1
6. Daniel, J. Frank. 0. Observations on ee Pelt of Gestation in
White Mice. ey ‘be Zool., 5.

7. Miller, Newton. 1911. Repeodiuctiaal: in the Brown Rat. Am. Nat.
XLV, p. 623,

THE DISTRIBUTION OF HYLA ARENICOLOR
COPE, WITH NOTES ON ITS HABITS
AND VARIATION

C. H. RICHARDSON, JR.

STANFORD UNIVERSITY `

_ Srupents of zoogeographical distribution are fre-

quently hindered by the scarcity and inexactness of the
published data in the particular group which they are
studying. Especially is this true of students of western
North American amphibians, for they must rely largely
upon the publications of the early exploring expeditions
in which localities were often stated in a most general
way and at times with doubtful accuracy.

Our present knowledge of the distribution of the tree
toad, Hyla arenicolor Cope, is very incomplete. Many
of the references to its occurrence are extremely indefi-
nite and unreliable and in no case has enough material
been gathered to give the limits of its range in any one
region. It was first discovered and named Hyla affinis
_ by Baird! in 1854, the description being based upon one
specimen from the state of Sonora, Mexico. Later Cope*
found this name to be preoccupied and replaced it with
arenicolor. At the present writing this tree toad is known
to inhabit parts of southern California, Utah, Arizona,
New Mexico, Texas, and Mexico. , Southern California is
included in its range on the strength of two specimens
collected in 1875 by H. W. Henshaw,’ and no additional
records of its occurrence within the state have been made
by the herpetologists who have explored this region.
Within the last few years, however, the University of
California Museum of Vertebrate Zoology has acquired
a number of specimens of Hyla arenicolor from various

*Proe, Acad. Nat. Sci. Phila., p. 61.

*Jowrnal Acad. Nat, Sci. Phila., 1866, p. 84.
* Yarrow, Bull. U. 3. Nat. Mus., No. 24, 1882, pp. 24, 175.
605

606 ` THE AMERICAN NATURALIST [Vor XLVI

localities in southern California. Through the kindness
of the director, Professor J. Grinnell, the writer has been
extended the privilege of examining these specimens and
the results are incorporated in the present article.
Specimens from the following localities have been
studied, all of which are in the Museum of Vertebrate
Zoology unless otherwise stated.

TABLE A
No. of
Speci- Locality Altitude Date ‘Collector
mens
8 |Mountain Spring, San Diego Co., Cal. About 1909
4,500 ft.| Mar. 25 |F. Stephens
1 |La Puerta, San Diego Co., Cal. About
4,500 ft.| June 5 |F. Stephens
1 |Warner’s Pass, San iy Co., Cal. 4,000 ft.| June 22 |F. Stephens
1 |Julian, San Diego Co., Cal 3,750 ft.| July 29 |F. Stephens
1906
3 |Pine Mt., near Escondido, Cal. 2,750 ft.| Sept. 4 |J. Dixon
1908
1 |Carrizo Creek, Santa Rosa Mts., Cal. 3,000 ft.| June 22 |J. Grinnell
2 |Dos Palmos Springs, Santa Rosa Mts.,
3,500 ft.| May 26 J. Grinnell
9 |Deep Canyon, Santa Rosa Mts., Cal 3,000 ft.| June 21 J. Grinnell
1 (Lower Palm Canyon, San Jacas Mts.,
Cal. 800 ft.| June 15 |J. Grinnell
2 |Oak Springs, upper Palm Canyon, San
Jacinto Mts., Cal. 4,750 ft.| June 11 |J. Grinnell
4 |Base of San Jacinto Mts. near Cabazon, May 5 |W. P. Taylor
Riverside Co. a Cal 1,700 ft. and 7 and C. H.
Richardson
r.
4 |Sierra Madre, Los Angeles Co., Cal. 1,500 ft.| May, 1904/J. Grinnell
1903
20 Arroyo Seco Canyon near Pasadena, Cal.| 1,500 to Aug. 3
i 2,000 ft. and 23 |J. Grinnell
1 |Tejunga Valley, Los Angeles Co., Cal. About 1910
1,500 ft.| Apr.1 |J. Grinnell
1 renee: University coll.) Upper Santa C. H: Rich-
Anita Canyon, Los Angeles Co., Cal. | 3,500 ft.| Aug. 7 ardson, Jr.

In California, Hyla arenicolor is now known to range
northward along the coast mountains from near the Mex-
ican boundary to the Tejunga Valley, Los Angeles
County. In San Diego County it occurs on both the coast
and desert slopes of the mountains; in Riverside County,
on the desert slope, and in Los Angeles County, so far
as known, only on the coast slope. No records of its

eke Nt
E ae
=: 4
TEs

No. 550] HYLA ARENICOLOR COPE < BOT

occurrence in San Bernardino County are at hand,
but a careful search will undoubtedly reveal its presence
there. Throughout this region, its habitat appears to be
confined to streams and mountain springs between 1,000
and 5,000 feet elevation. Here the writer has found it
associated with such trees as Alnus rhombifolia, Platanus
racemosa, and Acer macrophyllum and in this state at
least it may be regarded as an inhabitant of canyons
within the upper sonoran zone. It is apparently more
strictly aquatic than the smaller Hyla regilla Baird and
Girard, whose range in southern California is, in part,
coextensive with it. The former species has never been
found far away from the vicinity of water, while the
latter has often been seen under vegetation a considerable
distance from it.

The following meager notes indicate that the breeding
season of Hyla arenicolor extends from late spring until
fall: Two females from Sierra Madre, Los Angeles
County, May, 1904, one from Warner’s Pass, San Diego
County, June 22, 1909, and one from Pine Mountain, near
Escondido, San Diego County, Sept. 4, 1906, contain large
eggs. There is also a young specimen from La Puerta,
San Diego County, collected on June 5, 1909, in which the
tail has not been entirely absorbed.

Miss Dickerson‘ has described the rapid color changes
that take place in this species. The writer noted that a
light-colored individual which was captured on a granite
rock changed to a dark gray mottled with lighter mark-
ings when placed for a short time in a covered tin pail.

The widely scattered record stations given below, some
of which are too inexact to be of great value, suggest that
Hyla arenicolor lives in suitable places over practically
the entire southwestern part of the United States and a
considerable portion of the Mexican tableland as well.
Little has been written concerning its habitat preferences
within this region. Stejneger® has found it in the Grand

* The Frog Book, Doubleday, Page and Company, 1906, p. 122.

"N. A. Fauna, No. 3, 1890, p. 117.

608 THE AMERICAN NATURALIST [Vou. XLVI
Canyon of the Colorado, Arizona, at an elevation of
1,000 feet above the river and in the bottom of the
cai von. Ruthven’s* specimens from Sabino Canyon,
Santa Catalina Mountains, Arizona, were ‘‘found among
bushes on the floor of the canyon’’ in the ‘‘ willow-poplar

association.’’
TABLE B
No. of
Speci- Leoality Date Authority
mens
1
MPEG POON, PACKING jer ccs’ S wives oe eos Cope; Baird
Guana juato, Mexico eae are esas Gee PESIN 4 E
RIWROUIRIOTG, E S vo or es ee, S Duges
grt Sg City of Masta: and Chihuahua,
AE E T Win Bad ec ely Lara ir OH an P a Oe eee
Valley eal a OF TOMA MeRIOO: 2 Fle. Secs. Cop
1 |Del Rio, hin idee a es Obie pee a ea A Witmer aay
a O P E AA e T Se eee oe ee ee he Boulen
1 Santa Fe, sow MOR CG eee ees June, 1873 pt
a (ort Winmeate, New: Mexico... J. - 32.0. oo os oo ee, ope
BPO ein cis ea Owls ies A e 1872 arrow
IT Me aO a eaa ae aa a aa a e Miss Dickerson
Lepper Colorado Hive 2 fico S a o Cope
White River Can S Soe A E AN PEE ae ope
6 [Grand Snara of paa dolorido. Arizona | Sept. 3, 1889| Stejneger
Fort Whipe Arons osana ee Cope
2 abino iosal Santa Catalina Mts.,
E E E E TEE E EAT May 23,1903) Ruthven
2 Southern CRABS oe ae 1875 Yarrow

Specimens of Hyla arenicolor from southern Cali-
fornia agree quite closely with the published descrip-
tions. Cope’ states that the diameter of the tympanum
is equal to two thirds that of the eye fissure, but in ten
specimens measured (see following table) this ratio is
shown to be 47.7 per cent., or not quite one half. Boulen-
ger’s Hyla copiis described from two specimens obtained
at El Paso, Texas, and now considered synonymous with
Hyla arenicolor, has the diameter of the tympanum one
half that of the eye fissure, a fact which suggests that this
character is subject to considerable variation throughout
the range of the species. The ratio between the length
of the body and that of the hind limb is also subject to

* Bull. Am. Mus. Nat. Hist., 1907, p. 509.

"Bull. U. 8. Nat. Mus. , 1889 , No. 34, pp. 369-370.
* Annals and Magasiné Nat. Hist., 1887, p. 53.

609

HYLA ARENICOLOR COPE

No. 550]

‘TSO “OD OBI UBK ‘ssBq S IOUIGM| CZF" LY K SS ‘OL “3G ST T'o é 1801
TSO
“OD OdeIq ULG ‘oprpuoosy reed le ha ould; PPF o'p 3 G'IZ "29 G'g 8'SI "Le é 666
‘TRO “OD sepsuy SO ‘JIPEN Big og a 4 KA “IZ "g9 “8S “ST "88 é 16
[8D “OD sejesuy SOT ‘IPEN BLII 09° 9 'g SS ‘L9 "93 "OT “OF é 626
‘TEO “OD O33 uBg ‘SulIdg urezunoNW| = FHP" oy KA Si ‘ag G'IZ 9ST 88s 2 816
"T8D “OD OBA ULg ‘Burıdg urezuno w OF’ '} o'I “LI ‘og T'SI Ti ‘OS ad 946
‘TEO “OD o3arq uvg ‘Suridg urwgzunoWw| 607 FP S'I cot £09 ost ET £8 2 086
T8O “OD o3ərq ULg ‘Fundg ureguno W gg "7 Tg "SI "E9 £61 “ST OPE 2 646
‘TEO “OO o3ərq ULg ‘Sundg uleyunoyy; GLE a4 G'I ST S'ES 0G "OI "Se 2 116
"T8D “OD sejesuy so ‘srpen vog) 939° '} g'g ‘0G GLa SS SSL o'oe Ye 026
ə : of
jo "wet a[4ys0rQ Jo jo q3uy | snuy
£ d 94y Jo mnueduis], d quyrt nou . £ A
TT Saiz jo | Wea jooma] “OE | ee Ok | wom | a) S ON
“WET Oey YIPIM

(SUMLAWITUW, NI) SINAWTANSVAN JO TIAVL

‘610 THE AMERICAN.NATURALIST — [Vou. XLVI

pronounced variation in southern California specimens.
The heel of the limb when extended forward usually
reaches to the anterior border of the orbit,® but in young
individuals it often extends to the tip of the snout.

he =

THE DISTRIBUTION OF Hyla arenicolor COPE IN CALIFORNIA ACCORDING TO THE
LOCALITY RECORDS.

j Ei || | | l

The table also shows that there is a well-marked dif-
ference in the size of the sexes, the female being larger,
especially in the length of the body and hind limbs.

BIBLIOGRAPHY
Baird, 8. F.,
1854. DER of New Genera and Species of North American
Proc. Acad. Nat. Sci. Phila., p. 61.
1859. United States and Mexican Boundary Survey: Batrachia, p. 29,
4-7.

pl.
-Boulenger, G. A.
Catalogue Bratrach. Salient., p. 373.
1887. Descriptions of New Reptiles and Batrachians in the British
useum. Annals and Magazine Nat. Hist., p. 53.
1888. Description of New Brazilian Batrachians, Hyla copii. Annals
and Magazine Nat. Hist., p. 189.
* Cope, loc. cit.

No. 550] HYLA ARENICOLOR COPE 611

Cope, E. D,
1866. Journal Acad. Nat. Sci. Phila., p. 84.
1866. On the Reptilia and Batrachia of the apera a of the
Nearctic Region. Proc. Acad. Nat. Sci. Phila., p.
1875. Check-list of North American Patraihii and a gee Uz.
. Nat. Mus., No. 1, pp. 31, 90.
1880. On the dadana Position of Texas. Bull. U. S. Nat. Mus.,

No. 17, p. 47.
1885. The Riverside Natural History, Houghton, Mifflin and Company,
III, p. 338.

1887. Batrachia and ea anes of Central America and Mexico. Bull.
. S. Nat. Mus, No. 32,

1888. AMERICAN CEH p. '80.

1889. The Batrachia of S America. Bull. U. S. Nat. Mus., No.

34, p. 369, figs. 1-7.
1896. The Geographical Dber of Batrachia and Reptilia in
North America. AMERICAN NATURALIST, 30, pp. 1014, 1021,

1022.
‘Coues, E.
1875. U. S. Geological Survey West of the 100th Meridian, Zoology
5, p. 630.

Dickerson, Miss M.
106. The Pog Book. ‘Doubleday, Page and Company, p. 122.
` Plates XLVII and XLVIII.
Ruthven, A. G.
1907. A Collection of Reptiles and Amphibians from Southern New
Mexico and Arizona. Bull. Am. Mus. Nat. Hist., p.
Stejneger, L;
1890. Annotated List of Reptiles and Sni, Collected by C.
Hart Merriam and Vernon Bailey on San Francisco
Plateau and Desert of the Little sae Arizona, with
Descriptions of New Species. N. A. Fauna, No. 3, p. 117.
Stone, W.
1903. A Collection of Reptiles and nga ond from Arkansas,
Indian Territory, and Western Texas. Proc. Acad. Nat. Sci.
Phila., p. 539.
Yarrow, HC

1875. U. S. Geological Survey West of the 100th Meridian, Zoology

> P-
1882. Check- list of North American Reptiles and Batrachians. Bull.
. S. Nat. Mus., No. 24, pp. 24, 175.

THE UNEXPECTED OCCURRENCE OF ALEU-
RONE COLORS IN F, OF A CROSS BE-
TWEEN NON-COLORED VARIETIES

F MAIZE

PROFESSOR R. A. EMERSON

UNIVERSITY OF NEBRASKA

Brrore the Mendelian methods of analysis became
available, considerable wonder would doubtless have been
excited by the ‘‘mysterious’’ appearance in F, of one
colored grain—purple or red—to every five or six white
ones in case of a maize cross, both parents and F, of
which had only white grains. An occurrence of this
sort has recently been noted in one of my maize cultures
and the F, numbers are explained here as a trihybrid or
tetrahybrid ratio. The crosses in question were made
primarily for a study of size inheritance and fairly large
numbers have been grown. The varieties concerned are
two dwarfs of distinctly different types, Tom Thumb pop
and California Rice pop, and a tall type Missouri dent.
The facts with reference to aleurone color are these: Tom
Thumb pop, a ‘‘white’’ corn (i. e., having non-colored
aleurone), was crossed with Missouri dent, also a white
corn. Three generations of hybrid plants—four gener-
ations for aleurone and other endosperm characters—
have been grown without the appearance of any but
white grains. The same white-seeded Missouri dent was
also crossed with the white-seeded California pop. The
three hybrid generations grown to date have shown no
aleurone color. Furthermore, when the same white Tom
Thumb pop was crossed with the same white California
pop, only white grains appeared in F,. But both of the
two ears containing F, seeds—the only ones that have
been produced as yet—had a sprinkling of both purple
and red grains, too many to be explained as due to care-

612

No. 550] ALEURONE COLORS 613

less guarding against foreign pollen and too few to be ac-
counted for by any simple monohybrid or dihybrid for-
mula. The actual numbers of grains of the various sorts
were as follows:

a anes 43 purple, 10 red, 308 white.

ee SS rer 32 purple, 11 red, 222 white.
BNE soe esas 75 purple, 21 red, 530 white.

The fact is familiar that in crosses of purple with
white varieties of corn, there often appear in addition to
the monohybrid ratio of three purple grains to one white
one, purple, red and white grains in the dihybrid ratios
of 9:3:4 (Hast and Hayes! and Emerson?). It is also
well known that in similar crosses purple and white
grains may appear in F, in the reversed monohybrid
ratio of 1:3 or the dihybrid ratio of 9:7 (Hast and
Hayes’). East? has recently shown that for the produc-
tion of purple aleurone there must be present three Men-
delian factors, C, R, and P, and has demonstrated for
purple, red, and white grains the trihybrid ratio of 27:9:
28. C is a general color factor, that must be present
ordinarily in order that any color may develop, R a fac-
tor that has to do with the production of red aleurone
when C is present, and P a factor for purple that is ef-
fective only in the presence of both C and R. Thus
CRP gives purple and CRp red, while all the other
possible combinations give white. All this is on the as-
sumption that a fourth factor J, an inhibitor of color de-
velopment, is absent. Purple color of the aleurone may,
therefore, be said to depend upon the presence of three
factors and the absence of one, CRPi, red color upon
the presence of two factors and the absence of the two
others, CRpi, and whites upon the absence of either one
of the two factors C or R or upon the presence of a third
factor, I, cRP, CrP, or CRPI, ete.

*E. M. East and H. K. Hayes, Conn. Agr. Expt. Sta., Bul. 167 . 57-
100, 1911. yes, . Agr. Expt, l , » PP

— A. Emerson, Amer. Breeders’ Assoc., vol. 6, pp. 233-237, 1911.
E. M. East, AMER. Nar., vol, 46, pp. 363-365, 1912.

614 THE AMERICAN NATURALIST [Vou. XLVI

If the numbers obtained in F, of the cross of Tom
Thumb pop with California pop are to be regarded as
constituting a tetrahybrid ratio, all four aleurone factors.
must be heterozygous in F, the formula being CcRrPpli.
The F, generation would then be constituted as follows:

ye ae ee ee ae eS 27 purple
earra a n a E i oe a cs 9 red

27 CRpl Pe ee gen ee Bee at ae 220 white

scorpi |

If either C or R is homozygous in F,, the resulting F,
ratio should approximate 9 purple: 3 red: 52 white.

The actual numbers fell between these two theoretical.
ratios, as is seen from the following comparison:

Purple Red White Total
Tetrahybrid ratio ........ 22 538 626
Observed numbers ........ 75 21 530 626
Inod Patio 2. 88 29 509 626

From the ratio alone it is plainly impossible to say
whether the cross in question is a tetrahybrid or a tri-
hybrid. Of course behavior of the reds and purples in F,
will settle the matter. If, for instance, either C or R is.
homozygous, one third of the F, red grains should breed
true and two thirds produce reds and whites in the ratio
of 3:1, while if both are heterozygous, only one ninth of

No. 550] ALEURONE COLORS 615-

them should breed true, four ninths produce a 3:1 ratio,.
and four ninths produce a 9:7 ratio. Similarly, if either
one of these two factors is homozygous, of the F, purples.
one ninth should breed true, two ninths give purple and
red 3:1, two ninths purple and white 3:1, and four ninths.
purple, red, and white 9:3:4. But if both factors are
heterozygous, out of the twenty-seven F, purples only one-
should breed true; two yield purple and red 3:1; four,
purple and white 3:1; four, purple and white 9:7; eight,
purple, red and white 9:3:4; and eight, purple, red and’
white 27:9: 28.

The results of intercrossing Tom Thumb pop, Missouri
dent and California pop, so far as they are known at
present, might be obtained if the three varieties had
either of the following sets of formule, or any of the-
modifications of them suggested below:

Tom Thumb pop ICRP ICrP
Missouri dent IcohP or twrf
California pop icrp icRp
Among the allowable modifications of the above for--
mulæ are these: The formulæ for Tom Thumb pop and
California pop may be interchanged. Substitutions of
C for R and R for C may be made if carried throughout
the set. P may be present in any one or two varieties
and absent from any one or two. Where J is present im
Missouri dent and also in one of the other varieties, R
may be present in all three varieties, absent in any one
variety, or absent in Missouri dent and either one of the-
other varieties.

A NEW SUBSPECIES OF ZEA MAYS L.
DR. WALTER B. GERNERT
UNIVERSITY OF ILLINOIS

Waite harvesting a plot of yellow dent corn, a strain of the
Leaming variety grown on the Illinois Agricultural Experiment
Station fields in 1909, one of the workmen found a peculiarly
shaped ear which was laid aside in the drying-room as a cu-
riosity. The corn in which this ear was found came from a
strain that had been subjected for several generations to an
ear-row selection for high protein content by a mechanical in-
spection of the endosperm.:

This new type of ear which reproduces faithfully in its prog-
eny is cone-shaped in outline and gives the appearance ex-
ternally of being composed of a mass of kernels borne on num-
erous irregular branches (see ‘‘a’’ in the figure). A longitu-
dinal section (at ‘‘b’’) displayed kernels throughout the ear.

The ‘‘branched’’ form is a prolification of the fleshy type of
4 to 30 or more-rowed cob common to all varieties that to the
writer’s knowledge have béen described to date. For this new
type the writer proposes the name Zea ramosa, from the Latin
“‘ramosus—having many branches.’’ This name is proposed
in conformity with the bi-nomial classification of Sturtevant?
which is now generally recognized. We will not here discuss
the precedence nor the desirability of Sturtevant’s nomencla-
ture for the subspecies of corn which were all grouped at first
by Linneus under the general head Zea mays.

The new type Z. ramosa (branched) is as much deserving of
a specific name as are any of the six groups recognized by Stur-
tevant, namely: tunicata (pod), everta (pop), indurata (flint),
indentata (dent), amylacea (soft), saccharata (sweet). The
first of these six groups has a more or less monstrous develop-
ment of glumes into pods which inclose each kernel on the ear
with leafy bracts known as the husks. The classification of the
other five groups is based on differences in characters situated
in the endosperm of the kernel.

* Representative samples of the ears thus obtained for planting in the

next year were also analyzed chemically to determine the efficiency of the
method of mechanical selection. `

* Sturtevant, E. L., Bul. Torr. Bot. Club, 21: 319-343, 1894; Off. Exp.
Sta, U. S. D. A., Bul. 57: 7-108, 1899,

616

No. 550] ZEA MAYS 617

The ear of Z. ramosa, which is always of a definite form, is
borne at the usual place near the middle of the culm and is not
to be confused with sparsely branched ears sometimes found on
the culm nor with ears frequently found in the tassels on ordi-

). a, external
. 619). c, tassel ( . 620).
linois Agricultural Experi-

: iD TASSEL IN CoRN (Zea ramosa
view of parent ear. b, longitudinal section of same (p
Photographs by Flora Sims and by courtesy of the Il

ment Station.

THE NeW TYPE or EAR AN

nary corn plants. Such abnormalities which are fluctuating
in their inheritance have thick pasal branches of fleshy cob—
which may be as long or longer than the primary cob itselfi—
and may bear from two to a dozen or more rows of kernels on
each branch. Furthermore, no male florets have as yet been
found in any of the ears of Z. ramosa and they are always
covered with normal husks.

618 THE AMERICAN NATURALIST [ Vou. XLVI

A feature of especial interest in the new type is the fact that
the tassels of such plants are also invariably much branched and

cone-shaped. (A reduced photograph of the tassel is shown at
**e.’’) No instance has yet appeared in which this correlation
did not exist.

No. 550] ZEA MAYS 619

During the last three years the writer has had under obser-
vation a large number of varieties and their hybrids. He has
been able to isolate more than a dozen tassel types which are
strikingly different in shape and which are distinct from each
other in inheritance. Some
of the characters of these
types are plainly correlated
with certain characters of
the ear. Advantage can un-
doubtedly be taken of this
fact in analyzing the be-
havior of such fluctuating
phenomena as size, shape,
and number of parts.

This correlation between
tassel and ear permits the
selection of individuals in
the field before the silks

pollenated at will. In an
investigation not yet pub-
lished, the writer has found
that the tassels in a large
number of varieties. are
usually out and fully ex-
panded one or more days
before ` any pollen is shed
from the anthers, while the
tassels produce pollen at an
average of from one to three T

days before the silks appear on the same plants. ,

During the first season (1910) in which the branched ear was
-tested to see if it would reproduce the character only two out
of fifty kernels planted produced individuals bearing the
branched ears. It was at this time that the correlation between
the tassel and ear type was discovered. The fact that only two
Plants of- the Z. ramosa type were obtained this first year indi-
cated that either the character was reproduced only occasionally
or that it was a recessive character and that the parent was
Pollenated largely by neighboring plants bearing normal ears
which must be dominant to the branched form. It was pre-

c

620 THE AMERICAN NATURALIST [ Vou. XLVI

dicted? that the latter explanation was the true one, and the
results from another generation (grown in 1911 from hand
pollenated ears) have substantiated the prediction. Our data
for this year show that the branched form of ear and its ac-
companying tassel type are recessive to the fasciated, cylindri-
eal form of ear from which they originated.

Perhaps no one is ready to draw the limits upon that inde-
finable term ‘‘species,’’ but Mendelian studies have thrown a
bright light upon this mooted question. It is now very evident
that sterility in hybrids is not a safe guide for determining
what shall be a species. Darwin reported a number of cases to
show the fallacy of this theory which was at one time advanced
by Kölreuter, Gartner and others. Mendelian studies have
disclosed a number of cases of sterility (I have found several
in corn) which are not due to hybridization nor to species’ dif-
ferences.

Systematic classification should be founded upon either the
genotype or upon the Mendelian basis. The genotype basis
would be feasible for self-bred and apogamous, including par-
thenogenetic, types of reproduction; while the Mendelian basis
would undoubtedly be the most satisfactory for types of plants
and animals that are continually mix-fecundated. We are
learning that there are an almost inestimable number of char-
acters in corn and that they may be quickly distributed to all
the representatives of the six species-groups by hybridization.
As an example: the kernel colors; red, yellow, blue and their
absence (white) are found in all of the groups. If we were to
give each distinct character, wherever we find it, a specific clas-
sification we should have many more species than we now recog-
nize. This is especially true with regard to economic plants.

Such classification is desirable, however, and will soon be
needed from a Mendelian standpoint if from no other. As an
instance: we have evidence that there are more than twenty
reds or phases of red color in corn alone, and a system for their
classification is desirable. As was mentioned above, we have
isolated a dozen distinct tassel types, each possessing a number
of characters that may be easily redistributed by hybridization.
The inheritance of detail in both plants and animals is various:
l ***The Analysis of Characters in Corn and Their Behavior on Transmis-
sion,” a paper submitted May 13, 1911, to the graduate school of the
University of Illinois as a detects thais.

No. 550] ZEA MAYS 621

when true dwarfs and true giants are hybridized, size segre-
gates distinctly in their progeny; but when fluctuating shorts
and talls are hybridized, size exhibits a so-called ‘‘ blending’
behavior that is generally complex. Thus it is evident that for
recording such characters and their method of segregation we
already need for the sake of conformity, brevity and ease in
reference a definite, simple, systematic classification of charac-
ters irrespective of species, varieties or individuals. Bateson,
by grouping characters of similarity under one head; Tscher-
mak and others, by distinguishing ‘‘types’’ of segregation, have
already taken a step toward this end.

The newly discovered type of corn is so radically different
from all others yet reported, and since we are at present recog-
nizing six species-groups of Zea, it seems very appropriate to
add Z. ramosa as a seventh. And yet the writer will not be dis-
appointed if the proposed addition is not recognized.

That Z. tunicata and Z. ramosa both originated as mutations
we have no doubt; but as to the causes which led to the pro-
duction of these two peculiar types, we have no definite knowl-
edge. It has been proposed that new forms, aside from those
developed by hybridization, are due to accidents in mytotie di-
vision; and yet those same writers are perhaps not ready to
admit that even the greater proportion of the myriads of diverse
forms of plant and animal life that exist on the earth to-day are
accidents! This, of course, has nothing to do with the fact of
chance meeting of gametes in reproduction.

The writer has evidence (not yet published) upon various
Strains of pod varieties and their hybrids with other podless
varieties to show that the pod character, in that form, never was
the normal or original pod or glume in Z ea; and it is also evi-
dent that the new branched ear, as it is, is not a reversion to
a former one. As may be seen at “‘b’’ in the illustration, the
pithy core of the cob is not affected by the branching in the
outer zone. The branches are somewhat fleshy and contorted
as well as being very numerous. As stated above, no male
florets have yet been found in the ears of the branched corn.
Such evidence points to the conclusion that this is not a case
of at least total reversion.

As is generally the case in such instances, it is only a matter
of conjecture as to the causes that led to the production of this
individual which, in so far as is known, was different from all

622 THE AMERICAN NATURALIST [Vou. XLVI

others in the history of the strain. Mr. W. T. Craig, who has
been connected with the corn-breeding work at the University
of Illinois for a number of years, states that to his knowledge
no ear similar to this has ever been harvested on any of the
breeding plots at this station.

The selection in the particular strain in which the branched
ear was found has since been discontinued and thus we do not
know whether the type would ever have occurred again in the
same strain. Hybrid progeny from this parent strain are, how-
ever, yet being grown at this station; but no other individuals
like the one here described have been found.

Several more generations of the branched corn should be
grown before we can make any reliable statements as to its
economic value. It is hoped that the new type may be devel-
oped by hybridization and subsequent selection among the seg-
regates (which work is in progress now). As yet it does not
bear as much grain as the unbranched ear in the strain in which
it was found. The parent ear of Z. ramosa measured approx-
imately 5.5” in length and 9” in circumference. Very little dif-
ference was found in the size of the other parts of the plants
except that of the tassel, which is also slightly smaller on the
new type.

The branched ear is apparently an ideal form to feed whole
to livestock. The cob is of such nature that it may be readily
masticated with the kernels and without the necessity of grind-
ing or chopping before it is fed. It may also prove to be an
ideal type for ensilage. Whether it will yield well enough to
justify its production for any of these, or other purposes, re-
mains to be investigated. 3

NOTES AND LITERATURE

PATTEN ON THE ORIGIN OF VERTEBRATES, AND
THE GENERAL QUESTION OF THE VALUE
OF SPECULATIONS ON THE PHYLOG-—
ENY OF ORGANIC BEINGS?!

From the standpoint of range of topics covered, amount of
work performed, and time devoted to its execution, this work by
Patten may without exaggeration be spoken of as monumental.
Many of the facts set forth are original observations by the au-
thor and his students, and of those not original a large propor-
tion have seemingly been personally studied by him. Further-
more, nearly all the large number of figures are either original
or bear the stamp, by way of modification of borrowed figures,
of Patten’s well-known skill as an illustrator.

A list is appended comprising 26 titles of papers and ad-
dresses by the author or the author in collaboration with his
students; but unfortunately references to the works of other
investigators drawn upon are rather few, often somewhat in-
definite, and not well set out in the text. In a book so abound-
ing as is this one in argumentation, many of the main conten-
tions of which are open to debate, sources due to authority ought
to be given exactly and without stint.

Frequent as is the occasion in scientific books to estimate
worth from the two viewpoints of facts presented and theories
defended, rarely is the importance of keeping the two distinct
So great as in this case. Many of the chapters, notably V to
XII and XVI to XX, are veritable magazines of recorded ob-
servation to which workers in the field will, it would seem, find
it profitable to turn for years to come. This remark applies par-
ticularly to the sections dealing with the central nervous sys-
tem of Limulus; with the cutaneous, olfactory, and optical or-
gans of ‘‘Cerachnids’’ and Anthropods; with the dermal skeleton
of Limulus; with the endoskeleton of Arachnids; with the nerve .

***The Evolution of the Vertebrates and their Kin,’’? by Wm.
486 pp. and 309 figures, P: Blakiston’s Son & Co., 1912.

623

624 THE AMERICAN NATURALIST [ Vou. XLVI

supply to the heart of Limulus; and with the general structure
of the Ostracoderms.

The experiments on functions of the brain recorded in Chapter
XI should be a valuable contribution to the interesting prob-
lem of metamerism as expressed through the activities of the
central nervous mechanism.

Two defects in the descriptive matter are likely to interfere
with as extensive a utilization of the book by biologists as it
merits. The first to be mentioned is a want of directness and
definiteness in many of the descriptions that renders their com-
prehension extremely difficult, in some cases almost impossible.
This is due partly to the way references are made to the illus-
trations. Not infrequently a text description of a structure is
given, not very fully, in the course of which one or several
figures are referred to but without specifying the letterings for
the particular parts described. The reader, being in doubt, may
turn to the ‘‘Explanation of lettering’’ at the end of the vol-
ume, only to find that the illustrations in question either have no
letterings for the particular parts, or if sufficient patience in
digging is exercised, to find that the part is labeled with a dif-
ferent name from that used in the description. The account
of the ‘‘middle cord, the lemmatochord and the notochord’’
(Chap. XVIII) is an example of the difficulty here indicated.
Although I have spent much time on this chapter, I have not
been able to get a clear understanding of what is dealt with.
How many distinct structures are in hand? ‘‘The beginning of
the notochord may be recognized in practically all segmented
invertebrates, as the so-called middle cord, or median nerve, and
in its derivative, the lemmatochord,’’ p. 324. This statement is
general, i. e., is not made as applying to any particular animal.
It seems definite to the effect that ‘‘median nerve’’ and ‘‘middle
cord”? are synonymous, and that the structure indicated gives
rise to the lemmatochord. But Fig. 224A, p. 327, representing
the ‘‘nerve cords and lemmatochord of Cecropia,’’ presents to us
the “late pupal stage showing the fully formed lemmatochord,
derived from the condensed sheaths of the median and later
Per ds; also remnants of the median nerve.’’ (Italics by the re-
deto In Limulus the ‘‘middle cord” is said, p. 334, to be
_ arranged in two main lateral cords,” and Fig. 55, l.l.ch., p. 67,
ìs referred to as illustrating this statement. Turning to this fig-
ure and the explanation of letterings, we find that L.l.ch. stands

No. 550] NOTES AND LITERATURE 625.

for ‘‘lateral bands of the lemmatochord.’’ This same figure-
shows a “‘median portion of the lemmatochord”’ distinctly and’
widely separated from the lateral portions. The inference is:
that these lateral and median structures unite somewhere; but
no direct statement to this effect is given, at least in the section
on the middle cord of Limulus.

Again, Fig. 225, p. 328, presents five cross sections of the:
nerve cord of an adult scorpion. Section 5 is said to show the-
“‘merochord.’’ Neither in the description of this figure nor in
the text do we find a direct letter reference to the merochord.
A structure labeled m appears in the figure, but on turning to
the explanation of lettering ‘‘m’’ we find may stand for-
*‘mouth’’ or ‘‘muscle.’? But it is unfair to criticize illustra-
tions and their letterings and labels alone. They must be taken
in connection with the text. By reading a subsection headed
“The Bothroidal Cord or Lemmatochord’’ we find that the-
merochord in section 5 of Fig. 225 is marked l.ch., which stands
for lemmatochord. A sufficiently careful reading of the text
clears up the merely expositive difficulties contained in the figure :
The merochord is the lemmatochord of the ‘‘ posterior thoracic
neuromeres,’’? p. 328. This interpretation is compelled when
Fig. 5 is taken in connection with the text statement indicated.
But then the difficulty becomes substantial and not merely ex-
positive, for on page 328 we read: ‘‘In the scorpion, the median
herve itself is hardly recognizable. .. . The neurilemmas of the:
median and lateral cords form the bothroidal cord of the ab-
domen and the merochord of the posterior thoracic neuromeres.’”
But we have seen above that according to section 5 of Fig. 225
the merochord is a particular part of the lemmatochord. Hence
this part at least of the lemmatochord is formed from the neuro-
lemmas of the median and lateral cords. Under the topic ‘‘De-
velopment of the Lemmatochord’’ we read ‘‘The lemmatochoré
arises, in part, as an axial cord of cells extending forward from.
the primitive streak’’; and nothing under this heading or else-
where so far as I have been able to find, iarmonizes this state-
ment with the indirect assertion above pointed out that the-
lemmatochord is formed from the neurilemmas of the median
and lateral cords. The nearest approach to:such harmonization
is the statement, p. 330, that at the time of hatching, the lem-.
matochord at certain places ‘‘remains permanently attached to:

neurilemma of the middle cord.” As:one-of the many stu--

626 THE AMERICAN NATURALIST [Vou. XLVI

dents who have expended considerable time and ‘‘gray matter’’
on the problem of the forerunner of the vertebrate notochord
the reviewer would heartily welcome a demonstration that the
organ ‘‘may be recognized in practically all segmented inverte-
brates’’; but until a clearer, more convincing description is fur-
nished us of some structure in ‘‘ practically all segmented inver-
tebrates’’ with which the vertebrate notochord is to be com-
pared, the question of homology in the strict sense would not
even be raised were the best interests of comparative anatomy
duly considered.

The other defect in the presentation of matters-of-fact which
a well-wisher for the book may justly fear will tend to prevent
as wide use of it as it deserves, is the circumstance that several
modes of statement occurring over and over again are bound to
give even the fairest-minded reader the impression that many
of the facts were prejudged; that is, were collected and recorded
not primarily on their merits, but in behalf of a theory. The
nomenclature employed in several important connections is
likely to have this effect. The use of the substantive cephalon
with various prefixes in purely descriptive matter dealing with
the thoracic region of arthropods is an example.

In the chapter ‘‘ Minute rig of the Brain and Cord of
Arachnids” we read (p. 8

The more posterior thoracic commissures, and those in the hindbrain,
are shorter, and the neural and hemal fascicles are widely separated,
leaving a space between them, which represents the beginning of the
fourth ventricle (italies by the reviewer).

Such dogmatic and unealled-for statements inserted into
purely descriptive matter are very unfortunate, for they can but
militate against the factual value of the work in the mind of
every candid reader. The feeling of uneasiness engendered in
the reader by these gratuitous dogmatizings, as to the extent to
which facts dealt with have been unconsciously colored by
theory, is not allayed by the author’s avowed attitude toward
the facts of organic structure. On page 469 we find this:

Comparative morphology has no value except in so far as it points
out the historie sequence of organie forms and functions, and reveals to
us the trend of evolution and the causes that direct and control it.

This statement I insist is not true. It may be partly true, but
in the unqualified form given it by the author is not only untrue,
but is provokingly and harmfully untrue. If for Dr. Patten

No. 550] NOTES AND LITERATURE 627

comparative morphology has no value except in the ways indi-
cated, well and good. I neither question nor quarrel with the
assertion. For me, however, comparative morphology has great
value in numerous other ways; and there is much evidence, both
historical and contemporaneous, that it has other values for
many biologists. It is, I submit, a real even though uninten-
tional harm, not only to individuals, but to biological science, for
an able morphologist to make an assertion which carries the
clear implication that the interest in and the valuations placed
upon comparative morphology by the men who devoted them-
selves to it long before anybody knew there was such a thing as
a “‘trend of evolution,’’ were an illusory or spurious interest
and valuation. And I would express my firm conviction that
biologists of this present childhood period of the evolution theory,
as the era from Darwin to the present day may well be called, must
come to see that great as is the value of morphology as a record
of evolution, this is still only one of its values; and further, that
until such perception is attained, just estimation of the facts of
morphology as a record of evolution will be impossible. The in-
terest of the future morphologist in his raw material ought to be,
according to my understanding, that of the pre-evolutionary
morphologist plus that growing out of the later discovery that
the facts of structure mark the ‘historic sequence of organic
forms and functions.’’ It is just because many, indeed all of
the facts dealt with in this work have values for me over and
above those attaching to them as a record of evolution that I re-
gret that they could not have been presented in a fashion less
caleulated to raise doubts in so many instances as to whether
they would appear exactly as they do but for the circumstance
of having been interpreted by the author in the light of his par-
ticular theory of historic sequence.

This consideration will, I trust, give real weight to my words
When I say that the criticisms I am passing on Patten’s mode
of presenting facts is an apology for a work of truly great fac-
tual worth, and not at all an attempt to discredit it. Nor would
I have any one understand me to be an advocate of ‘‘mere facts”?
of morphology; facts, that is, without any reference to their
wider bearings. My point is that all facts of morphology, as of
all other departments of biology, have so many ‘‘wider bear-
ings” that to write down as without value all except some one
set of these bearings, even so important a set as that of evolution,

.

18 to narrow the horizon of biological science. Against tenden-

628 THE AMERICAN NATURALIST [ Vou. XLVI

z
cies of this sort, wherever occurring, and unfortunately they
occur in many sections of the vast realm, I am ready at all
times to do battle.

o far this review and commentary has been made entirely
from the standpoint of observed and observable facts dealt with
in the volume. Now the point of view must be shifted to the
theoretical side.

Patten’s central thesis, as is well known, is that vertebrates
have descended from arachnids. The ‘‘arachnid theory of
origin of vertebrates,’’ or, for short, the ‘‘arachnid theory,’
the phraseology used by the author. The theory was first a
nitely set forth in 1889, the title of the original publication being
**On the Origin of Vertebrates from Arachnids.” A somewhat
abbreviated quotation of the author’s outline of the theory will
be justifiable. He writes:

This theory has formed the basis of all my subsequent work, and as
far as it went, is practically the same as the one presented here. In that
paper it was maintained that the vertebrates are descended from the
arachnid division of the arthropods, in which were included the typical
arachnids, the trilobites, and merostomes. The ostracoderms were re-
garded as a separate class, uniting the arachnids with the true verte-
brates. Limulus and the scorpion were the types most carefully studied,
because they were the nearest and most available living representatives
of the now extinct merostomes, or giant sea scorpions, Z were re-
garded as the arachnids standing nearest to the ostracoderm

Other evidence and conclusions were as follows: (1) In die arachnids
a forebrain vesicle is formed by the same process of marginal over-
growth as in the vertebrates. . . . (2)The kidney-shaped compound eye
of arachnids has been transferred to the walls of the cerebral vesicle in
vertebrates, giving rise to the retina, which still shows traces of omma-
tidia in the arrangement of the rod-and-cone cells. . (3) The arach-
nids have a cartilaginous endocranium similar in laine and location to
the primordial cranium of vertebrates. (4) They have an axial, sub-
neural rod comparable with the notochord. (5) In arachnids the brain
contains approximately the same number of neuromeres as in verte-
brates. . . . (6) The segmental sense organs (median and lateral eyes,
olfactory and auditory organs) are comparable with those in vertebrates-
The coxal sense organs are associated with special sensory nerves and
ganglia, comparable with the cranial dorsal-root nerves and ganglia
(suprabranchial sense organs) of verterates. (7) The basal arches of

appendages are comparable with the oral and branchial visceral
arches in vertebrates, (8) The tendency toward concentration of
neuromeres has narrowed the passage way for the stomodeum and
modified the mode of life in the arachnids. This ultimately led to its

No. 550] NOTES AND LITERATURE 629

permanent closure, the infundibulum and adjacent nerve tissues in
vertebrates representing the remnants of the old stomodeum with its
nerves and ganglia. . . .° (12) The process of gastrulation in verte-
brates and arachnids is confined to the procephalie lobes, in the place
where at a later period the primitive stomodeum appears. The so-
ealled “ gastrulation” of vertebrates. and arachnids is an entirely
different and independent process, that is, the process of adding by
apical or teloblastiec growth a segmented, bilaterally symmetrical body to
a primitive radially symmetrical head. (13) The arachnids resemble
the vertebrates in more general ways, as in the minute structure of
cartilage, muscle, nerves, digestive, and sexual organs. (Pp. xvii and
xviii of the Introduction.)

None of the statements in this list, either those quoted or those
not quoted, bring out clearly one of the most striking features of
the theory, namely, the supposition that the ventral surface of
the arachnid became the dorsal surface of the vertebrate. This
is well shown by several series of figures, as, for instance, that
on page eight, of imaginary arthropods and vertebrates with
creatures intermediate between them.

During the twenty years and more that the theory has been
before zoologists it seems to have won very few adherents; in-
deed, to have had little influence on biological thinking of any
sort. Nor does it seem probable that this final marshaling of
the evidence will accomplish much more as regards the main
contention. So far are we still from certainty as to exactly how
and when the back-boned animals originated that even the prob-
able evidence toward such knowledge is not great. Indeed, if
one will consider fully the nature and difficulties of the prob-
lem he will see that the chance of ever reaching certainty is
almost nil.

What would constitute a demonstration of the parenthood of
vertebrates ? Obviously the most indubitable proof, that of
direct observation, is out of the question. The ‘‘supreme canon
of historical evidence that only the statement of contemporaries
can be admitted,” must, in the nature of the case, be completely
ignored here. The best chance of reaching a demonstration is
m finding a series of fossil animals intermediate between some
unmistakably primitive vertebrate and the assumed ancestor,
containing no gaps great enough to raise serious doubt in the
mind of any competent authority as to the genetic relationship
__ “Number nine in this list is absent in the text. Numbers 10 and 11 are
Purposely omitted in the quotation.

630 THE AMERICAN NATURALIST [Von XLVI

of the two forms separated by the gap. The greatness of our
lack of a series connecting any vertebrate with any invertebrate
whatever, must impress one more and more the more he knows
and thinks about the problem. The only other conceivable mode
of demonstration would be to induce, experimentally, the trans-
formation of some living invertebrate into an undoubted verte-
brate. The possibility of accomplishing such a fact is so slight
that no biologist is likely to try it.

This brings us to a point where the transcendent importance
comes to view, of the logic of the interpretation of evolution,
understanding by evolution the transformation of one kind, or
species, of organism into another kind, or species. Except in
the relatively simple and rare cases of proof through direct ob-
servation, experimental or other, the fact that all evidence rests
back absolutely on resemblance—on similarity in form and com-
position of the parts of organisms—is of the utmost importance.
No evolutionist hesitates in either word or intention to accept
this tenet; yet when it comes to the actual ascertainment of like-
nesses and differences, and to speculating on their phylogenetic
significance, the danger of shifting from the inductive to the
deductive mode of reasoning, and of going down before the fal-
lacy known in the logic books as petitio principi, or surreptitious
assumption, is so imminent and subtle that very few of us, even
those most wide awake for pitfalls, avoid it wholly. Almost if
not quite all the numberless hypothetical ancestral organisms
that have been summoned to the aid of speculation on descent
during the last half century and more, are, I am convinced, vic-
tims of this evil, some in greater, some in less degree. The fal-
lacious reasoning usually runs something like this: Assuming
such-or-such an animal once existed, it is easy to see how a par-
ticular organ or part of the animal actually before us may have
arisen from that ancestry. Says Patten:

We have merely to strip off the superficial disguise of our hypothetical
arachnid ancestors and see whether either their underlying structure,
their mode of growth, . . . does or does not harmonize with the assump-
tion that they are the ancestors of the vertebrates (p. 3).

Before this ‘‘mere’’ stripping off begins, would it not be well
to consider ‘our hypothetical arachnid ancestors’’ rather closely?
If such an animal actually once existed, let it be granted for the
moment that it might have undergone such a transformation as
1s conjectured. Surely the first point to be pressed is as to the

No. 550] NOTES AND LITERATURE 631

evidence that the supposed creature did exist. And what is the
evidence? Why, the observed facts of arachnid structure, ex-
actly these and no others, else the ancestor would not be held as
hypothetical. No hypothesis can, of itself, add any new facts.
The hypothetical ancestor has, consequently, done nothing for
the case except to disguise the obvious difficulty there is in seeing
how an actual arachnid can be transformed into an actual verte-
brate.

The only really safe rule for using hypotheses in biology is
that no hypothesis shall be made except to help toward answer-
ing a question by formulating a clear. provisional answer to
that question. The making of hypotheses and using them before
they are themselves proved for the solution of other problems
than those to which they immediately pertain, is perilous busi-
ness. I may without danger, even with profit, construct an
imaginary enteropneust to aid my efforts to answer the question
of how the enteropneustic branchial apparatus or the organ in
this group called a notochord, arose phylogenetically. But be-
fore this imaginary creature can do me real service I must es-
tablish at least a strong probability that such an imaginary ani-
mal once actually existed. Nor must I fail to notice that such
probability can be established only by bringing new evidence
into the case. The facts on which I based the hypothetical ani-
mal can not be used over again to prove the reality of that ani-
mal. But if I must be thus cautious in resorting to an hypothet-
ical enteropneust for the interpretation of actual enteropneusts,
how much more need would there be for caution should I venture
to invoke an hypothetical enteropneust for interpreting actual
vertebrates !

Reflections of this nature opened my eyes several years ago
when I was struggling with the enteropneust hypothesis of verte-
brate descent, to the very slight chance there is of ever solving
the problem of vertebral origin.

Do I then regard all hypotheses on the problem as equal in
value because all alike are equally futile? By no means. I con-
sider each one of them to be of real worth, particularly those
that have been worked out as have Patten’s arachnid hypothesis,
Gaskell’s crustacean hypothesis, the annelid hypothesis and the
enteropneust hypothesis. They are of worth because each shows
quite clearly certain possible developmental courses that may
have been followed in the origin and progress of the great verte-
bral stock. To be explicit, it seems to me that Patten has shown

632 THE AMERICAN NATURALIST [ Vou. XLVI

that in certain structural details there is a similarity between
‘the exoskeletons of Limulus and some of the Ostrachoderms
‘which makes quite possible, not somewhat probable, genetic
relationship between these animals. Similarly the remarkable
likeness between the gills of the enteropneust and amphioxus, if
considered by itself, would make genetic kinship between these
‘animals highly probable; however, when considered along with
‘the whole organization and mode of life of each group, the
probability is much reduced, to such an extent, indeed, as to
become hardly more than a strong possibility.

If we take due cognizance of the extent to which our faith in
the general theory of organic evolution rests on similarities of
form and function among individual organisms which come
under our observation, we are in position to feel the weight of
‘the purely inductive evidence furnished by the many resemb-
lances between vertebrates and invertebrates brought to light by
the investigators in support of the various hypotheses of verte-
brate ancestry, that not only is the theory of evolution true as
applied to the back-boned animals, but that the ancestors of
‘these animals must have been in many respects like certain in-
-vertebrated animals with which we are familiar. The inductive
proof ef the truth of organie evolution, even of the general
‘course of evolution within large provinces of the living world,
approaches much closer to certainty when taken en masse than
‘does, of necessity, that pertaining to any single instance or small
group of instances.

So the moral drawn from examining the theoretical side of
‘the volume before us has this in common with that drawn from
‘examining the factual side: It has real worth, but that worth
would stand forth much more sharply, and probably would be
‘received by biologists with far greater sympathy, had it been
always and clearly distinguished in presentation from pure
‘matters of fact. And the worth would have been greater still
‘had the main and numerous subsidiary and corollary hypotheses
been frankly treated as far from demonstrated or even demon-
‘strable, but as varying in degree of reasonableness from very
possibly true to rather probably true.

Wm. E. RITTER

La JOLLA, CALIF.,
July 26, 1912

. XLVI, NO. 551

VOL

NOVEMBER, 1912

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-Hu dson, New Y
Articles containing research work bearing on the problems of oome evolu-

tion are ne welcome, and will be

given preference in publica

One hundrea reprints of cee aes are supplied to authors as of charge.

Further osare will > supplied at cos

Subscriptions Dukea should be sent to the Pi on The
subscription price is att dollars a year. Fore
Canadian postage twenty-five cents additional.
forty cents. The advertising rates are Fou

Tha saree tie seas seg
e . sn or single s
r Dollars for a page. ‘op

THE SCIENCE PRESS

_ NEW YORK: Sub-Station 84

Lancaster, Pa.

Entered , April 2, 1

+ tha Dace MAM

Garrison, N. Y.

, Pa., under the Act of

Congress of March 8, 1879.

FOR SALE
ARCTIC, ICELAND and GREENLAND
BIRDS’ SKINS,

Well Prepared Low Prices

Colle
Husavik, North Iceland, Via alai and

TENTH EDITION.
THE MICROSCOPE,

by HENEY GAGE, of University, Over
350 octavo and above 250 figures in this new
and fally revi: tion. Price $2.00, postpai:

COMSTOCK PUBLISHING CO., Ithaca, N. Y.

Reptiles, Amphibians
and Fishes

oot

Archives of Psychology

RECENT NUMBERS
19. An par sates Study of Certain Tests for
dividual Di

In fferences. Mary THEODORA WHIT-
LEY. Pp. 146. $1.25; cloth, $1.50.
An testi: various tests of
association, tion, movement, etc., with dications
of the relative ue of rent forms of test,

Visual Acuity with Lights of Different
rasat aly Davip EDGAR RICE.
Pp. 58. vention Gaeta 75 cents.

An

Korper ong vd
red, green, blue, and uncolored of the de-

ation of
ored illumination, and
see Sah ae am n cl seagate

as, i No. 1s = Seaman

THE
AMERICAN NATURALIST

Vout. XLVI November, 1912 No, 551

THE MENDELIAN NOTATION AS A DESCRIP-
TION OF PHYSIOLOGICAL FACTS

PROFESSOR E. M. EAST

Bussey INSTITUTION, HARVARD UNIVERSITY

As I understand Mendelism! it is a concept pure and
simple. One crosses various animals or plants and re-
cords the results. With the duplication of experiments
under comparatively constant environments these re-
sults recur with sufficient definiteness to justify the use
of a notation in which theoretical genes located in the
germ cells replace actual somatic characters found by
experiment. This is done wholly to simplify the descrip-
tion of the experimental results. If one finds that the
expression DR X DR—1DD+2DR+1RR# adequately
represents the facts in numerous breeding experiments,
he is then able to use the knowledge and the expression
in predicting the results of other similar experiments.
Mendelism is therefore just such a conceptual notation
as is used in algebra or in chemistry. No one objects to
expressing a circle as 2?-+-y?—r*. No one objects to

*I do not speak ES of the new biological facts discovered by Mendel
or by his followers. Facts are always facts. Alternative inheritance and
character recombinations were important facts, but I think no one will deny
that the greatest value of Mendel’s facts arose from the mathematical treat-
ment he gave them. This mathematical notation remains conceptual just
as does the chemical formula, but it must have as much basis of fact as there
are pertinent facts extant.

633

634 THE AMERICAN NATURALIST [ Vou. XLVI

saying that BaCl, + H,SO,—BaSO,+2HCl. No one
should object to saying that DR + RR—1DR-+1RR.
We push things into the germ cells as we place the
dollars in the magician’s hat. Hocuspocus! They dis-
appear! Presto! Out they come again! If we have
marked our money we may find that that which appears
from the magician’s false-bottomed hat is not the same
as that which we put in. But it looks the same and is
good coin of the realm. We have a good right there-
fore to poke our characters into the germ cell and to pull
them out again if by so doing we can develop—not a
true conception of the mechanism of heredity—but a
scheme that aids in describing an inheritance. We may
do this even as we may use algebraical and chemical no-
tations, if we remember that x? + y? does not give us a
circle, that a chemical equation does not represent a true
reaction or prove the atomic theory, that we have not
pulled something new and astonishing out of the germ
cell, that a unit factor represents an idea and not a real-
ity, though it must have a broad basis of reality if it is
to describe a series of genetic facts.

The facts of heredity that one describes in the higher
organisms are the actual somatic characters, variable
things indeed, but still things concrete. Their potential-
ities are transmitted to a new generation by the germ
cells. We know nothing of this germ cell beyond a few
superficial facts, but since a short description of the
breeding facts demands a unit of description, the term
unit factor has been coined. As I hope to show, a factor,”
not being a biological reality but a descriptive term,
must be fixed and unchangeable. If it were otherwise
it would present no points of advantage in describing
varying characters. The only obvious reason for poking
it into the germ cell is to distinguish thus the actual
parent (the cell) from the putative parent (the carrier).

° I hope this statement is not confusing. The term factor represents in &
way a biological reality of whose nature we are ignorant just as a structural
molecular formula represents fundamentally a reality, yet both as they are
used mathematically are concepts.

No. 551] THE MENDELIAN NOTATION 635

If we forget ourselves and begin to speak of unit factors
as particles, only a confusion follows similar to that
caused by Nägeli, Spencer and Weismann. Nothing is
gained and even facts are obscured.

THE Score or MENDELISM

How far may we carry this conceptual notation? My
answer is: just as far as the notation interprets the facts
of breeding and is helpful. Interest in the scope of Men-
delism is now focused on two phases, complete and par-
tial coupling and the interpretation of so-called size char-
acters. Complete coupling in the transmission of char-
acters apparently non-related has been shown in a large
number of cases. Perhaps those best worked out in ani-
mals are the sex-coupled or sex-limited characters ob-
served by Morgan in Drosophila. In plants, cases ob-
served by Emerson and by Bateson and his coworkers
are equally clear. Emerson has shown beyond a rea-
sonable doubt that the characters he describes are inde-
pendent of each other, and can not be represented by one
factor. Bateson has recently corroborated the observa-
tion on other characters. Besides this phenomenon,
Bateson has discovered partially coupled characters.
All three of these writers, have subsidiary hypotheses to
account for their facts. Bateson, when discussing per-
fect couplings, merely says that the characters come out
in F, coupled in the way they went in in the grandpa-
rents, which naturally is only a restatement of the facts.
Morgan and Emerson deal in pictures of carrying bod-
ies. Both of their theories fit their own facts as they
necessarily would. Emerson and, I may say, myself be-
lieve Morgan’s theory incompatible with that of Emer-
Son. Morgan believes his theory adequate for both
cases. Without discussing the merits of these particular
hypotheses I think it is agreed that some characters do
0 into the F, generation and come out from it together
that are in other cases independent. The importance
of the phenomenon is greater than the theory at present.

636 THE AMERICAN NATURALIST [ Vou. XLVI

It has been questioned whether one has the right to con-
tinue to couple characters in large numbers to interpret
facts, because by proper coupling one may interpret
almost any fact, and place himself in a logically unassail-
able position. But this is no reason for not coupling
factors as much as one pleases if it is helpful and if all
of the facts fit. A propos of this statement I might say
that I have recently remade the historical old cross first
made by Kölreuter in 1760, Nicotiana rustica X N icotiana
paniculata. These species differ in many details—habit
of growth, size, shape and hairiness of leaf, inflorescence,
and size and shape of flower and fruit. Both of the pa-
rent species have been reproduced exactly from a par-
tially fertile F, in a total number of less than 200 F,
plants. One may formulate an hypothesis of selective
elimination of gametes combined with selective fertiliza-
tion that helps to describe the facts, but unless large
numbers of factors are coupled together I believe it to
be impossible to account for all the facts by the usual
Mendelian notation.

Before leaving this subject it might be mentioned that
Bateson’s theory, originated to account for partial coup-
ling, keeps the idea of factors segregating from their
absence, but instead of A and a being formed in equal
quantities as in ‘‘regular’’ Mendelian notation, they are
to be formed in series represented by the scheme n—14B:
14B:1aB:n—1ab. I do not believe one should hasten
to accept this description, although Bateson’s F, gen-
eration facts certainly fit and have been recently sup-
ported by Baur. My reason for making this statement
is that as yet Bateson’s F, facts do not fit the theory.
Some of them would even make necessary two or more
different kinds of factorial distribution in the same plant
varieties. On this score the helpfulness of our notation?

*Here is a good illustration of the Mendelian notation as a concept.
Supposing the gametic distribution n —14B:14b:1aB:n—1ab were to
fit all the facts in the case, then no one could object to its use. If it were
to be demonstrated that segregation occurred at the reduction division, how-
ever, the scheme no longer fits the facts and must be abandoned.

No. 551] THE MENDELIAN NOTATION 637

is impaired and this is the only excuse for its existence.
Furthermore, while it has not been proved that the phe-
nomenon we call segregation occurs at the reduction
division, the presumption is in favor of that view. The
work of Webber, Correns, Lock, Emerson and myself on
Xenia in maize indicates that segregation does not take
place immediately after reduction, while the work of the
Marchals on regeneration in mosses indicates that it
does not take place before reduction.

Now to turn to the kinds of variation that may be de-
scribed by the Mendelian notation. Owing to its youth,
we can all remember how we wondered, as each new case
came up, whether Mendelian phraseology would fit.
Since qualitative characters were the ones that could be
divided into definite categories they were the ones at-
tacked. One by one they were analyzed. The phraseol-
ogy did fit. Qualitative characters however form a very
small proportion of the characters in animals and plants.
The numerous characters are the quantitative, the size
characters. If Mendel’s law is to be worth anything as a
generality, therefore, it must describe the inheritance of
these characters. :

To some of us Mendel’s law from the first seemed
destined to be a notation generally useful in describing
inheritance in sexual reproduction. This conclusion was
indicated by the simple fact that Mendel’s law described
many cases in both the animal and the vegetable king-
dom. It was inconceivable that this should be the re-
sult of coincidence. It was therefore still more mecon-
ceivable that only a small portion of the facts in each
kingdom should come under the scope of Mendelism.

A basis for the inclusion of quantitative characters
was obtained when Nilsson-Ehle and the writer showed
that certain qualitative characters gave ratios of 15:1,
63:1, etc., in the F, generation, and in other ways be-
haved so that they might be described only by assuming
more than one independent gametic factor as the germ
cell representative of the character, if the orthodox idea
of segregation were retained. From these phenomena

638 THE AMERICAN NATURALIST [ Vou. XLVI

it was immediately seen that where dominance is ab-
sent and such multiple factors are assumed, size char-
acters can be interpreted as coming under the Mendelian
law. When dominance is complete the mathematical
representation of an F, generation is (3/4+ 1/4)"
where n represents the number of factorial differences
involved; as the manifestation of dominance becomes
less this formula approaches the type (1/2 + 1/2)”.
The difference between the heredity of qualitative char-
acters and quantitative characters is therefore only one
of degree, for there is absence of dominance in cases of
simple monohybrid qualitative characters and there is
presence of multiple factors in cases of qualitative char-
acters showing dominance. But it is manifestly absurd
to expect size characters to appear in natural groups as
do. many qualitative characters. The marked effect of
environment and our ignorance of the exact effect to
attribute to each factor precludes it. One can determine
whether size inheritance compares with the inheritance
of qualitative characters only by the use of arbitrary bio-
metrical methods. In theory, homozygotes with size dif-
ferences when crossed should give an intermediate F,
of low variability and an F, of high variability. Vari-
ous F, populations should differ in their mean and in
their variability. The difference in the variability of
F, over F, should decrease as the heterozygosity of the
parents increases. Sometimes parents of the same size
should differ in the factors they contain and the F, gen-
eration should contain individuals smaller and individ-
uals larger than either of the parents. Each of these re-
quirements has been satisfied by experiment. East and
East and Hayes have tested it for number of rows per
cob, height of plant, length of ear and size of seed in
maize, Shull for number of rows in maize, Emerson for
fruit sizes in maize, beans and gourds, Tammes for
various characters in flax species, Tschermak for time of
blooming in beans, Hayes for number of leaves in to-
bacco, Belling for certain characters in beans, Phillips
for body size in ducks, MacDowell for body size in rab-

No. 551] THE MENDELIAN NOTATION 639

bits. In these investigations every test possible for the
theory has been satisfied. No criticism could be made ex-
cept that certain of the characters used varied consider-
ably in the mother varieties and therefore were pre-
sumably not homozygous for all character factors. This
criticism is apparently answered by a recent investiga-
tion of the writer’s, as yet unpublished, where two
species, Nicotiana forgetiana and Nicotiana alata grandi-
flora were crossed. As seen by the table, the corolla
length is very slightly variable in either species, nor is
it affected to any extent by environment, yet each species
was absolutely reproduced by recombination in the F,
generation.
TABLE I
FREQUENCY DISTRIBUTIONS FOR LENGTH OF COROLLA IN A CROSS BETWEEN
Nicotiana forgetiana (314) anv N. alata grandiflora (321).

: Class Centers in Millimeters
Designation < EEE E SI ; i ; j
20 | 25 | 30 | 35 | 40 45 | 50 | 55 | 6o | 65 | 70 | 75 | 80 | 85 | 90
| meen rt aan ook
314 9 |133| 28 | | |
1/19/50 | 56|32| 9
(314 X321) Fi 3 | 30| 58| 20 | |
— (314 X321) F 5| 27 |79 |136]125|132/ 1021105! 64 |30 15! 6) 2)

Coefficients of variation are: 314 = 8.86 + .33 per cent.; 321 = 6.82 + .25
per cent.; (314 X 321) F, = 8.28 + .38 per cent.; (314 X 321) F, = 22.57 +
39 per cent.

H
I do not believe that biologists have sufficient facts as
yet to warrant any concrete meaning being given to their
notation as regards germ-cell structure, but I do main-
tain that the Mendelian notation satisfies the facts of
size inheritance as well as it satisfies the facts of quali-
tative inheritance. As a description, it goes the whole
way. If qualitative inheritance is Mendelian, quantita-
tive inheritance is Mendelian; if quantitative inheritance
is not thus described, qualitative inheritance is described
not a whit better.
All writers do not agree with this statement; never-
theless, speaking for myself only, I believe it to be be-
yond question. Castle (Amer. Nart., 46: p. 361) says:

It is quite possible that we are stretching Mendelism too far in

640 THE AMERICAN NATURALIST [Vor. XLVI

making it cover such cases. Dominance is clearly absent and the only
fact suggesting segregation is the increased variability of the second as
compared with the first hybrid generation. This fact however may be
accounted for on other grounds than the existence of multiple units of
unvarying power.

If size differences are due to quantitative variations in special
materials within the cell, it is not necessary to suppose that these
materials are localized in chunks of uniform and unvarying size, or
that they occur in any particular number of chunks, yet the genotype
hypothesis involves one or both of these assumptions. oth are un-
necessary. Variability would result whether the growth-inducing sub-
stances were localized or not, provided only that they were not homo-
geneous in distribution throughout the cell. Crossing would increase
variability beyond the first generation of offspring because it would
inerease the heterogeneity of the zygote in special substances (though
not its total content of such substances) and this heterogeneity of struc-
ture would lead to greater quantitative variation in such materials
among the gametes arising from the heterozygote. Thus greater varia-
bility would appear in the second hybrid generation.

I can not agree with this statement as I understand it,
though this disagreement may be due to my own limita-
tions. We do not stretch Mendelism and we do not make
it cover such cases. The facts of breeding have been ob-
tained and the Mendelian notation expresses them. That
is all that it is necessary to claim. It is not precisely
true, however, to say that increased variability in the
second hybrid generation is the only fact to be expressed.
It is of paramount importance that various F, individ-
uals giving F, populations differing in mean and in
variability, should be included in the Mendelian descrip-
tion. They are included.

Again, Castle states that the genotype conception as-
sumes the localization of the hypothetical factors either
in chunks of uniform and unvarying size, or that they are
carried by a particular number of chunks. I am unaware
of any such assumptions. It is true that some such pic-
ture has been suggested as a diagram helpful to the
imagination in its conception of the scheme as a me-
chanical process, but this is purely and simply a dia-
gram. The real matter under discussion is that the
breeding facts are adequately described in a notation `
essentially Mendelian.

No. 551] THE MENDELIAN NOTATION 641

Of course Castle’s scheme of expressing the facts by
heterogeneity in the germ cell might serve. He pro-
duces increased variability in the second hybrid genera-
tion by greater differentiation among the gametes aris-
ing from the heterozygote. But one can also describe
inheritance of qualitative characters in the same way,
and one gains no system by it. It is a return to the type
of expression used by Nägeli, Naudin and De Lage in
pre-Mendelian days. It is simply a trans-nomination
possessing no advantages.

Before leaving this phase of the subject, I must speak
of Davis’s recent fine paper (Amer. Nart., 46: p. 415) on
his crosses between (Enothera biennis and Cnothera
grandiflora. As I have had the advantage of seeing his
cultures many times in the past two years, I am in a fair
position to draw my own conclusions as to the meaning
of his data. In regard to his F, generation from the
hybrid plant marked 10.30 L b he says:

1. In the immensely greater diversity exhibited by the F, generation
over that of the F, is clearly shown a differentiation of the germ plasm
expressed by the appearance in the F, plants of definite tendencies in
directions toward the two parents of the cross. This seems to the writer
the essential principle of Mendelism and does not necessarily involve
the acceptance of the doctrine of unit characters and their segregation
in either modified or unmodified form.

2. Certain characters of the parent species have appeared in the F,
segregates in apparently pure condition, but the very large range of
intermediate conditions indicates that factors governing the form and
Measurements of organs (if present at all) must in some cases be con-
cerned with characters so numerous and so small that they can not be
Separated from the possible range of fluctuating variations. If this is
true such characters seem beyond the possibility of isolation and analysis
and the unit character hypothesis for these cases has little more than a
theoretical interest.

3. Both cultures certainly showed marked progressive advance in the
range of flower size, the largest flowers having petals somewhat more
than 1 em. longer than those of the grandiflora parent. There 7e
similar advance in the size of the leaves and the extent of their erinkling.
These progressive advances would seem to demand on the unit charaeter

Se either the modification of the old or the creation of new
rs,

4. The absence of classes among the F, hybrids (except for the

642 THE AMERICAN NATURALIST [Vot XLVI

dwarfs) further works against the unit character hypothesis as of
practical value in the analysis of a hybrid generation of this character,
It should be remembered, however, that there were in this cross no
sharply contrasted distinctions of color, anthocyan (stem) coloration
proving most unsatisfactory for the purpose of a genetical study.

These four paragraphs are practically a résumé of
Davis’s genetic facts: I take exception only to some of the
implied conclusions. It is quite evident that Dr. Davis
believes that many breeding facts are expressed in
shorthand by the Mendelian notation. His statements,
however, imply a feeling of loss of caste or something of
the kind if he makes definite use of Mendelian phraseol-
ogy. His F, generation was exactly what would be ex-
pected when several Mendelian units without dominance
segregate and recombine. The advance in size of corolla
was predicted by me in 1910 (Amer. Nar., 44: p. 81) as
a direct consequence of size inheritance. It has since
been demonstrated by Tschermak for time of blossom-
ing of beans and clearly analyzed by Hayes for number
of tobacco leaves. It demands neither modification of
old nor the creation of new factors. It occurs when-
ever AABB (size factors) is crossed with CCDD, and
each factor is allelomorphic to its own absence, to use
the ordinary phraseology.

As to the difficulty of precise analysis into factors, I
agree with Dr. Davis, but that there is no advantage in
showing that this behavior is described in typical Men-
delian terms I can not admit. One holds the same prac-
tical advantage here—though the case is complex—that
one holds in all Mendelian inheritance. He knows that
somatic appearance is not the criterion of breeding ca-
pacity, but that it is determined in some way by gametic
constitution, although no germ cell architecture is pre-
supposed. He knows that recombination of some kind
of factors occurs and has some idea of the number of
progeny to be grown to obtain the desired combination.
In other words, the blend in F, does not indicate com-
plete loss of extremes.

No. 551] THE MENDELIAN NOTATION 643

Menvew’s Law anp Gatton’s Law

The above statement leads into a discussion of Men-
del’s law of heredity as compared with Galton’s law, for
in itself it is almost a statement of the difference. As
Bateson was the first to emphasize, organisms inherit
from parental germ cells only, therefore a law of an-
cestral heredity is a fallacy and a misnomer. The
simple illustration that of two individuals alike in ap-
pearance one is homozygous for a character and the
other heterozygous for the same character, shows the
superficial reasoning that leads to the correlation coeff-
cient as a measure of heredity. Parental and filial pop-
ulations may show correlation, but that is only a matter
of averages and not a measure of the inheritance.

Professor Castle has recently disclosed the probable
Mendelian basis for Galton’s data on coat color of
Bassett hounds by showing the inheritance of tricolor
coat in guinea-pigs, yet he makes the surprising state-
ment that ‘‘as regards height, however, and other size
characters, Galton’s law is quite as good a basis for pre-
dicting the result of particular matings as is Mendel’s.’’
The arch priest of biometry, Karl Pearson, does not
claim that Galton’s law can predict the result of individ-
ual matings. Similarly, Mendel’s law predicts only by
averages. It says that where DR meets DR, there will be
on the average 1DD:2DR:1RR produced. Where the
classes are larger the prediction in increasingly compli-
cated. But the prediction is as good for size characters as
for qualitative characters of the same complication. And
there are such qualitative complications, as is manifest by
Castle’s formula of AACCUUIIYYBBBrBrEE for a
wild rabbit’s coat color. The difference between Gal-
ton’s law and Mendel’s law is that the true criterion of
the germ plasm of any individual is its breeding power
and not the somatic appearance of its back ancestry.
This is as true of size characters as of any other char-
acters.

644 THE AMERICAN NATURALIST [ Vou. XLVI

THE GENOTYPE CONCEPTION or HEREDITY

Expressed in Johannsen’s words, the basis of the
modern conception of heredity is: ‘‘Personal qualities
are the reactions of the gametes joining to form a zygote;
but the nature of the gametes is not determined by the
personal qualities of the parents or ancestors in ques-
tion.” The quotation expresses well the idea that I
have just tried to convey, and from it one sees plainly
that it is the correlation that necessarily appears to a
greater or less extent between the somatic qualities of
two generations when they exist in large numbers that
gave the basis for Galton’s superficial law.

This quotation is Johannsen’s slogan for the geno-
type conception of heredity. As there stated, it is
merely a generalized expression of the essential fea-
tures of the Mendelian notation. Johannsen, therefore,
was the first to admit the broadness of its scope. In his
exposition of his position, however, he adds two sub-
sidiary propositions that we will now discuss; the first
is the perennial question of the possibility of the inherit-
ance of acquired characters, the second is a question
which from its illusivenesg is likely to take on a peren-
nial habit—that of the relative constancy of unit char-
acters, :

In regard to the first question I must be content here
with a mere general statement. Like Osborn I would
emphasize the possibly delusively static condition of
organisms when tested during the infinitesimal time
usually devoted to experiment. The inheritance of ac-
quirements in some subtle way unknown to us may have
been of immense importance in evolution. On the other
hand, some sort of an orthogenesis may account for all
the facts without the inheritance of acquired characters.
It scarcely seems possible that everything is mere chance,
though one who has studied plant teratology is as-
tounded at the almost infinite number of characters that
have appeared that were absolutely dangerous to the
individual in its contest for survival. Be that as it may,
I simply wish to acknowledge unbelief in any so-called

No. 551] THE MENDELIAN NOTATION 645

proof that the inheritance of acquired characters is im-
possible. At the same time one must admit that no un-
questionable proof of such inheritance has ever been
submitted. Experimental evidence is woefully negative.
It seems only reasonable, therefore, considering the
available corroborative evidence, to relegate the expres-
sion of new characters to variations that have affected
the potentialities of the germ cells. We can simply di-
vide variations into the classes inherited and non-in-
herited without any admission as to their cause. We
can call the inherited variations mutations if we will, or
we can give them any other name. We must simply re-
member that they are both large and small.

One can hardly agree with Osborn that large varia-
tions which are not in an orthogenetic line have had
little value in evolution, or that teratological phenom-
ena are of little consequence. The production of iden-
tical quadruplets in the armadillo can scarcely be a grad-
ually perfected character. Zygomorphism in flowers is
lost as a unit and although this does not prove its birth
as a unit, still that is to be presumed. One could fill
pages with such data, but this is hardly the place for it.
We will therefore consider the relative constancy of
what we know as a unit character.

THe Constancy or Unit FACTORS

The first thing one does if he wishes to oppose the
idea of a unit character is to ask for a definition. A per-
fect definition of a unit character is as difficult to formu-
late as for a flower, yet one can obtain an idea of a
flower by proper application. If one describes a unit
character as the somatic expression of a single gametic
- factor or heredity unit, he at once gets into trouble. As
the factor and not the character is the descriptive unit,
a unit factor may affect a character but that character
may never be expressed except when several units co-
operate in ontogeny. I should prefer to disregard the
word character therefore in formulating the problem.
The real problem is: Are the facts of. heredity ade-

646 THE AMERICAN NATURALIST [Vor. XLVI

quately described by unvarying hypothetical factors?
It is my thesis that if they can not be so described, the
Mendelian notation fails.

Johannsen was the first to show the relative constancy
of characters by his beautiful experiments on beans.
Since that time, experiments designed to show change, if
present, have yielded negative results on bisexual animals
such as poultry (Pearl), on plants such as peas (Love),
beans (Johannsen’s later work), maize (Shull, East,
Emerson), on asexual animals such as hydra (Hanel),
paramecium (Jennings) and on asexual plants such as
bacteria (Barber and others), and potatoes (East).

Three critics have appeared. Karl Pearson took up
the gage of battle because Johannsen’s work shows the
utter untenability of the correlation coefficient as a
measure of heredity. He has produced no evidence to
uphold his view. Harris, following Pearson for a like
reason, has concluded against Johannsen, but has not
yet presented his data for public criticism. There re-
mains the work of Castle, which he believes is supported
by the work of Woltereck. The question to consider then
is whether the work of these two investigators justifies
the contention.

Castle states that by selection he has modified a unit
character. No one questions that under certain condi-
tions changes in characters are made manifest by selec-
tion. It has been done again and again. The question
as I see it is the following: Are not the facts presented
by Castle and the facts of the pure-line workers described
most concisely and in a way most helpful to investiga-
tion, by the reactions of fixed and unchanging units? If
they can not be thus described the use of units is an ab-
surdity, for one can not measure or describe by changing
standards.

Castle’s principal work on selection is with a fluctua-
ting black and white coat pattern—the so-called hooded
es, writing of these experiments, Castle says (l. c.,
p. de ¢

No. 551] THE MENDELIAN NOTATION 647

I shall speak first of the case least open to objection from the geno-
type point of view, which requires:

1. That no cross breeding shall attend or shortly precede the selection
experiment, lest modifying units may unconsciously have been intro-
duced, an

2. That only a single unit-character shall be involved i in the experiment.

These requirements are met by a variety of hooded rat which shows a
particular black and white coat pattern. This pattern has been found
to behave as a simple Mendelian unit-character alternative to the self-
condition of all black or of wild gray rats, by the independent investi-
gations of Doncaster, MacCurdy and myself. The pigmentation how-
ever in the most carefully selected race fluctuates in extent precisely as
it does in Holstein or in Dutch Belted cattle. Selection has now been
made by Dr. John C. Phillips and myself through 12 successive genera-
tions without a single out-cross. In one series selection has been made
for an increase in the extent of the pigmented areas; in the other series
the attempt has been made to decrease the pigmented areas. The result
is that the average pigmentation in one series has steadily increased, in
the other it has steadily decreased. The details of the experiment can
not be here presented, but it may be pointed out (1) that with each
selection the amount of regression has grown less, 2. e., the effects of
selection have become more permanent; (2) that advance in the upper
limit of variation has been attended by a like recession of the lower
limit; the total range of variation has therefore not been materially
affected, but a progressive change has been made in the mode about
which variation takes place.

. The plus and minus series have from time to time been crossed
with the same wild race. Each behaves as a simple recessive unit giving
a 3:1 ratio among the grandchildren. But the extracted plus and the
extracted minus individuals are different; the former are the more
extensively pigmented.

e series of animals studied is large enough to have significance.
It ineludes more than 10,000 individuals.

The conclusion seems to me unavoidable that in this ease selection has
modified steadily and permanently a character unmistakably behaving
as a simple Mendelian unit.

This conclusion, from the writer’s standpoint, is not
only avoidable, but unnecessary. No direct or implied
denial of these facts is made, but a shift is made in the
Point of view. It seems to me a logical necessity that
hypothetical units used as measurement or descriptive
Standards be fixed. The problem to be solved is the
simplest means of thus expressing the facts. If the most

648 THE AMERICAN NATURALIST [Vot. XLVI

definite characters—i. e., certain pure-line homozygotes
are sufficiently constant in successive generations to
be expressed by a fixed standard, well and good. The
whole heredity shorthand is then simple. If such is not
the case, the character must still be described by some
fixed standard, but in that case recourse must be had to
complex mathematical expressions and not to a single
unit to describe the most constant somatic expressions.
Furthermore, if these mathematical expressions served
any practical purpose, it would be necessary to prove
that all somatic variability of homozygotes under uni-
form conditions (if there is any) may be expressed by
very few formulas.

Such an attitude does not seem to be in harmony with
the progressive spirit of the times. I believe that we
may describe our results simply and accurately by hold-
ing that unit factors produce identical ontogenetic ex-
pressions under identical or similar conditions. If under
identical conditions the expression is different, then a
new standard, a new unit, must be assumed; that is, fac-
tor A by any change becomes factor B. The results of
the pure-line investigations are the warrant for this in-
terpretation, for they are the investigations of success-
sive generations of somatic expressions with the least
attendant complication. From them one may assume
that a succession of individuals homozygous in all char-
acters and kept under identical conditions will be alike.’
To be sure there are numerous changes in the expression
of characters when external and internal conditions are
not so uniform as the above, but I believe that these
changes can all be described adequately and simply by
ascribing them to modifying conditions both external
and internal. When external we recognize their usual
effect in what we called non-inherited fluctuations, when
internal we recognize their cause in other gametic fac-
tors inherited independently of the primary factor but

* Possibly even under these conditions rare variations that are exceptions
to this rule might occur. In other words, mutations might occur having 20
external cause and therefore to be left for vitalistic interpretation, but this
would not affect the general situation.

No. 551] THE MENDELIAN NOTATION 649

modifying its reaction during development. This is a
physiological conception of heredity, as it recognizes the
great coöperation between factors during development.
It is a very simple conception of heredity, moreover, for
it allows a multitude of individual transmissible differ-
ences with the assumption of a very few factors. Some
illustrations will be given later that will show the idea
underlying this theory. Let us now see whether Castle’s
work can be described properly by it.

Castle started with a peculiar character. It fluctuates
continually and has never been bred to as small a varia-
bility as have many other characters. I have worked
with a somewhat similar character in maize. It is a
variegated pericarp color. In experimenting with it I
have raised over a thousand progeny in one generation,
a thing manifestly impossible with rats. Both solid
colored ears and white ears have been obtained, and
while at present it would be unwise to draw definite con-
clusions, it appears that both solid red ears and white
ears of this kind give again variegated progeny. In
other words, neither the red ear nor the white can be-
have like a normal red or white ear, but as if the pattern
had fluctuated so widely that it can not appear on the
ear (this explanation was suggested by Emerson). At
any rate, we may conclude that the rat pattern fluctuates
widely and is therefore markedly affected by some con-
dition either internal or external.

Castle began therefore with a character in a fluctua-
ting condition, possessed by a race which had not re-
cently been crossed with a different race. This does not
mean, however, that the various individuals forming his
original stock did not differ in several factors that in
their different combinations might have an effect upon
the developing pattern. Suppose for the moment that
this were actually the case. If he had been able to pro-
duce a fraternity by a single mating numbering several
thousands, he would have produced individuals with all
of these combinations of other genes. It is probable
that he would then have obtained his progressive ex-

650 THE AMERICAN NATURALIST [Vou. XLVI

tremes in one generation, extremes that were never seen
when but few individuals were produced. This sort of a
thing is not hypothetical. It is mathematically demon-
strable that with the same variability (a + b)” expanded
gives an increase in the number of classes as the total
number of individuals increases. It is, moreover, sup-
ported by the experimental evidence of De Vries on se-
lecting for higher number of rows in maize. I, myself,
by using greater numbers obtained an increase in protein
in maize in one generation comparable to that obtained
by the Illinois Agricultural Station in three generations.

Castle further argues that decrease in regression
toward the original mean supports his view. On the
other hand, this is exactly what should take place on as-
suming the truth of the fixed factor conception, as has
been shown by Jennings.

Again, the selected races when crossed with wild races
both act as simple recessives, but the extracted plus in-
dividuals are more pigmented. This is what I should
expect. The extracted plus individuals would be more
pigmented when existing in small numbers, because the
modifying factors are several. If several thousand
progeny were grown, however, recombinations would
show a more varied result. And as a matter of fact, ex-
tracted recessives from the plus race are not precisely
comparable in their fluctuation to the selected race with
which the wild was crossed. They are more variable than
the progeny of an inbred hooded individual of the same
grade as the parent used in the cross with the wild. I
do not think that one has a right to say, therefore, that
there were no modifying genes present in various com-
binations in the extracted recessives.

When the selected lines were crossed together, more-
over, the resulting progeny were somewhat intermediate
and variable. The grandchildren were more variable.
This is what should result from our assumptions. The
animals are homozygous as far as having a pattern is
concerned, but they differ in several genes that affect
the development of the pattern.

No. 551] THE MENDELIAN NOTATION 651

Taking into consideration all the facts, no one can
deny that they are well described by terminology which
requires hypothetical descriptive segregating units as
represented by the term factors. What then is the object
of having the units vary at will? There is then no value
to the unit, the unit itself being only an assumption. It
is the expressed character that is seen to vary; and if
one can describe these facts by the use of hypothetical
units theoretically fixed but influenced by environment
and by other units, simplicity of description is gained.
If, however, one creates a hypothetical unit by which to
describe phenomena and this unit varies, he really has
no basis for description.

The facts obtained when working with pied types are
complex. They are evidently not thoroughly understood
as is evidenced by a different interpretation made by
every worker who has investigated them. Doncaster
and Mudge see two types of Irish rat. Why not three or
four? Crampe obtained hooded rats from cross of self-
colored and albino, the hooded coming only from hetero-
zygotes having some white. No adequate explanation
has been given. Cuénot concluded regarding pied mice
with several degrees of piedness that each was recessive
to the other of next higher grade. In fact the behavior
of self colors and spotted colors among mammals as
among plants is pretty well ‘‘confused,”’ as in several
species spotted types dominant to self color are known.

Castle’s other experiments in selection—the forma-
tion of a four-toed race of guinea-pigs starting with one
animal with a rudimentary fourth toe, and the perfec-
tion of a silvered race of guinea-pigs from an animal in
which the character was feebly expressed—need not be
considered here. Both were necessarily crossed with
normals at the start, and gradual isolation of races hav-
ing the proper gene complex for complete expression of
the characters is to be expected. There have been nu-
merous selection experiments of this type—such as those
of De Vries, the Vilmorins, the Illinois Agricultural Ex-
“periment Station, ete.—that have yielded results.

652 THE AMERICAN NATURALIST [ Von. XLVI

But these results, with one possible exception, were
open to the criticism that they probably had to do with
mixed lines and could therefore be described by the no-
tation we have used. The experiments on pure lines have
given no such results. One should not be asked to accept
the results of the unguarded experiments and disregard
the results of the guarded investigations.

The one possible exception alluded to above refers to
the experiments of Woltereck (Deut. Zool. Gesell., 19:
110-173, 1909) on parthenogenetic strains of Hyalo-
daphnia and Daphnia where there can be no question of
gametic recombination. This experiment is not beyond
criticism as will be seen later, but if it were our position
would not be affected. The results would still have to be
described by some fixed standard but the description
would be complicated. Since it is not beyond criticism,
there is yet no reason for such a complication.

Woltereck’s work was primarily to show whether or
not acquired characters are inherited. It was a second-
ary object to find out whether small variations or distinct
sports occurred in the species. Those who use the work
as an argument for unit factor modification, therefore,
should also accept his inheritance of characters acquired.

Woltereck tested the effect of selection on seven char-
acters. Selection gave no results in five cases. The first
supposedly successful case is for difference in head
height. In different pure lines he found an enormous
effect of environment. He therefore endeavored to plot
curves for different kinds of environment, food, tempera-
ture, generation number, etc. By comparing these
curves he makes an argument for the inheritance of
small acquired variations. In the absence of control cul-
tures, and from the fact that culture conditions very
uniform to Dr. Woltereck may have been somewhat ex-
treme to Mr. A. Daphnia, the argument has only the
value of the other numerous scholastic defences of in-
herited acquirements. It is criticized by Tower in a re-
cent publication. Woltereck did obtain one inherited

No. 551] THE MENDELIAN NOTATION 653

head variation. It apparently arose suddenly. He calls
it a mutation.

The only result that can be considered seriously from
the standpoint taken in this paper is the result when se-
lecting for a rudimentary eye. Daphnia has been dis-
tinguished from Hyalodaphnia by the presence of a rudi-
mentary eye. The distinction does not seem to be valid,
for Woltereck noticed rudimentary eyes several times in
pure line cultures of Hyalodaphnia and they have also
been seen by others in wild cultures. He regards the
phenomenon as a reversion to a preexisting condition.
He found that the presence of the rudimentary eye is
periodic. In the spring it appears, in the summer it
again disappears. Wither kind can produce progeny of
the other kind. From this fact it seems reasonable to
believe that environment or generation number has much
to do with the expression of the character, although
Woltereck in one place inclines to the opinion that ex-
ternal factors affect it but little. He performed several
experiments on the effect of light and temperature, how-
ever, and says that provisionally they gave no result
free from objection—‘‘ . . . gegaben einstweilen kein
einwandfreies Resultat.’? Almost any interpretation
can be given this statement. :

From a pure line in which this variable eye spot ap-
peared, he isolated a mother and grandmother with the
character well developed. Ninety per cent. of the
progeny had the eye well developed. The rapidity of his
results and the fact of periodicity in the expression of
the character makes any cumulative effect of selection
exceedingly questionable. One is not justified therefore
in accepting it as proof without corroboration.

CONCLUSION
Tn conclusion, it may be asked if it is not reasonable
to accept simply as a nomenclature the description of
the whole facts of inheritance in sexual reproduction
given by the Mendelian system? Is it wise to turn back-
ward and to give up this handy and helpful notation

654 THE AMERICAN NATURALIST [Vor XLVI

right in the midst of a useful career? The experiments
least open to objection (the pure-line experiments) have
shown the wisdom of assuming a stable unit factor, this
factor being representative of the stability manifested
by a character complex when no interfering conditions
intervene. Let us accept this simple interpretation pro-
visionally, appreciating the fact that the stability of the
characters that have been represented by fixed units may
be only a static appearance due to limited experiments;
but that this appearance justifies our neglecting any
infinitesimal fluency of our factor standards in experi-
ments of like duration, since taking them into account
would necessitate a change of standard, a new fabric of
hypotheses and a more complicated system. Let us take
a physiological view of heredity. Factors are assumed
to be stable. Characters are somewhat unstable owing
to the effect that other factors have upon their expres-
sion. Factor A, for example, is potentially able to pro-
duce a typical expression in ontogeny under certain defi-
nite conditions of environment, but the presence or ab-
sence of B or C or D or B, C, and D are responsible for
slight changes in the expression of A. This conception
gives us a picture of heredity in real accordance with
physiological facts, in contradistinction to the non-bio-
logical and fixed physical conception—the mosaic organ-
ism conception—that critics often say is held by some
geneticists.

One may answer that this conception is all right for
quantitative characters, but do the facts uphold it for
qualitative characters? They do. I will give examples
from my own experiments on the inheritance of the
purple aleurone cells in maize. Here one obtains prog-
eny by the thousands and sees phenomena that are ob-
secured by lesser numbers.

Crosses of the purple variety with three different
whites have given three different results. One shows
that the purple may be represented by the schematic de-
scription PPRRCC. Crossed with pp rr ce it gives
purples, reds and whites in the F, generation, as all three

No. 551] THE MENDELIAN NOTATION 655

factors are necessary for the production of the purple
color. How many other factors (present also in the
whites) may be necessary one can not say. In another
white, the R factor is present and purples and whites in
the ratio of 9:7 result. In another white, both P and R
are present. In another white, both P and C are present.
Both give monohybrid ratios when crossed with the
purple.

This is not the sum total of whites, however; several
others have been found. One has an intensifying factor.
We get darker purples together with the normal purples,
but no one can doubt that the purple is still the same pig-
ment modified in its expression. Another white has a
dominant inhibiting factor. In the heterozygous condi-
tion it does not always inhibit the color entirely, but in
the homozygous condition color never develops. The
dominance of this factor is proved by the fact that ex-
tracted colored recessives are still heterozygous for pres-
ence of color.

In still other whites I have. demonstrated the presence
of at least three modifying genes M,M,M,. They are
independent of each other, yet each and all affect the
purple color. One is dominant, as if it were a partial
inhibitor, the others are recessive, as if they were the loss
of intensifying factors. Purples of all different degrees
can be isolated and breed true. The lightest is such that
the color can be distinguished only with a lens. But they
are all strictly alternative in their transmission and
Somewhere near the expected ratios of darks, lights, very
lights, ete., appear. It is too much to ask that exact ra-
tios be obtained for with this kind of modification all
Shades appear, yet conclusive evidence has been ob-
tained by F, and F, generations. ee :

The qualitative characters do act the same as quanti-
tative characters, therefore, and one can not make a real
distinction between them.

A FIRST STUDY OF THE INFLUENCE OF THE
STARVATION OF THE ASCENDANTS UPON
THE CHARACTERISTICS OF THE DE-
SCENDANTS—IP?

DR. J. ARTHUR HARRIS

CARNEGIE INSTITUTION OF WASHINGTON

III. Presentation oF Data AND COMPARISON OF CoN-
STANTS FoR Navy, WHITE FLAGEOLET AND Ne PLus
LTRA Brans.—Continued

B. Number of Ovules per Pod.

The nature of this and the following character has been
discussed elsewhere.? There the data from which all the
physical constants necessary in this study may be de-
duced, but not the constants themselves, are set forth.
Tables IX-XI give these constants based on countings
of ovules formed and seeds matured in 130,074 pods.

That the starvation of the individual affects not merely
the number of pods which it produces but the character-
istics of these pods as well is evident from a study of
these tables, but is best brought out by a special kind of
graph, _

Reducing absolute to relative frequencies, we take the
difference

Starved less fed
for each ovule grade. Such differences are shown in
Diagram 8 for NDD-NDH, NDDC-NDHC, NHD-NHH,
NHDC-NHHC, USD-USH, USDC-USHC, FSD-FSH
and FSDC-FSHC. The differences for the ancestral
series, which for the moment alone interest us, are shown

*The first part of this paper appeared in this journal, Vol. 46, pp. 313-

343, 1912. The reader must consult it for all questions of purpose, ma-
terials, methods, ete.

* Harris, J. Arthur, ‘‘On the Relationship between Bilateral Asymmetry
and Fertility and Feeundity,’’ Roux’s Archiv. In press.

656

No. 551] INFLUENCE OF STARVATION 657

by the positions of the circles while those for their off-
spring grown upon the comparison field are represented

TABLE IX
Š “Mas pad Probable Standard Deviation Coefficient of Variation
Series and Probable Error and Probable Error
NH 5.7170 = .0106 .7762 = .0075 13.5763 = .1335
NHH 5.4126 = .0049 8569 + .0035 15.8323 + .0653
NHHH 5.3706 = .004 7669 = 0035 14.2803 = .0656
4.7192 + .0080 890 = .0057 18.8383 + .1245
NHDD 4.9985 + .0074 8064 + .0052 16.1313 = .1069
.0034 + .0120 6829 = .0085 17.0587 = .2181
D 4.3503 = .0138 .7695 = .0098 20.1638 + .2340
NDDD 4.7161 + .0163 7083 + .0080 15.0181 + .1630
H 5.1375 = .0070 7957 + .0049 15.4881 = .0980
NDHH 5.1692 + .0067 7057 = .0048 13.6517 = .0936
C 5.6607 + .0109 7852 = .0077 13.8708 = .1389
NHHHC 5.5853 + .0107 8099 = .0076 14.4999 81
Cc 5.6295 + .0137 8466 + .0097 15.0390 + .1762
NHDDC 5.6573 = .0104 .7371 = .0073 13.0289 = .1315
C 5.3701 = .0147 .7409 = .0104 13.7965 + .1969
NDDDC 5.5337 = .0127 .7397 = .0090 13.3674 = .1652
NDHC 5.4800 + .0 .7016 = .0073 12.8031 + .1362
NDHHC 5.4046 = .0096 .7133 = .0068 13.1973 = .1282
TABLE X
i ficient of Variation
Series | Mean ape Probablo | Siar otabieHerot_| ‘and Probable Error
USS 5.5279 = .0072 8694 + .0051 15.7280 + .0945
5. as 8196 + .0067 15.9178 = .1333
USHH 5.2392 + .0116 .7151 = .0082 6483 +
= .020 8705 = .0147 18.2481 + 3174
USDD 4.7991 + .0174 .7531 = .0123 15.6925 = 8
USC 5.5866 = .0093 .6 12.4908 + .1194
USSC 5.6992 = .0109 .6992 + .0077 12.2692 + .1367
USHC 5.5548 + .0125 oy ms 3.2290 =.
USHHC 5.5713 = .0117 .7622 + .0083 13.6816 = .1511
USDC 5.5287 = .0117 .7377 = .0083 13.3431 =+ .15
USDDC 5.5244 + .012 .7236 = .0086 3. = .1579
TABLE XI
of Variation
Beries Mean y Aeg paar sai edges ager yooh ee Pi oe
aridmemesroanesivd sorae E Lt PEATE LRTI
FSS 5 0080 1.0530 = .0057 18.5181 + .1033
5.5580 + .0077 .7639 = .0054 13.7443 + .099
FSHH 5.5150 = .0075 .6873 = .0053 12.4619 + .0974
4.9579 = .0142 8022 = .0101 16.1798 + .2080
FSDD 5.0193 = .0116 .6799 = .008 Sane
60 = .0097 7728 12.6768 + .1145
FSSC 6.1840 + .0106 7816.+ .0075 6390 +.
FSHC 6.0761 = .0121 '8273 = .0086 13.6149 = .1437
FSHHC 6.0245 = . 8173 = .0069 13.5660 + .1168
FSDC 6.0525 = 0137. | .7853 = .00 _ 12.9756 = 1621
—FSDDC | 6.0616 + .0107 18144 = 0076 | 184355 = 1208 __

658 THE AMERICAN NATURALIST [Vor. XLVI

+20
HOP
ojro as TCRA os
—10}- L
NDD—NDH NHD—NHH
AND AND
—29- NDDC—NDHC | NHDC—NHHC
+nob :
+10 :
—l0 i
USD -USH FSD -FSH
—2ol USDC — USHC | FSDC—FSHC
L 1 L L i L L La L L L | i 1 ae FOO

DIAGRAM 8. Differences in the percentage frequencies of various numbers of
ovules per pod in luxuriant and depauperate cultures and in their offspring.
by solid dots. A profound influence of the starvation
conditions is evident from these graphs. Relatively, the
lower grades are much in excess, the higher grades much

in defect in the starvation series.

The same fact is quite patent when one deals with the
means of the series instead of comparing individual
classes in the same lot. Clearly from Diagram 9:

No. 551] INFLUENCE OF STARVATION 659

SCALE OF MAGNITUDE FOR SCALE OF MAGNITUDE FOR
ANCESTRAL SERIES COMPARISON SERIES

so 475 3.00 525 3.50 3.75

-
S
=
-
K
a
~

= d
=
=
oO?
s

STRAIN AND SERIES OF PLANTS

USDD

FLAGEOLET

FSD ie
FSDD

Diacram 9. Mean number of ovules per pod. „Compare explanation of diagram 6.

(a) The means are in every case conspicuously lower
for the starved than for the fed ancestral series.

(b) The means for the comparison series are closely
Similar: there is no striking superiority of fed over
Starved ancestry.

Here, accordingly, as for number of pods per plant,
we must have recourse to numerical differences and their
Probable errors. The intra-ramal comparisons are made
in Table XII. In three of the four cases, the starved
seeds produce pods with more ovules, but in all ae
Seed was a year older for one generation of ascendan

660

THE AMERICAN NATURALIST

TABLE XII

[Vou. XLVI

Description of Material

Ancestors Starved for
Two Generations

Ancestors Starved for
Three Generations

Ancestors starved for one generation:
USDC series:

SRE OB e be Oe Eee A E Ole ee ak ee

USDDC series:

—2'0101 + 2198

NDDDC series:

= 4291 æ .2571

starvation than for two.

Compare the results for num-

ber of pods per plant, and for seeds per pod.

The inter-ramal differences appear in Tables XII-
XVI. Of the 28 comparisons, direct and cross, of mean
number of ovules, 18 are negative and 10 positive; that
is, in 18 cases, the plants with starved ancestry have a
lower number of ovules per pod. Thus the deviation

TABLE XIII

Description of Material
.

Ancestors Starved for
One ‘oan

Ancestors Starved for
Two Generations
NHDDC

TSE E ee
Ancestors well fed for one generation:
N. ey ries

eS 8 Se EN a E A ee ga E E E gti
ROW ORR 8 6 6s oy

eh oe te ee ee ee

EE ORES EO aia MA g CR rer Ge WE Te

HON E D
Coefficient = ea aS
NDHHC seri
M

2 SOS a NE EO Oe ee E ick OS

th we a eta We) ae ee ae

oe 06 heh 8-8 88 E S

Cee ee ae

CS ee eg Rw i gs E eg a Ty

"eee eee eees

+ .1495 = .0172
1450 = .0121

42.2359 + .2297

— .0312 +,

H
©
o
ns

+ .2249 =,

2 = .0174
0

+ .1773 = .0147
4 5 = .0103
+ .2258 = .1892

— .0034 = .0151
— .0481 + .0106
— .8419 + .1913

+ .2527 + .0141
+ .0238 = .0100
— .1684 =

+ .0720 = .0149
28 = Ol
= 1.4710 = -1908

No. 551] INFLUENCE OF STARVATION 661

from the equality of division, if there were no influence
of the environment of ascendants, is 4+1.79. Of the 28
differences, 16 are thrice their probable errors; 9 are

significantly negative and 7 significantly positive.
Taking averages, regarding signs, we have:

— .0819

Navy, Within Strains, A = + .0150
A

Ne Plus Ultra,

White Flageolet, A = — .0378
TABLE XIV
Ancestors Starved for | Ancestors Starved for
Description of Material nerations Three Generations
DDO NDDDC
Ancestors well fed for one generation:
NDHC series:
“a AEE ARER cee re eS — .1099 = .0180 + .0. .0164
tandard deviat 7 on oak Puen + .0393 = .0127 + .0381 + .0116
rae Of variston... os. + .9934 = .2394 + .5643 = .2140
ees leh vg for es petals es
ISO Rap Rea geome WE Er ee ott as — .2906 = .0183 — .1270 = .0167
Piandard OES e WA of — .0443 = .0129 — ,0455 = .0118
aent = ais a ee a — .0743 + .2410 — .5034 = .2159
NDHHC se
ick A Ey IE ge cre NR et a oN — .0345 = .0175 + .1291 = .0159
Standard deviation.............. + .0276 = .0124 + .0264 = .0113
Coefficient of moar signee E + .5992 + .2349 -+ A701 = 2000 |
Ancestors well fed for three generations
NHH.
Bo A ec es — .2152 = .0182 — .0516 + .0166
andard deviation.............. — .0690 + .0129 = 0702 © 011s
Coefficient of variation.......... — .7034 = .2404 | —1.1325 + .2154 _
TABLE XV
x ES
ved Ancestors Starved
Description of Material D oy Pictish for Two Generations
USDC USDDC

LOE i ate ae
Ancestors well fed for one generation:
USSC series:

WHS iii ivy ee — .1705 = .0160 | — .1748 = .0163
Standard deviation.............. + .0385 = .0113 0244 + .0115
oleae of nena a ia +1.0739 = .2047 | + .8297 = .2088

a we ee ee wl Be were

— 0261 + .0171
+ .0029 + .0121
+ 1141 = .2223

— 0112 + .0123
— 1301 =

USHHC se
MM ee — .0426 = .0166 | — .0469 = .0168
Btandard duviation 222.22. — [9245 = 0117 | — .0386 = .0120
2. Cosficient of ¢ annie ee ae — 3385 = .2145 | — 5827 + .2186

662 THE AMERICAN NATURALIST [ Vou. XLVI

TABLE XVI
Ancestors Starved Ancestors Starved
Description of Material for One Generation for Two Generations
FSDC FSDDC
Ancestors well fed for one generation:

C series:

Ch EN P ee ae ae — .1315 = .0173 — .1224 + .0151
Standard deviation..........3.... + .0037 + .0122 + .0328 + .0107
nr acral of soot un penn + .3366 + .2037 + .7965 1769

aes

oe POG so nee — .0236 = .0183 — .0145 + .0162
Standard deviation eroana eaS — .0420 0130 — .0129 = .0115
Coefficient of variation.......... — .6393 + .2166 — .1794 + .1916
Ancestors at fed for two ‘anata:

parte series:
bE RSE loegepsi eure lye a ee ON ie + .0280 + .0169 + -0371 = .0145
AEE déviation; 04.54.44 a — .0320 = .0119 0029 = .0102
Coefficient of variation.......... 5904 7 1305 = .1723

SCALE OF MAGNITUDE FOR : SCALE OF MAGNITUDE F
ANCESTRAL SERIES COMPARISON pase
250 200 350 400 ino noo 3.50 350 wo

NAVY ee

N

NHH b

NHD' bera

NHDD' Hien

NAVY
ND E]
NDD.
D oe
eet a
= NDH T
a NDHH Ò
©
un ULTRA
oy
æ us
A uss Q
a mt ve
USH > J
fi
z USHH -=O
= usD el
a usDD ao th
FLAGEOLET

So pe L -

FSS oo fice ei a

eel

UH remna p= TO Dap S

FSD eoe. F ----} 8

FSDD Ei a EA Re

DIAGRAM 10. Mean number of seeds per pod. Compare explanation of diagram 6.

No. 551] INFLUENCE OF STARVATION 663

The results for Navy are slightly positive, for the other
two varieties more conspicuously negative. The mean
of the four varieties is — .0291.

Regarding only the 10 direct inter-ramal comparisons,
we note that 7 are negative and 3 positive; 3 significantly
negative and none significantly positive. The mean of
the negative differences is —.0771, of the positive
+ .0794, of all — .0316.

The data are available for any one caring to work out
the relationships for variabilities. The discussion of this
point is reserved until further series are gotten.

C. Number of Seeds per Pod

Tables XVII-XIX give the essential biometric con-
stants for number of seeds per pod. Diagram 10 justi-
fies the same general conclusions for the mean number
of seeds as were drawn from Diagram 9 for mean num-
ber of ovules per pod. In one case, however, the average
for seeds is lower on a feeding plot than on the starva-
tion fields.®

Appealing again to constants and their probable
errors, we have the results set forth in the tables of fun-
damental differences, XX—XXIV.

The intra-ramal comparisons, Table XX, show three
positive and one negative difference. Two of the positive
differences are probably and the third possibly statis-
tically significant. Note, however, that the age of the
seed may be a disturbing factor. Compare the results
for number of pods and number of ovules.

Of the 28 inter-ramal comparisons, Tables XXLXXIV,
15 are negative and 13 positive.

In the usual manner, we get for the means:

Navy, Within Strains, A = + .0078
Ne Plus Ultra, A= 0711
White Flageolet, = — .0383

* Why FSS plants did not mature their seeds well I have never been able
to make out. The fact was noticed at harvest time.

664 THE AMERICAN NATURALIST [ Vou. XLVI

For Navy the positive difference is trivial; the negative
difference for Ne Plus Ultra and for White Flageolet is
much larger.

Considering statistical significance to be indicated by
a difference thrice its probable error, we find that 8 are
significantly negative and 4 significantly positive. These
4, and 2 of the 8 significantly negative differences, fall in
the comparisons between the two strains of Navy and
hence can not be given much weight. Of the 20 inter-
ramal comparisons, direct and cross, within the same
strain, 11 are negative and 9 are positive in sign. Of the
10 direct comparisons 5 are positive and 5 negative in

TABLE XVII

Berlen Mean and Probable Standard Deviation Coefficient of Variation
Error and Probable Error and Probable Error
NH 4.2555 = .0193 1.4113 = .0136 33.1634 = .3538
NHH 4.2672 = .0073 1.2893 = .0052 30.2134 = .1323
NHHH 4.277 0082 1.2802 + .0058 29.9268 = .1463

D 3.1269 = .0117 1.2964 + 41.4611 = .306
NHDD 3.7493 = .0116 1.2672 + .0082 33.7974 = .2420
D 3. = .0196 1.1143 + .0139 35.9804 + .5017
DD 3.1872 + .0189 1.1984 = .0134 37.6013 = .4752
NDDD 3.5823 = .0176 1.1709 = .0124 32.6870 = .3823
DH 4.17 0108 1.2363 = .0076 29.5786 = .1982
NDHH 4.2512 = .0112 1.1774 = .0079 27.6963 + .2003
4.0115 + .0194 1.3959 = .0137 34.7973 + .3812
NHHHC 3.8979 + .0181 1.3683 = .0128 35.1028 + .3656
3.9417 + .0221 1.3616 + .0156 34. +
NHDDC 3.9333 = .0191 1.3585 = .0135 34.5397 = .3816
DC 3.7550 + .0254 1.2798 = .0179 34.0838 + .5301
NDDDC 3.8191 = .0227 1.3215 = .0161 34.6033 + .4679
.7333 = .0192 1.2989 + .0136 34.7927 + .4058
NDHHC |_ 3.7751 + .0173 1.2835 + .0123 34.0005 =
TABLE XVIII

Series Mean and Probable Standard Deviation ew ye of Variation
Error and Probable Error d Probable Error
USS 3.8781 = .0114 1.3760 = .0081 35.4817 + .2330
4.0097 = .0146 1.2599 = .0103 31.4219 + .2810
USHH 3.8698 = .0214 1.3253 + .0151 34.2475 + .4347
2.6322 + .0291 2235 = 6 46.4812 = .9367
USDD 3.2268 + .0297 1.2823 = .0210 39.7382 = .7453
4.1269 = .0175 1.3177 = .0124 31.9296 + .3297
USSC 4.2113 = .0204 1.3126 + .0144 31.1673 + .3739
USHC 4.1045 = .0226 1.3280 + .0160 2.3553 + .4283
USHHC 4.0832 =. 1.3333 + .0145 32.6536 = .3899
3.9718 = .0213 1.3403 = .0150 33.7456 + .4192

__USDDC | 4.1619 + .0214 1.2752 = .0151 30.7127 + .3969 __

No. 551] INFLUENCE OF STARVATION 665

TABLE XIX
Mean and Probable Standard Deviation Coefficient of td scree
Series Error and Probable Error and Probable Err.
FSS 3.2918 = .0113 1.4772 = .0080 44.8741 + .2868
H 4.3907. = .0 1.3078 = .0093 29.7865 = .2288
FSHH 4.4563 = .0143 1.3095 = .0101 29.3856 = .2450
D 3.4624 = .0230 1.2993 = .0163 37.5263 + 5323
FSDD 3.8548 = .0021 1.1991 = .0145 31.1063 = .4109
C 4.2865 = .0172 1.3677 = .0122 31.9072 = .3113
FSSC 4.3643 = .0198 1.4532 = .0140 33.2987 = .3541
HC 4.1941 + .0200 1.3610 = .0141 2.4494 + 3701
FSHHC 4.1997 + .0167 1.3941 = .0118 .1948 + .3102
4.1574 + .0236 1.3596 = .0167 32.7024 + «4428
FSDDC 4.2710 = .0178 1.3537. = 0126 31.6950 = .3220
TABLE XxX
SS ` Ancestors Starved for | Ancestors atric ok ag)
Description of Material Generations Three Generatio
Ancestors a for one generation: E
USDC serie: USDD6 series:
ORs ee A Gt Care es oe OS + .1801 + .0302
Standard deviation.............. — .0651 + .0213
anpassa of aeei EE E —3.0329 = .5773
FSDC ser FSDDC series:
DN a a a + .1136 + .0296
Standard deviation.............. — .0059 = .0209
Sop ee of meini, Re a —1.0074 = .5475
HDC se .HDDGE series:
We a os — .0084 = .0292
Biandari d deviation... oc u — .0031 = .0206
fficient of variation.......... — .0047 = .5828
Ancestors era for two generations: :
DDDC series:
i Be ee e a eee + .0641 + .0341
Standard deviation.............. + .0417 + .0241
Coefficient of variation.......... | +.5195 + .7070 _

sign, but none of the positive differences are statistically
Significant, while 3 of the negative differences are from
4 to 8 times their probable errors. The mean of the 5
negative differences is — .1371, of the 5 positive Mier.
ences + .0482, of the whole series — .0445.
D. Weight of Seed

hort the experiments partially described in this
Paper, attention has been given to the weight of the indi-
Vidual seed. From the practical standpoint, the total
weight of the seeds produced by the plant would have

n a more desirable determination, but for several
reasons this was not feasible.

666

THE AMERICAN NATURALIST

[Vou. XLVI

For all the ancestral series, the seeds were weighed
individually in units of .025 gram, but the excessive
labor involved precluded this for the twenty comparison

crops.

Instead, the seeds of each series were mixed

thoroughly among themselves and random drawings

TABLE XXI
neestors Starved Ancestors Starved
Description of Material for One Generation or Two Generations
` NHDC NHDDC
Ancestors well Reg for one generation:
a ie :
RUE iaey he Oa ewe kee + .2084 + .0293 .2000 + .0270
Standard deviation. E eee a + .0627 = .0207 + .0596 = .0192
ont or VATIBUION «sci ccs sis — .2483 = .5989 — .2530 + .5570
isolate well fed for tea pR
E E O rs tie — .0698 = .0294 — .0782 = .0272
Standard deviation 5.2 o1 — .0343 = .0208 — .0374 + .0192
one a i Warintion: 622... — .2529 + .5826 — .2594 + .5394
gr hy
Coy Gime eres aa as + .1666 + .0281 + .1582 0258
Standard aena Rees eee yA + .0781 = .0199 + .0750 + .0183
ib Of Variation «063555 o. + .5439 = .5691 + .5392 + .5248
ies ee fed for si generations
HC series:
BANG. reat se ck Ot -+ .0438 = .0286 + .0354 = .0263
tandard déviation... ..<:.. — .0067 = .0202 — .0098 = .0186
Coefficient of A r E eg — .5584 = .5725 5631 + .5285
TABLE XXII
Ancestors Starved for | Ancestors Starved for
Description of Material Two Generations Three Generations
NDDC NDDDC
Ancestors well fed for one generation:
ND saps ies ë
E T a oak + .0217 = .0318 + .0858 = .0297
Sta aba MSVIALION coo — .0191 + .0225 ti .0226 = .0211
Coefficient of variation.......... -7089 76 1894 + .6194
Ancestors = Sn for ie a generations: e
Mo o — .0256 = .0320 — .1924 + .0299
Standard GOT Cs 1161 = .0225 — .0744 = .0211
cone sa ai BE eee — .7135 = .6529 — .1940 = .6035
ND tami C se
a R a — .0201 = . +. =
or Govindan., ae ae — .0037 + .0217 | + .0380 + .0203
Coefficient of variation.......... + .0833 = + .6028 = .5905
Ancestors well f fed fo or ‘hres generations:
NHHH
Mit a = i =, = = .0290
Standard deviation.............. =+ pobre — .0468 + .0206
Coefficient of variation.......... — .0190 + .6440 4995 = .5938

No. 551] INFLUENCE OF STARVATION 667

TABLE XXIII

Ancestors Starved for | Ancestors Starved for
Description of Material One Ber ages Two 0 Geasrations

Ancestors tie fed for one generation:
U 2” rie

Peder ei E E ew yi — .2395 + .0295 — .0594 + .0296
kadad Geviation 2. E + .0277 = .0208 — .0374 + .0209
Spa o yaristion: i sarr o -+2.5783 + .5617 — 4546 + .5453

tesla

Pa Bites E ene ee ee — .1327 + .0310 +. 0474 + .0311
Standard GeViauon. 62.5035 2a + .0123 + .0219 — .0528 = .0220
Coefficient of variation.......... +1.3903 = .5993 —1.6426 + .5840

Ancestors wel agg or two EA DES
US.
Sen: er i es ee ee ee — .1114 = .0295 + .0687 + .0296
tandard deviation.............. + .0070 = .0209 — .0581 + .0209
Coefficient of variation.......... +1.0920 + .5725 +1.9409 + .5564

TABLE XXIV

Ancestors Starved for | Ancestors Starved for
Description of Material One ph ag Two A
Ancestors well fed for one generation:
FSSC series: ‘ :

RR Bae ek eee un ee — .2069 + .0308 — .0933 = .0266
Standard deviation.............. — .0936 + .0218 — 0995 = .0188
“Symone of variation: : fo i... — .5963 + .5670 — 1.6037 + .4786

FSHC se

nin r a a — .0367 = .0309 | + .0769 = ppan
Standard deviation.............. — .0014 = .0219 — .0073 = .0189
Coefficient of variation.......... + .2530 = .5771 7544 + 4906

Ancestors well fed for two generations:
HHC series:

Reo i hore oh Gh ees ware — .0423 + .0289 + .0713 + .0244
Standard deviation.............. — .0345 + .0204 | — .0404 = .0173
Coefficient of vacation ee eae — A924 + .5406 —1.4998 + .4471

made for mass weighings.t It is on these samples of
2,000 or more seeds (weighed after drying for several
months)* that the averages are based.

Seed weight will be touched rather lightly in this paper.
This is in part due to the fact that the seeds for the com-
Parison series could not be weighed individually, thus

* This was the plan for all but USC and FSC, where 100 seeds, or as

many as were available, were weighed en masse for each line. From these,
the general population mean was calculated. Li
*The plants were harvested and dried in an empty greenhouse in $
early autumn of 1910; stored in an unheated building for the early zak 9
the winter; counted rs uring a period extending from January to
April; and left in the laboratory till weighing, some time in June.

668 THE AMERICAN NATURALIST [Vou. XLVI

affording data from which the variabilities and probable
errors might be calculated. It is in part due to the fact
that (as will be clear later) the starvation conditions
available seem not to have affected seed weight as they
did the other characters with which we deal.

SCALE OF MAGNITUDE FOR SCALE OF MAGNITUDE FOR
100 aso 200 250 3300 35o _.100 JA75 223 27S es
NAVY
NH Q.
NHH ->
NHHH J
NHD a
NHDD o
NAVY
D a
n NDD ---}--@
P A NDAD
< WUDU
<
a NDH +
S NDHH 2 ee ee ee oe Seen eer es lh a
A ULTRA
n USS Ly EN ae 3
a a
z USH -----+-;-
z USHH EE <a
= La AI
x usp #-4----- ----}--- -iS
n - --
FLAGEOLET
FSS ro es --+-------@
FSH a >
FSHH Ł - A
FSD F- e
r D Fspp a
Ww ee ee eee a

DiacramM 11. Mean weight of seeds. Compare diagrams 6, 9 and 10.

The mean weight for both ancestral and comparison
series are shown in decimals of grams in Table XXV.

That seed weight is a character influenced by environ-
mental conditions is apparent from the scatter of the
means in the ancestral field of Diagram 11 as compared
with their closeness to line in the comparison panel. But

No. 551] INFLUENCE OF STARVATION 669

the diagram also shows at once that* seed weight has not
been influenced by the starvation conditions as other
characters have. Whereas, with a single exception, num-
ber of pods per plant, number of ovules per pod and num-
ber of seeds per pod were all conspicuously reduced, seed
weight is sometimes higher, sometimes lower, on the
starvation plots. Of the 28 differences, 21 are negative
in sign, as compared with 28 for pods per plant and
ovules per pod and 27 for seeds per pod.

Expressing numerically the same differences for the
ancestral series as are usually taken for the comparison
series, we find for comparisons within the strains:

TABLE XXV
Ancestral Series Comparison Series
Series N Mean Series N Mean
F 3,000 19322 FS 7,562 18587
FSS 3,740 15636 FSSC 2,000 18026
2,122 20039 FS ,000 17893
FSHH 1,788 16813 FSHHC 2,000 18145
1,989 16140 FSD 2,000 17816
FSDD 1,643 16783 FSDDC ,000 19027
U 2,391 34491 US 9,879 28450
USS 3,271 35349 USSC 2,000 29933
1,165 34261 USHC 2,000 314
USHH 31642 USHHC 2,000 30147
1,002 25474 USDC 2,000 29170
USDD 29948 USDDC 2,000 29284
7,334 23186 H 2,000 25430
NHHH 5,601 20274 NHHHC 2; 244

6,630 20073 NH. 2,000 26616
NHDD 5,029 22293 NHDDC 2,000 24669
2,362 NDDC 2,000 25849
NDDD 1,946 .20374 NDDDC 2,000 (25207
NDH 3,227 ‘20879 NDHC 2,000 "26248
NDHH 2,433 .20261 NDHHC 2,000 24815

Navy, A = — .0047

Ne Plus Ultra, A= — .0604

White Flageolet, A= — -0103

General Average, A =— 0231

Thus there appears to be a distinct influence of the de-
Ppauperization of the individual upon the weight of the
seeds which it produces. But modification of weight is
very slight indeed as compared with that of the other
characters. :

670 THE AMERICAN NATURALIST [ Vou. XLVI

Considering now the weight of the seeds produced by
the comparison plants, we note that of the 28 differences
which may be taken (within and between strains) be-
tween plants of luxuriant and those of depauperate an-
cestry, 15 are negative and 13 are positive in sign, a devi-
ation of only 1+ 1.79 from the expected 14:14 ratio.

Of the individual differences, the largest is .023 gram,
while most of them fall, towards zero. Averaging we
get:

Positive Differences, + .00915 gr.
Negative Differences, — .00831 gr.
All Differences, — .00020 gr.

Surely values as low as this can not give much weight to
the assertion that depauperization of the parents has
had any influence upon the weight of the seed of the off-
spring plants.

Looked at in a preliminary and superficial way (and
it hardly seems worth while to go into the matter more
minutely until other data are tabled and reduced), the
data seem to indicate that the weight of the seed is a
character much less directly dependent upon cultural
conditions than are the vegetative characters of the
plant. Conditions which reduce these latter may not
materially affect seed weight.

Possibly, the environmental complexes available were
such as to affect certain characteristics of the ancestral
plants, while leaving others, i. e., seed weight, unmodi-
fied. Possibly, seed weight is a character little affected
by external conditions of any kind. These are questions
which can only be solved by further experiments de-
signed to determine whether some environmental com-
plexes regularly affect seed weight while others do not,
and to ascertain what influence, if any, such reduction
has upon the characteristics of offspring seeds.

E. Combination Characters
: Some characteristics are combinations of two or more
individual measurements. Such are, for example, the

No. 551] INFLUENCE OF STARVATION 671

correlation between two dimensions, or the ratio of the
one to the other.

The only case of this kind to be considered here is the.
coefficient of fecundity, which is simply the ratio of the
total seeds matured by a population of pods to the total
ovules formed. The values are given in Table XXVI.

For the ancestral series all but 2 of the 20 compari-
sons within strains show lower fecundity in the starved
series. Taking averages:

Navy, Within Strains, A —— .0790
Ne Plus Ultra, A=—.1178
White Flageolet, A= + .0075

Thus Navy and Ne Plus Ultra mature about 7-11 per
cent. more of their ovules on feeding than on starvation
fields. Apparently, White Flageolet is not affected.

TABLE XXVI

Ancestral Series Comparison Series

Series Pods C.F. Series Pods C.F.
FS — osha FSC 2,876 7387
FSS 7,809 5789 FSSC 2,457 7057
4,541 7899 FSH 2,117 6903
FSHH 3,837 8080 FSHHC 3,180 6971

FSD ; 984 FSD 1,506 86
FSDD 1,556 7680 FSDDC 2,646 .7046
U. —— Sc 2.569 "7032
USS 7015 USSC 1,888 7389
3,406 7787 ‘SH 1,570 7389
USHH 1,743 7386 USHHC 1,936 7329
USD 802 5518 US 1,810 7184
USDD 851 6724 USDDC 1,619 7516
; 78 NHHC 355 7087
NHHH | 11,230 7965 NHHHC 2,614 6979
5,581 6625 1,733 7002
NHDD 5,449 7501 NHDDC 2,308 6953
1,827 .7326 1,159 .6992
NDDD 2,018 7596 NDDDC 1,542 6901
5,955 8136 2,077 6813
NDHH 5,019 (8224 NDHHC 2,494 .6985

For the comparison series grown from these seeds, we
find that of the 20 comparisons, direct and cross, within
the strains, 12 show lower and 8 show higher fecundity
in the offspring of starved plants. The means are:

672 THE AMERICAN NATURALIST [Vou. XLVI

Navy, A = —- .00036
Ne. Plus Ultra, A= — .00615
White Flageolet, A = — .00196

Discussion of such differences is obviously superfluous.

IV. Recaprrunation, Discussion anD TENTATIVE Con-
CLUSIONS

The purpose of the series of investigations, described
in part, is to ascertain whether the depauperization
of the individual through the environmental com-
plexes constituting ‘‘poor’’ agricultural conditions, in-
fluences the characteristics of its offspring, and if so,
how much. The problem of the chemical and physical
‘‘causes”’ of the depauperization has received the most
intensive experimental consideration. The question of
the influence of the surroundings of the ascendants upon
the characteristics of the descendants has been much
more a matter of speculation. Yet the second of these
problems is of obvious importance to the agriculturist
and of interest to the evolutionist concerned with en-
vironmental factors. The time seems ripe, therefore, for
its consideration on the basis of extensive quantitative
experimental data.

Notwithstanding the great progress which has been
made in the investigation of the relationship of the
chemical and physical properties of the substratum to
the characteristics of the plant, the diversity of results
and the clash of theories show that we have only entered
the edge of this field of research. In consideration of
these facts, and especially in view of the all but unsur-
mountable difficulties of controlling in large experiments
the conditions of growth of flowering plants, it has
Seemed necessary in first studies to choose merely good
and bad growing conditions as indicated by yields in
actual cultures. Thus the methods are avowedly and in-
tentionally of the rough and ready sort. If in such ex-
periments an unquestionable influence of the conditions

No. 551] INFLUENCE OF STARVATION 673

of growth of the ascendants upon the characteristics of
the descendants be demonstrated, it will be worth while
to determine the weight to be given to individual physical
and chemical factors in the ascendant environment. If,
on the other hand, there be no detectable effect of ances-
tral depauperization, then the cost of elaborate batter-
ies of experiments had better be devoted to some other
problem.

This first study is based upon three varieties of one

species, Phaseolus vulgaris. The conclusions should not,

therefore, be extended to other forms with different de-
mands upon the soil, habits of growth or type of seed.
This is true not merely on general principles, but is espe-
cially important because of the well-known capacity of
this species for growth under adverse conditions.

The characters considered are number of pods per
_plant, number of ovules formed and number of seeds ma-
_ tured per pod, ratio of total seeds ripened to total ovules
daid down—the coefficient of fecundity—and weight of
eds,

‘astants are based upon the countings of number of
ies and seeds in about 130,000 pods and weighings of
er 110,000 carefully selected seeds. But these obser-
lions were drawn from only 21,000 individual plants.
he results in the body of the paper show, these num-
rs are too small rather than unnecessarily large for
roblem of this delicacy. I believe, however, that they
e sufficiently large to bring the probable errors of
‘random sampling low enough that dangers of erroneous
‘onclusions lie rather in the inevitable experimental
and to a less extent observational) errors.

_ Bearing in mind the difficulties to be surmounted and
the consequent possibilities of error, we draw the fol-
lowing tentative conclusions.

Environmental conditions which greatly reduce num-
ber of pods per plant, number of ovules formed per pod
and number of seeds matured per pod, affect to a less
degree the relative number of seeds matured, i. e., the

674 THE AMERICAN NATURALIST [Vou. XLVI

coefficient of fecundity, and have but little effect upon
seed weight.

_ The influence of the modification of the ascendants
upon the characteristics of the descendants is extremely
slight. There seems, nevertheless, to be a definite reduc-
tion in the number of pods per plant and number of
ovules per pod. There is also a possible lowering of the
absolute and relative number of seeds per pod. Ap-
parently, there is no modification of seed weight.

Cop SPRING HARBOR, N. Y.

STRUCTURAL RELATIONS IN XENOPARASITISM

W. A. CANNON

DESERT BOTANICAL LABORATORY

Ar various times normally independent plants have
been experimentally caused to grow and develop within
the tissues of other independent plants, deriving from
this arrangement food and food-materials and organizing
tissues and organs.! Although in themselves short-lived,
the artificial parasites offer interesting suggestions as to
the possible conditions under which true parasitism may
arise in nature.? Itis clear, for instance, that the mutual
relation of parasite and host is extremely complex, both
from a purely physiological point of view and from a
structural one. On the one hand, it presupposes suitable
osmotic relations and not unfavorable chemical reactions,
and on the other, among other things, the fitting and exact
adjustment of the tissues of the parasite, and it signifies
atrophies as well.

hen we observe the leading structural changes which
normally occur in the growth of a haustorium of a habi-
tual parasite, such, for example, as the mistletoe,’ we find
a course of development which is full of suggestions. A —
young haustorium is composed mainly of undifferentiated
ground tissue, but there are the beginnings of conductive
tissue within, and a protective epidermis without. Upon
the commencement of the parasitic relation the most
marked changes occur. In the first place epithelial cells

*“* Artificial Parasitism, ete.,’’? G. J. Peirce, Bot. Gaz., 38: 214, 1904.
‘‘The Condition of Parasitism in Plants,’’ D. T. MacDougal and W. A.
Cannon, Publ. No. 129 Carnegie Inst. of Wash., 1910. ‘‘An Attempted
Analysis of Parasitism,’’ D. T. MacDougal, Bot. Gaz., 52: 249, 1911.

***An Attempted Analysis of Parasitism,’’ D. T. MacDougal, Bot. Gaz.,
52: 249, 1911.

‘The Anatomy of Phoradendron villosum,’’? W. A. Cannon, Bull. Torr.
Bot. Club, 1901.

675

676 THE AMERICAN NATURALIST [ Vou. XLVI

are formed directly from parenchyma, and then after
penetrating the host, such of the periphery of the haus-
torium as touches non-living cortical host cells, organizes
cork. Finally, upon the attainment by the haustorium of
the woody cylinder the conductive tissue of the hausto-
rium opposes cell for cell the conductive tissue of the host,
and in such parasites as possess sieve-tubes, the sieve-
tubes hold a similar relation to the sieve-tubes of the
host. It happens therefore in habitual parasites that a
portion of the development of the haustoria occurs after
the parasitic relation has been entered into, so that the
direction of the development of much of the tissue of the
haustorium is fortuitous, depending in part on the posi-
tion occupied by the tissues of the host.

DURATION oF THE XENOPARASITIC RELATION

Although induced parasitism means naturally a limited
period during which the artificial relation can be con-
tinued, this period varies greatly with the different nutri-
tive couples. A review of this phase of the subject will
not be given here, as it is completely presented in the
papers referred to above, but two or three of the most
pertinent parasitic relations will be cited. Peirce grew
Pisum sativum on Vicia Faba to maturity (Peirce, ‘‘ Arti-
ficial Parasitism,” l. c.). MacDougal (see above) records
many experiments of which the following may be given:
Cissus laciniata was grown on Opuntia blakeana from
February 1, 1908, until April 19, 1909, and another cul-
ture, which is especially treated in this paper, lasted from
early autumn, 1911, to June 10, 1912. In the instances
where Cissus was employed roots were freely formed, the
stem attained considerable length and organized tendrils
and leaves. From these facts a large capacity for adjust-
ment on the part of the induced parasites is exhibited,
and also a degree of physiological adaptability is shown
which reveals something of the plasticity of such plants
and argues a fair suitability for the dependent relation.

: “On the Structure of the Haustoria of Some Phaneorgamic Parasites, ’’
G. J. Peirce, Ann. Bot., 7: 324, 1893.

No. 551] RELATIONS IN XENOPARASITISM 677

XENOPARASITISM OF Cissus LACINIATA

The induced parasite Cissus laciniata exhibits in the
structure and form of its roots (the shoot was not studied)
certain deviations from the normal which are of signifi-
cance andinterest. A history of the experiments in which
this species was used as a parasite is given in another
place, suffice it to state here that a cutting of the Mexican
grape (Cissus laciniata) was introduced into the tissues of
Opuntia blakeana and allowed to remain several months.
A shoot with leaves and tendrils was formed. After the
culture had been running some time a root of the grape
was seen to emerge from the surface of the cactus, to
grow downwards, and to penetrate the soil. It was sev-
ered so that the Cissus had connections with the cactus
only. On June 10, 1912, the newly organized leaves were
‘seen to be relatively small and the tendrils not to develop.
The culture was thereupon taken down and the roots of
the parasite dissected out so that their relations to the
host tissue might be learned.

All of the roots of Cissus which were situated within
the tissues of the cactus were found to be fleshy. A main
root was traced from the base of the cutting through the
tissues of the cactus for a distance of 3 em. when, as
above mentioned, it issued from the cactus and found its
way into the soil. This root gave off one branch about
1 cm. from its point of origin, which extended for a dis-
tance of 3 cm. into the tissues of the cactus. The root last
mentioned gave rise in turn to a branch which attained a
length of 1.5 cm. In addition to these roots there were
several short ones which reached little belond the surface
of the parent root. All roots except the one especially
mentioned as not behaving in this manner were wholly
enclosed within tissues of the host.

STRUCTURE or FREE-LIVING Roots :
The portion of the roots which are free-living offer
useful points of comparison, for which purpose. the anat-
omy will be briefly reviewed.

678 THE AMERICAN NATURALIST [ Vou. XLVI

A root 2.0 mm. in diameter shows the usual divisions
into central cylinder and cortex. The endodermis is
well marked. The epidermis is discolored and bears the
remains of root-hairs. Cork has not begun to form, how-
ever. The cortex is composed of cubical parenchyma;
the parenchyma of the central cylinder offers no un-
usual features. Little starch or crystals are to be seen.

xin

ui
ih Nab

ANATOMICAL RELATION OF THE Cissus- Opuntia COUPLE. On the left appears
the extra-cortical portion of the root with the limit indicated by the arrow.
On the right is the wound tissue of the cactus, and between this and the root
lies disorganized cactus cells.

STRUCTURE OF THE Parasitic Roots

The roots of Cissus, which developed within the tissues
of the cactus, varied in diameter from 2 to 5 mm. and
showed characteristics which were in certain regards
quite different from those of the free- living roots ex-
amined.

If a cross-section of a root 2 mm. in diameter is studied
the usual differentiation into cortex and central cylinder
will be noted. The cortex is composed of relatively large
cells a few of which contain stellate crystal aggregates
and raphides. A layer of cork, over half dozen cells in
thickness, bounded by the dead remains of the epidermis,

No. 551] RELATIONS IN XENOPARASITISM 679

lies on its periphery. The remains of root-hairs were
looked for but were not found. A well defined endo-
dermis with granular contents, a portion of which is
starch, limits the cortex on its inner surface. The cen-
tral cylinder has relatively wide medullary rays and a
large pith containing much starch. Opposite each

CROSS-SECTION oF Parasitic Root or Cissus laciniata SHOWING THE FORMATION
CORK AND THE DISORGANIZED EPIDERMIS.

bundle, and about 2 cells inside the endodermis, there is
a plate which may be composed of leptome, and which in
some favorable material appears to be thickened wall
only.

A root 5 mm. in diameter has essentially the same
structure as the smaller one above described. The main
differences lie in the heavier cork and the thicker cor-
tex. The plate which lies opposite each fibro-vascular
bundle, also, is heavier. The endodermis is noticeably
poorer in starch.

680 THE AMERICAN NATURALIST [Vou. XLVI

TISSUES oF THE Host

The structure of the flat stems of Opuntia, broadly
speaking, consists of thin-walled, large parenchyma,
through which there course strands of conductive tissue.
Protection of the stem is afforded by a heavy cuticular-
ized epidermis.

When the parasitic relation is entered into, wound
tissue, with heavy outer walls in certain cells similar to
those of the cork, is formed about the injury caused by
the introduction of the cutting. The cutting sends out
adventitious roots which penetrate the parenchymatous
tissue of the host, and sooner or later these roots are
surrounded by wound tissue which the host promptly
organizes as a result of the unusual stimulation. By
this formation the water-storing ground tissue of the
host is separated from the living cells of the parasite.

Tissuz RELATIONS or ParastrE anp Host

In rapidly growing roots, contact is made with the
living parenchyma of the cactus, and the parasite is in
physical position to absorb foods and food materials. In
instances, however, where root growth is slow, wound
tissue is formed by the cactus, and the parasitic relation
is not favorable for absorption. Following the forma-
tion of wound tissue cork is organized by the parasite,
so that the cushion of non-living material separating host
and parasite in the older portions of the culture, comes
to be derived from both species.

When one compares the structural relations of a haus-
torium of a habitual parasite with the analogous absorb-
ing organ of such a xenoparasite as Cissus, several sug-
gestive inferences may be derived. The relation may be
presented briefly in the following parallel:

Xenoparasite Parasite
No special digestive cells, Epithelium - developed.

Root-hairs suppressed. No root-hairs formed.

No. 551]

Foods and food materials enter
haustorial root through epidermis.

Cork formed after establishment,
following organization of wound
tissue by host.

Tissues articulate with the corre-
sponding host tissues.

Meristematic tissue localized.

RELATIONS IN XENOPARASITISM |

681

Foods and food materials enter
haustorium through parenchyma,
sieve-tubes, and vessels.

Terminal portions at least of all
permanent tissues formed after
establishment.

The same.

Meristematic tissue not localized.

The parallel given above suggests, as intimated in an-
other paragraph, that any species which is to become
dependent on another species possesses to a large degree
the power of adaptability and morphological plasticity, so
that the direction of the development of its tissues or
organs can to a degree be modified. Atrophies result, and
the assumption of unaccustomed functions, and tissues
are organized in harmony with tissue formation, or other
physiological activity on the part of the host.

SHORTER ARTICLES AND DISCUSSION
ON TRICOLOR COAT IN DOGS AND GUINEA-PIGS |

AFTER reading Dr. Castle’s short article on this subject I
want to make a few remarks. His explanation of the peculiar
inheritance of the lemon and white and tricolor colors in Gal-
ton’s Bassett hounds will have to be somewhat modified. For it
is impossible to compare tricolor dogs and tricolor guinea-pigs.
Tricolor dogs are never irregularly spotted with black and yel-
low, as tricolor guinea-pigs, cats or rabbits, but they are in .
reality either black and tan, or else sable, spotted with white.
My attention being drawn to the subject of tricolor dogs by
Galton’s paper, I have never neglected an opportunity to ob-
serve dogs of this color, in dog-shows and from illustrations.
Some tricolor breeds, as the fox terrier, are black and tan,
spotted with white, others, as nearly all the hounds, are sable,
still others, such as collies, may be either black and tan or
sable, spotted with white. I have never seen an exception, such
as a dog with a yellow spot on the back and a black foot.

For the rest, I think Dr. Castle’s explanation is quite correct ;
it all depends upon the place of the spots upon a sable dog,
whether these will be yellow or black. A spot on a dog of
sable color, e. g., a fox hound, will always be black if it is on the
animal’s back. If on the muzzle, or on a foot, or far down on
the side, it will always be yellow, a spot, e. g., on the shoulder
may be partially black, partially yellow, shading off from one
color into the other.

It is of course possible that some of the Bassett hounds in
the pack recorded by Galton were real yellow and whites, and
we know from the evidence of breeders of dachshunds that yel-
low can be dominant over black and tan or sable in dogs. So it
may be possible that in that pack two real lemon-and-whites
(e. g., such as had a yellow spot on the back) have sometimes
given tricolor young, but in those cases in which two tricolor
parents gave lemon and white offspring, I feel sure, such young
were of that color only because they happened not to be pig-
mented in a spot where sable dogs show black color.

682 :

No.551] SHORTER ARTICLES AND DISCUSSION 683

It would certainly be interesting to try and find illustrations
of Galton’s hounds, especially of the lemon and white ones of
tricolor parentage.

In rabbits, there exist three wholly different classes of tricolor
animals. In the first place there are the real tricolors, those
animals which, if they were not partially albinistic, would be
irregularly spotted with black, agouti, blue or chocolate on a
yellow ground. They are comparable to the tricolor black-
yellow-white, blue-cream-white, ete., guinea-pigs and to the tri-
color cats. Secondly, there are those animals which are black
and tans, or blue (or chocolate) and tans, spotted with white.
‘These are comparable to the tricolor fox terriers, tricolor goats
and the so-called tricolor mice, which are sable, spotted with
white.t
Thirdly, there are those rabbits which, if not partially albin-
istice, would be ‘‘tortoise-shell,’? and which are comparable to
the spotted ‘‘tortoise’’ mice.

I think Galton’s hounds may have all been alike except for
the distribution of the pigmented patches on the coat. Those
hounds with the less white would then be called black and tan,
or sable; those with much white would be called tricolor, or
lemon and white, or even black and white, according as to where
the colored patches fell.

I am not so sure Galton’s black and tan hounds must neces-
sarily have been partially albinistic, as in dogs the partially
albinistic ones are generally so because of the presence of a
factor (or factors) absent from wholly colored ones. (In other
words, spotting with white is dominant in some dogs.)

The distribution of the colored area over partially albinistic
animals assuredly depends upon the cooperation of so many
factors (amongst which there are very probably some non-
genetic ones) that on our hypothesis the production of tricolor
young from yellow and white parents, and vice versa, becomes
very well possible.

AREND. L. HAGEDOORN

VERRIÈRES LE BUISSON ae

*“¢The Genetic Factors in the Development of the Housemouse,’’ A. L.
pei Zeitschr. f. indukt. Abst. und Vererbungslehre, 1911, Bd. VI,

e. á

NOTES AND LITERATURE

THE CLASSIFICATION OF THE LIVERWORTS

Boranists have long felt that the classification of the liver-
worts was very much in need of revision, and any serious attempt
to establish a classification that will better express the real inter-
relationship of the Hepatice is very welcome. The valuable
series of papers recently published by Dr. Cavers, on the classi-
fication of the Bryophytes, is a distinct contribution to the sub-
ject, and is a decided advance over any classification that has
been proposed hitherto.

Dr. Cavers is well known to students of the liverworts through
a series of papers of exceptional merit, published at intervals
during the past few years. The present publication presents at
length the conclusions he has reached as a result of his studies
on these important plants.

It is still too early to expect a definitive classification of the
liverworts, as there are still a good many important types whose
development is incompletely known, and it is also highly prob-
able that there are still forms awaiting discovery which we may
reasonably expect will throw light upon some relationships which
are still obscure.

Dr. Cavers has made a careful study of the work of the most
recent investigators, as well as of the older standard works, and
while one may take exception to a few of his deductions, still, as
a whole, one will agree with his main conclusions, and will
welcome this contribution of his as a decided advance in our
knowledge of the inter-relationships of the Bryophytes. The
Bryophytes (or “mosses,’’ using this term in its widest sense)
are forms of peculiar interest to students of plant-morphology,
especially to those engaged in the problems of the origin of the
higher types of plants; since the Bryophytes occupy an inter-
mediate place between the aquatic algæ and the ferns which are
typically land plants. While there is decided difference of
opinion as to how the ferns originated, the weight of evidence

1‘‘The Inter-relationships of the Bryophyta,’’ by Frank Cavers, D.Sc.,
sh mi (New Phytologist, Reprint No. 4), Cambridge; at the Botany School,

684

No. 551] NOTES AND LITERATURE 685

is strongly on the side of their direct derivation from some liver-
wort-like ancestor. It is this question that makes a thorough
study of the liverworts of such great importance in seeking for
an explanation of the origin of the vascular plants.

Aside from this, however, the Bryophytes, especially the liver-
worts or Hepatice, are exceptionally interesting, as they show in
a remarkably clear way many adaptations to special environ-
mental influences.

The Bryophytes are divided, usually, into two main groups—
the Liverworts (Hepatice), and the True Mosses (Musci). One
peculiar order, Anthocerotales, the ‘‘Horned Liverworts,’’ is
sometimes considered to represent a third class, coordinate with
the Musci and Hepatic. Cavers does not accept this view, but
considers them to represent an order only of the Hepatice.

Aside from the Anthocerotales, the liverworts usually are
divided thus into two orders, Marchantiales and Jungerman-
niales. There are, however, several genera that to a certain
extent combine characters of both of these orders and sometimes
have been assigned to one, sometimes to the other. Of these
genera Spherocarpus may be cited. This is, on the whole, prob-
ably the simplest known liverwort, and is represented in the
United States by several species in the warmer parts of the
country. Much like Spherocarpus is a peculiar liverwort,
Geothallus, known as yet only from San Diego in Southern Cali-
fornia. A third genus, Riella, evidently related to these, is an
aquatic type, only recently found in America. All of these are
very simple liverworts and probably stand near the base of the
liverwort series. They may, perhaps, be regarded as synthetic
types connected with both of the main series of liverworts.
Cavers proposes to unite them into a special order, Spherocar-
pales, and this conclusion will probably be accepted as repre-
senting best their position in the system. In the Spherocar-
pales, as interpreted by Cavers, the sporophyte or neutral gen-
eration is of simple structure, and the elaters which in the typical
liverworts accompany the spores are represented by undiffer-
entiated sterile cells.

From some form probably not very unlike Sphwrocarpus, but
with perhaps a still simpler sporophyte, it is probable that the
two lines of development, the Marchantiales and the Jungerman-
niales have diverged. Within these two orders the course of
development can be easily traced, as nearly all the stages in the

686 THE AMERICAN NATURALIST [Vov. XLVI

evolution of the two groups are represented by existing genera.
It is hard to say which of the two orders should be considered
the more primitive, as the lower members of each are of about
equal complexity, and can be derived equally well from some
form allied to Sphwrocarpus.

Spherocarpus has been associated most commonly with the
Ricciaceæ, the lowest of the Marchantiales, but there are certain
genera of the Jungermanniales that in many ways show a close
resemblance to the Spherocarpacex, and make it almost certain
that there is a real relationship existing between them. These
similarities are found both in the character of the thallus and
reproductive organs, as well as in the early history of the embryo.
They may be only cases of parallel development, but it is quite
as likely that they are true homologies. Two genera, Petalo-
phyllum and Fossombronia, which have always been placed in
the Jungermanniales, are especially suggestive of a possible con-
nection with the Spherocarpales, and it is by no means impos-
sible that it may turn out that these genera, and possibly some
others, should be removed from their association with the J unger-
manniales and transferred to the Spheerocarpales.

THE MARCHANTIALES

The Marchantiales constitute a very natural order, whose
simplest members, the Ricciacer, are sometimes separated as a
distinct order. There does not seem to be any valid reason for
this, however, as the Ricciacee are connected with the more
n specialized Marchantiaceæ by a number of intermediate
orms.

_ The Marchantiales are comparatively few in number, probably
nee more than three hundred species being known; but their
relatively large size and characteristic appearance make them
the most conspicuous of the liverworts, the common and wide-
spread Marchantia polymorpha being the most familiar liver-
Wort to most students of botany. The dichotomously branched
thallus, with its elaborate systems of tissues, probably may be
said to represent the highest type of a strictly thallose plant.
Within the Marchantiales are many interesting cases of
adaptation, and a very complete series of forms exists showing
the evolution of the elaborate and highly specialized thallus of
Marchantia and similar genera, from the much simpler type

No. 551] NOTES AND LITERATURE 687

found in Riccia. The elaboration of the sporophyte can also be
followed. Riccia, as is well known, has the simplest known sporo-
phyte, in this respect being in a much lower plane than Sphero-
carpus, although the thallus in the latter is much less specialized
than in Riccia.

The evolution of the sporophyte, as every botanist knows, is
associated with a reduction in the amount of tissue devoted to
spore-production, and a corresponding increase in the purely
vegetative or sterile tissue of the sporophyte. The latter, how-
ever, in the Marchantiales always remains relatively simple in
structure.

In the lower Marchantiales the sexual organs are borne upon
the dorsal surface of the unmodified thallus, but in the more
highly specialized types like Fimbriaria or Marchantia, charac-
teristic receptacles are developed, usually composed of a number
of very short branches resulting from the repeated dichotomy of
the original thallus apex. The classification of the Marchan-
tiales has been based largely on the character of the receptacle
and the sporogonium.

Cavers recognizes five families of very unequal size, viz., Ric-
ciacew, Corsiniacex, Targioniacee, Monocleacee and Marchan-
tiacee. The latter, which aside from the Ricciacee, comprises
the greater part of the Marchantiales, was divided by Leitgeb into
three subfamilies, Astropore, Operculate and Composite, but
it is very doubtful whether these can be maintained.

The Ricciaceæ, the great majority of which belong to the
genus Riccia, are undoubtedly the simplest, and probably the
most primitive, members of the order. The extremely simple
sporophyte is almost entirely devoted to spore production, there
being no sterile tissue beyond a very imperfect single outer layer
of cells. No other liverworts approach the Ricciacee in the sim-
plicity of the sporophyte.

The second family, Corsiniacex, is intermediate in the struc-
ture of the sporophyte, between the Ricciacee and the higher
Marchantiales,

The third order, Targioniacex, includes the two small genera,
Targionia and Cyathodium. These are very characteristic liver-
worts represented in the United States by a single species Tar-
gionia hypophylla, common in the coast region of California, but
absent from the eastern states. This species occurs also m
southern and western Europe. Cyathodium includes a few

688 THE AMERICAN NATURALIST [Vou. XLVI

species of delicate liverworts inhabiting dark crevices in rocks,
or shallow caves. All the species show evidences of marked
structural modifications due to their unusual habitat. C. fæti-
dissimum is a characteristic species of the Indo-Malayan region.

The simple genus Monoclea with two species represents very
distinct the family Monocleacee. In his great work on the Hepat-
ice, Leitgeb referred Monoclea to the J ungermanniales, and this
view was adopted by Schiffner in his treatment of the Hepatice
in Engler & Prantl’s ‘‘ Natiirliche Pflanzenfamilien.’’ This asso-
ciation with the Jungermanniales was mainly on account of the
structure of the thallus, which is quite destitute of the air-
chambers which distinguish most of the Marchantiales. There is
also in Monoclea no definite archegonial receptacle, and the soli-
tary sporogonium has a long seta like that of many Jungerman-
niales,

All the more recent students of Monoclea, however, are agreed
that the plant really belongs to the Marchantiales, this being
shown both by the structure of the thallus, and that of the repro-
ductive organs. The absence of air-chambers is with little ques-
tion to be looked upon as a secondary condition, due to the semi-
aquatic habit of the plant. A similar disappearance of the air-
chambers is known in the unmistakable marchantiaceous genus
Dumortiera.

Leitgeb, in his important memoirs on the Hepatice, recognized
three types of archegonial receptacle. Only in one of these was
the receptacle compound in its structure. More recent studies,
including those of Cavers, indicate that this compound or ‘‘com-
posite’’ type is much more general than Leitgeb supposed.
Cavers states that probably all of the genera of the Marchan-
tiaceæ, except Clevea and Plagiochasma, will be found to have
receptacles of the composite type.

In tracing the phylogeny of the Marchantiales, Cavers distin-
guishes two main divergent groups which are connected with the
Ricciaceæ by Corsinia and Boschia, respectively. The first series
includes, among other genera, Clevea, Plagiochasma, Reboulia
= Fimbriaria, the latter representing the culmination of this
series.

The second series, starting with Boschia, shows two main
branches, one including the Targioniacer and Monoclea, the
other the most highly developed genera, like Fegatella, Dumor-

No. 551] NOTES AND LITERATURE 689

tiera and Marchantia. The latter genus is the most highly spe-
cialized of all the Marchantiales.

THE JUNGERMANNIALES

Much the greater number of liverworts belong to the Junger-
manniales. The classification of this large order is very much
in need of revision, as it is at present in a very unsatisfactory
condition.

They are generally divided into two series of very unequal size,
this division being based upon the position of the archegonium
—and are denominated the Anacrogyne and Acrogyne. In the
former the growing point of the shoot persists indefinitely, while
in the latter, in the fertile shoots, it is sooner or later trans-
formed into an archegonium, and the sporogonium is therefore
terminal.

The name Metzgeriacee was later proposed by Underwood, as
a substitute for Leitgeb’s Anacrogyne, the Acrogyne being alone
called Jungermanniacee. Cavers thinks these two divisions are
largely artificial, and it must be admitted that there is much to
be said for his view.

Comparing the Jungermanniales, as a whole, with the Mar-
chantiales, it is seen that in the former specialization has been in
the direction of external differentiation, i. e., in most of them a
more or less definite axis, bearing leaves, is present, but the tis-
sues remain quite uniform. In the Marchantiales, on the other
hand, the plant is a thallus, but the tissues are of various kinds.

The simplest of the Jungermanniales, e. g., Aneura, Pellia,
ete., have a very simple thallus, either composed of quite similar
cells, or with a midrib which may possess a strand of special
conductive tissue. The simplest type of thallus is quite like that
of Spherocarpus, and may very well have originated from some
similar type.

In these thallose Jungermanniales there is frequently a tend-
ency toward the development of marginal lobes, which may bear
a quite definite relation to the primary divisions of the single
apical cell of the thallus. Such marginal lobes are undoubtedly
homologues of the leaves found in the more highly specialized
leafy liverworts—the “ Acrogyne,’’ of Leitgeb. Sometimes these
leaf-like organs of the anacrogynous liverworts are very distinct,
as in Treubia, and the transition to the typical leafy liverworts

690 THE AMERICAN NATURALIST — [Vou. XLVI

like Porella or Frullania, is a very gradual one. It is very clear
that this tendency towards leaf-development has arisen in a
number of quite disconnected genera, and this of course suggests
a multiple origin for the Acrogyne.

Cavers proposes four families of the lower, or anacrogynous,
Jungermanniales, viz., Aneuraceæ, Blyttiaceæ, Codoniacee, Calo-
bryaceæ. He thinks that the first three are more or less arti-
ficial, and it is very certain that it will be necessary when some
of the less known genera are more fully investigated, to make a
radical revision of these families. The Calobryaceæ, on the other
hand, forms a sharply defined and natural family, comprising
two genera, Haplomitrium and Calobryum. Cavers concludes
that there are two main lines of development within the Anacro-
gynæ, one including the Codoniaceæ and Calobryaceæ, the other
the Blyttiaceæ and Aneuraceæ, suggesting that the two latter
families might perhaps be better united into a single one. There
seems to be little question that the two families are closely related
through such forms as Umbraculum and Podomitrium.

There is much uncertainty as to the limits of certain genera.
This is especially the case with the genus Calycularia, to which
have beèn assigned species which further investigation has shown
to belong to quite different families. The writer has had occa-
sion recently to examine carefully the structure of Calycularia
radiculosa, a rare species from Java. Schiffner concluded that
this species should be removed from the genus Calycularia, of
the family Codoniaceæ, and united with Mérkia, a member o
the Blyttiacee. While it is certainly distinct from the true
species of Calycularia, it is equally ‘certain that it can not be
assigned to Mérkia. It will probably have to be separated into
a distinct genus with characters intermediate between those of
the Codoniaceæ and the Blyttiaceew. In short, it is very clear
that at present a satisfactory classification of the group is not
feasible.

The Aneuracee and Blyttiacee show an interesting type of
specialization of the thallus which is wanting in the Codoniacer
and Calobryacexw, where the tendency is toward the development
of leaf-like lobes foreshadowing the leaves of the leafy liver-
worts. In Podomitrium and Umbraculwm, assigned respectively
to the Blyttiacee and Aneuraceæ, the thallus is differentiated
into a prostrate cylindrical rhizome and erect dichotomously
branched fan-shaped shoots, which resemble very closely the deli-

No. 551] NOTES AND LITERATURE 691

cate leaves of certain filmy. ferns, for which these liverworts
might easily be mistaken.

In the development of the sporophyte the Anacrogyne show a
decided advance over the Marchantiales. There may be devel-
oped a considerable amount of sterile tissue in the capsule, aside
from the ordinary elaters, and this sterile tissue sometimes
assumes the form of a sort of columella or ‘‘elaterophore,’’ sug-
gesting the columella found in the Anthocerotacex, and possibly
homologous with it. This elaterophore may be either apical
(Aneura) or basal (Pellia).

While recognizing the entirely independent origin of leaves in
several lines of the Anacrogyne, nevertheless Cavers is inclined
to believe that all of the true leafy liverworts (Acrogyne) can
be traced back to a single type which he thinks is best repre-
sented by Fossombronia, which genus he places at the top of the
series Codoniacew. It may be said, however, that there are some
strong arguments in favor of a polyphyletie origin for the Acro-
gyne—a view which has been defended by several students of
the group.

There are, as we have already stated, good reasons for believing
that Fossombronia should not be associated with Pellia and the
other Codoniacex, but associated with the Spherocarpales, as the
highest member of a series of which Sphwrocarpus and Geothal-
lus are lower members. This interpretation would not interfere
with the acceptance of Cavers’s view that some at least of the
leafy liverworts have been derived from forms like Fossombronia.

The acrogynous Jungermanniales, or leafy liverworts, include
much the larger part of existing liverworts. Of about 250
genera and 4,500 species of known liverworts, all but 60 genera
and 700 species belong to the acrogynous Jungermanniales.
They are nevertheless comparatively uniform in type, and Cavers
believes that they may all be traced back to a common ancestral
type allied to Fossombronia. ;

With very few exceptions they show a single tetrahedral apical
cell and usually three series of leaves corresponding to the three
lateral faces of the apical cell. The ventral leaves (amphigas-
tria) are not infrequently absent, and both dorsal and ventral
leaves often show various modifications, among the most striking
of which are hollow sacs presumably developed for water storage.

The tissues are very simple, and only very rarely is there any
specialization of cells for conduction or other purposes. In size

692 THE AMERICAN NATURALIST [ Von. XLVI

they range from almost microscopic forms like some of the
minute epiphyllous Lejeuniacex, to stout species like some of the
tropical Frullanias, which form pendant masses several feet in
length.

In all the Acrogyne the archegonia are in groups terminating
the fertile branch, whose further growth is arrested by the trans-
formation of the apical cell into an archegonium.

The sporogonium is always well developed, usually showing a
well-marked foot and seta. Perfect elaters are always present.
A small number only of the Acrogyne have been studied crit-
ically with reference to the development of the sporophyte, and
much more work must be done before the real affinities of some
of the genera can be determined satisfactorily.

On the basis of our present knowledge of the group, Cavers
proposes a classification based largely upon the work of Spruce.
He recognizes two main divisions, the first including a single
very large family, Lejeuniaceæ, with nearly 2,000 species; the
second contains seven families, of which three, viz., Porellaceæ,
Pleuroziacee and Radulacex, are regarded as natural families,
the other four as more or less artificial, the limits between them
being difficult to define.

The inter-relationships of the Acrogyne are extremely difficult
to follow. A number of students of the liverworts, notably
Spruce and Schiffner, believe that the group is of polyphyletie
origin, the Lejeuniacew representing a quite distinct line derived
from forms allied to the Aneuracee. There are striking resem-
blances both of gametophyte and sporophyte, the former in some
cases having a protonemal stage of long duration, and very much
resembling one of the simpler thallose liverworts. Cavers be-
lieves, however, that these resemblances are simply parallel devel-
opments, and not true homologies; and, as already stated, that
the Acrogyne represent a single line of development. Of these
forms he states that Lophozia probably comes nearer to the as-
sumed ancestral type.

From the Lophozia type, three branches are traced, one
through Plagiochila developing a large number of genera, among
which Cephalozia, again, is the starting-point for the develop-
ment of a number of specialized genera like Z oopsis, Lepidozia,
and Trichocolea. The second line leads through Marsupiella
and Nardia to a number of genera, of which the highest are
Stephaniella, Gyrothyra and Symphyomitra. The third line,

No. 551] NOTES AND LITERATURE 693

beginning with Lophozia, leads through Sphenolobus to the great
family Lejeuniaceæ, and to the characteristic genera Porella and
Frullania, which may be considered to represent the most per-
fectly developed characters of the whole order.

THE ANTHOCEROTALES

The Anthocerotales, Cavers’s fifth order of Hepatic, com-
prise a comparatively small number of liverworts of very
peculiar structure, and very readily distinguished from all other
plants. The differences between them and the other liverworts
are so marked that they are sometimes considered a class—Antho-
cerotes—coordinate, on the one hand, with all the other liver-
worts, on the other with the true mosses. | :

The structures of the four genera which are comprised in the
order are so much alike that they can all be assigned without
question to a single family, Anthocerotacee.

The gametophyte is of simple structure, and all the cells much
alike, each as a rule containing a single large chromatophore
resembling that of many green alge. The reproductive organs,
both archegonia and antheridia, show certain peculiarities, which
in some ways have their nearest approximation among the lower
ferns, and in connection with the characters of the sporophyte
suggest a real connection between the ferns and the Antho-
cerotalea.

The sporophyte differs much from that of the other liverworts.
In all of thé Anthocerotacea, except possibly some species of
Notothylas, the spore-producing tissue all arises from the outer
region (amphithecium) formed by the first periclinal divisions
in the capsule, and much the greater part of the tissue of the
sporophyte remains sterile. In all cases a large foot is present,
and above it a zone of actively dividing cells is present, which
may retain its activity for several months, so that the sporophyte
may attain a length of 10 centimeters or more. As the outer
tissues are in most cases well provided with chlorophyll, and
sometimes with stomata, a complete photosynthetic apparatus
is established much in advance of anything found in the other
Hepatice.

This long-continued growth of the sporophyte is associated
with a central strand of conducting tissue (columella), which is
reminiscent of the axial vascular bundle of the young sporo-

694 THE AMERICAN NATURALIST [Von XLVI

phyte of some of the lower ferns to which the sporophyte of
Anthoceros shows the closest resemblance known in the Bryo-
phyta.

Within the Anthocerotacer is an interesting series connecting
the small sporophyte of Notothylas with its relatively large
development of sporogenous tissue, and the large sporophyte of
Anthoceros with a small amount of sporogenous tissue and a
highly developed photosynthetic system.

It is, at present, impossible to say whether or not the type of
Notothylas is a reduced one. Cavers believes it is a primitive
type from which the more highly developed genera, culminating
in Anthoceros, have been derived. He is inclined to minimize
the importance of certain striking features, e. g., the peculiar
chloroplasts and the endogenous antheridia, and thinks the dif-
ferences between the Anthocerotales and the other liverworts are
not sufficient to warrant their separation into distinct classes.
He considers the columella of the Anthocerotacee may be con-
nected with the true liverworts through the Spherocarpales,
which they resemble in a number of particulars.

It may be noted, in passing, that there is a possibility of a
connection of the Anthocerotacee with some of the lower Mar-
chantiales. The Targioniacex, especially Cyathodium, for ex-
ample, show some interesting analogies in the sporogonium with
Notothylas, and the gametophytes also agree in the presence of
large lacunæ, and the chromatophores of Cyathodium are also
of unusual size.

Cavers’s conclusions may be summarized as follows: From
some common ancestral form, ‘‘Sphxro-Riccia,’’ two lines of
development diverged, one leading to the Marchantiales, the
other to the Spherocarpales, which in turn gave rise to the lower
Jungermanniales. From some member of the latter, perhaps
Fossombronia, all of the leafy liverworts arose. Somewhere near
the Spherocarpales it is assumed that the Anthocerotales
branched off.

| We are inclined to believe that some modifications of this
arrangement are likely to be made. It is quite possible that
Fossombronia should be removed from the J ungermanniales, and
associated with the Spherocarpales; and if Cavers’s assumption
1s correct, that the leafy liverworts (Acrogyne) have arisen
from a prototype resembling Fossombronia, this would entirely
divorce the two great divisions of the Jungermanniales.

No. 551] NOTES AND LITERATURE 695

It is doubtful whether the derivation of the Anthocerotacese
from the Spherocarpales will be generally accepted. For the
present, at least, the order must be regarded as a very isolated .
one, and perhaps best considered to represent a distinct class,

DovueLtas HOUGHTON CAMPBELL

STANFORD UNIVERSITY

INVERTEBRATES

Unper the able leadership of Professors Zeigler and Woltereck
there is appearing from Klinkhardt’s press in Leipzig a series
of excellent small monographs of familiar animals designed for
the student, teacher, investigator and amateur who desires to
secure a brief but authentic account of the results of systematic,
histological, morphological, anatomical and embryological investi-
gations on representative types of animals. Two volumes have
already appeared, the frog by Dr. Hempelmann, and the rabbit
by Dr. Gerhardt, and the series of invertebrates has been
introduced by two volumes, volume 3 of the series on ‘‘Hydra
und die Hydroiden’”’ by Dr. Steche, of Leipzig, and volume 4 by
Professor Meisenheimer, of Jena, on ‘‘Die Weinbergschnecke.”’

Dr. Steche’s volume is designed not merely as a monograph
on Hydra along the lines on which the series is planned, but
adds to these the features of an introduction to the experimental
treatment of biological problems as offered by the lower ani-
mals. Hydra is an exceptionally favorable subject for this
treatment by virtue of its hardiness, ease of obtaining and of
maintenance, and simplicity of structure. Few invertebrates
have served as a basis of so many and so varied experimental
tests and have been the object of so many investigations as
Hydra. With this wealth of results before him it is not to be
wondered at that this modest volume is open to the charge of
Some sins of omission. The choice of topics treated is, however,
most catholic and this author has wisely avoided controversial
difficulties. The histological and embryological sections are less

1 ‘í Monographien einheimischer Tiere,’’? Herausgegeben von Professor
Dr. H. E. Ziegler, Stuttgart, und Professor Dr. R. Woltereck, Leipzig, Bd. 3;
‘‘ Hydra und die Hydroiden. Zugleich eine Einführung in die experimentelle
Behandlung biologischer Probleme an niederen Tieren,’’ von Dr. Otto Steche,
vi + 162 pp., 65 figs. in text and 2 pls., M. 4, geb. M. 4.80; Bd. 4, “Die
Weibergschnecke, Helix pomatia,’? von Professor Johannes Meisenheimer,
140 pp., 1 pl. and 72 figs. in text, M. 4, geb M. 4.80.

696 THE AMERICAN NATURALIST [Vou. XLVI

developed than seems desirable, but in compensation the sec-
tions on biology and experimental subjects such as regenera-
tion, regulation, grafting, graft hybrids, effects of external fac-
tors on growth and regeneration, polarity and heteromorphosis
are well, though concisely, developed. Several pages of prac-
tical suggestions as to collection, rearing, feeding and preparing
Hydra will be found very useful as will also the key to the
species. ‘The author conservatively clings to the widely current
names viridis, grisea and fusca and rejects the older names of
Pallas which strictly have priority.

Half of the book is given to the hydroids. Noteworthy in this
are several superb figures of hydroid colonies from the Hel- |
goland Nordsee Museum. A brief list of titles closes the volume
from which we note the omission of Nutting’s and Mayer’s
monographs.

The volume by Professor Meisenheimer upon the garden snail
follows closely the program of the series, with perhaps less of
emphasis upon the experimental and physiological aspects and
more space taken for the presentation of the static phases which
are greatly increased necessarily over those of a simple animal
such as Hydra. But there appears still to be call for more
expansion on the dynamic aspects of the subject in the case of
this volume. The chapter upon the relation of the snail to the
environment and to man is a concession in the right direction,
and the prevalence of the biological standpoint throughout the
anatomical chapters in some measure supplies the physiological
data pertinent to the structural phases. These are very clearly
and methodically set forth with abundant illustrations, many
of which are new. A chapter on other land pulmonate mollusks
affords an all too brief basis for comparison of the snail with
other mollusks,

; Both of these volumes will be exceedingly useful to zoologists
in all countries, for the objects with which they deal are cos-
mopolitan. A similar series of monographic booklets on labo-
ratory types based on American material would be of great
value for American students and investigators.

CHARLES ATwoop Kororp

DECEMBER, 1912

. XLVI, NO. 552

VOL

THE
AMERIC

AN

JOURN

A M0

The American Naturalist

MSS. intended for publication and books, etc., intended for Sai should be
sent to the Editor of THE posa ICAN NATURALIST, Garrison-on-Hud n, New York
les containing research work bearing on the panu of ‘organi evolu-
tion are especially welcome, sce will be given preference in publica
One hun nts of ener ect are supplied to authors Res of charge.
Further Crease will be supplied at co
> ead ge and pi cainin should be sent to the poetais The
n price is four dollars a year. Foreign postage is and
elt an cents additional. The charge for single peiri 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

eats 4 tter, April 2, 1908 ter, Pa., under the Act ot
Congress of March 8, 1879.

FOR SALE TENTH EDITION.
ARCTIC, ICELAND and GREENLAND
BIRDS’ SKINS, TH E MICROSCOPE,
Well Prepared Low Prices Seed —_ GAGE, A n io on chan, a
: e figures in
GD ESEN. acc, OD beri revised edition. Price $2.00, postpaid.
Husavik, North Iceland, Via prea England | COMSTOCK PUBLISHING CO., Ithaca, N. Y.
RECENT NUMBERS

Reptiles, Amphibians
B res 19. An Empirical Study of Certain Tests tor
tos Fishes À sprit as aes pegs ie $1.50. THEODORA W

An Alark get Sh of various “tests of memory,

association, ception, pia Sibe ete., with indications
of the nepe Lag value of different forms ‘of test.

Archives of Psychology

fasg

| i ts of Different

W. F.H. ROSENBERG, _ | Colors and te ne Dayıp EpGAR RICE.
An etermination of visual acuity under

57, , Haverstock Hill, London, N. ee red, green, blue, and uncolored illumination, and of the de

‘dence of this acuii on intensi fe ondoan
na oe bon drene Pirima ah

cents. — pies
ores ol freeing on ey

and material, an

THE
AMERICAN NATURALIST

Vout. XLVI December, 1912 No. 552

THE MENDELIAN INHERITANCE OF FECUND-
ITY IN THE DOMESTIC FOWL?

DR. RAYMOND PEARL

MAINE AGRICULTURAL EXPERIMENT STATION

Tue investigation here reported was concerned with
the detailed analysis and interpretation of a rather ex-
tensive series of data regarding the inheritance of fe-
cundity in the domestic fowl. The basic data are derived
from trap-nest records extending over a period of years.
They include records from (a) pure Barred Plymouth
Rocks; (b) Cornish Indian Games; (c) the F, individ-
uals obtained by reciprocally crossing these two breeds;
and (d) the F, individuals obtained by matjng the F,’s
inter se and back upon the parent forms in all possible
combinations. The fully-pedigreed material made use of
in the present connection includes something over a
thousand adult females, each of which was trap-nested
for at least one year, and many for a longer period. This
material covers four generations. The birds of the
fifth generation have just completed their winter records
at the time of writing. Besides this fully pedigreed ma-
terial, the collection and study of which has occupied

1 At the request of the editor of the AMERICAN NATURALIST the following
summarized account of the principal results of an investigation carried out _
by the writer has been prepared. A detailed account has been published
in the Journal of Experimental Zoology, Vol. 13, No. 2, pp. 153-268,
August, 1912.

697

698 THE AMERICAN NATURALIST [Vou. XLVI

five years, there was available as a foundation, without
which the results here discussed could not have been
reached, nine years of continuous trap-nest records for
Barred Plymouth Rocks, involving thousands of birds,
which had been subjected during this long period to mass
selection for increased egg production.

Altogether it may fairly be said that the material on
which this work is based is (a) large in amount, (b) ex-
tensive in character, and (c) in quality as accurate as it is
humanly possible to get records of the egg production of
fowls.2 On these accounts the facts presented seem
worthy of careful consideration, and to have a perma-
nent value quite apart from any interpretation which
may be put upon them.

The essential facts brought out in this study of fe-
cundity appear to be the following: ‘

1. The record of fecundity of a hen, taken by and o
itself alone, gives no definite, reliable indication from
which the probable egg production of her daughters may
be predicted. Furthermore mass selection on the basis
_ of the fecundity records of females alone, even though
long continued and stringent in character, failed com-
pletely to produce any steady change in type in the di-
rection of selection. 2

2. Fecundity must, however, be inherited since (a)
there are. widely distinct and permanent (under ordi-
nary breeding) differences in respect of degree of fe-
cundity between different standard breeds of fowls com-
monly kept by poultrymen, and (b) a study of pedigree
records of poultry at once discovers pedigree lines (in
some measure inbred of course) in each of which a defi-
nite, particular degree of fecundity constantly reappears
generation after generation, the ‘‘line’’ thus ‘‘breeding
true’’ in this particular. With all birds (in which such
a phenomenon as that noted under b occurs) kept under
the same general environmental conditions such a result

* Pearl, R., ‘‘On the Accuracy of Trap-nest Records,’’ Me. Agr. Expt.
Sta. Ann. Rept. for 1911, pp. 186-193.

No. 552] INHERITANCE OF FECUNDITY 699

can only mean that the character is in some manner in-
herited.

The facts set forth in paragraphs 1 and 2 have been
presented, and, I believe, fully substantiated by exten-
sive evidence, in previous papers from this laboratory.
It is now further shown that:

3. The basis for observed variations in fecundity is
not anatomical. The number of visible oocytes on the
ovary bears no definite or constant relation to the actu-
ally realized egg production. This is shown by the fig-
ures presented in Table I. These give the counts of the
number of ọocytes on the ovary visible to the unaided eye
in the case of a number of individuals. It will be under-
stood that it is not contended that such counts give an
accurate measure of the total oocyte content of the
ovary. The figures, however, are so greatly in excess of
what a hen actually ever lays that it may be quite safely
concluded that in normal cases (where no accident or
operation has induced regenerative processes in the
ovary) all the eggs which will ever be laid (and usually
more) are included among those visible to the eye, on an
adult fowl’s ovary.

From this table it is evident that when one bird has a
winter record of twice what another bird has it is not
because the first has twice as many oocytes in the ovary.
On the contrary it appears that all birds have an ana-
tomical endowment entirely sufficient for a very high de-
gree of fecundity, and in point of fact quite equal to that
possessed by birds which actually accomplish a high
record of fecundity. Whether or not such high fecund-
ity is actually realized evidently depends then upon the
influence of additional factors beyond the anatomical
basis.

4. This can only mean that observed differences (vari-
ations) in actual egg productions depend upon differ-
ences in the complex physiological mechanism concerned
with the maturation of oocytes and ovulation. ;

THE AMERICAN NATURALIST [Vou. XLVI

700

"POT WOYM uorprpuos Surer ur 40u SPIE q
‘uorgonpord ye}0} oy} ATTeatjowrd yuosordos somSy oy, 'OTGT ‘T 19qWəAOoN 0} dn sprooer ony
qmq ‘Ər 10f sp10əə1 OY} JOU ere səseə əsəqy ur Spode ZZə oy, ‘Apnys Juosord oy} ur əsn 107 sjenprarpur pəzsou-deig osey} ow
aavs Aypury A19A OM ‘ÅPSIOATUN [[OULOH JO PNY “HT SIWLf IOSSOFOIg 0} po}qopUl We I Spalq Furmooy 9914} oy} PUL sy} 107,
‘norjdiosqe Jo ssovord ut SJO g SOPNPOUT SIIL s

98G CES ge g tt; hae Arenue (* : Poe? ey ee eM a tae ee oe uəy gaum g . nri
G94 LIL 9g € 6 pi IT, ‘°° Annur ‘‘uey voummy | t
000‘Z G18'T OL g og 90T TI, ‘03 YPN OL 1 Yorn reee wG
ose'T EZET L9T 9 i24 gI Ol et Ane 60, ‘13 mdy ‘owen UBIPUy ystu | g9
GPI‘ ZZ0‘Z 08 a EF 0 OT, ‘Zs tequieda(T 60. 3i oune "sees * sq10yseyT OI M | Sess
TOL‘T 929'T cL ¢ II 0 OL, ‘E1 10qW999q 60. ‘IZ APIA rettet UIB OTM | ers
09g 6L7'£ 80T T LIZ oe OT, ‘ST tequie09q7 60, ‘86 SOW [°° " "Tr uroq3oT OA | 2190'S
Zorg OFS 1&3 ¢ GL ¥¢ OT, ‘OZ Joquieveqy 60. ‘ST ABI **,usL0ysa'T ONM | OFG'e
IZSIT OFT raat ol 13 0 OLL Amp 60. ‘08 GAP |' 0 : Minoti 01IVE | CSP
9LS'T ccr't Z6 g VG g TI. ‘21 nyw OT, ‘61 ABN | YoY Youd potueg | O10'S
TOTZ 620° 6? 9 LI 0 Il. ‘FZ PERN OI.‘ unf | poy ymos poeg | 600'8
FOl'T SSI'T GP 9 gz 0 IL‘ PN or ‘gz ounf | poy ymouwsjq poeg | 810'8
908'S ietz 6z L 6f g Tle? judy OT, ‘8% mdy |*‘yo0y yynoursyg poeg | 198'T
PLT 960'T 89 8 ZI g Il. ‘FI YN OT, ‘6 oun yoy ynours[g poeg | ¢o0's
¥16 ‘688 Z9 G 8 0 TI: ‘Ol YPWN OL‘T unf | poy ymowsq poaeg | Og0's
999'T 96¢'T 19 L ZI 0 II, ‘08 Your Ol.‘ ong |** yoy yynoursyg poeg | 410'8
8Z2'T 6FL'T eg 6 LI g Il, ‘8G Your Ol. ‘IT unf |P oy yynours[g poeg | 20'S

|
iI9,ouUse UT A9JIMBICT Ul sz
PRIA 10d | THT 190p GOA LOM 3940 10 °W? paiolo
-un £9}Á900| [ 8944900 | 1 89474000 pains tery § POEM 9A Suryoyey] Jo ov pooig ‘ON PAIL

yunog UBBO

Saug NIVLYID dO AYVAQ FHL NI SHLAOQO AIAISIA, A0 YIJWAN THL ONIMOHY

I TIY

No. 552] INHERITANCE OF FECUNDITY 701

5. A study of winter egg production (taken for prac-
tical purposes as that from the beginning of the laying
year in the early fall to March 1) proves that this is the
best available measure of innate capacity in respect to
fecundity, primarily because it represents the laying
cycle in which the widest difference exists between birds
of high fecundity and those of low fecundity.

6. It is found to be the case that birds fall into three
well-defined classes in respect to winter egg production.
These include (a) birds with high winter records, (b)
birds with low winter records, and (c) birds which do
not lay at all in the winter period (as defined above).
The division point between a and b for the Barred
Plymouth Rock stock used in these experiments falls at
a production of about 30 eggs.

7. There is a definite segregation in the Mendelian
sense of the female offspring in respect to these three
fecundity divisions. This is demonstrated by extensive
statistics in the complete report of this work. Here a
single table only may be given by way of illustration, the `
one chosen being taken because all three classes are rep-
resented among the progeny of the particular type of
mating with which it deals.

TABLE II

SHOWING THE RESULTS OF ALL Mares or Crass 4 B.P.R. fd X CLASS
1 B.P.R. 99. GAMETIC CONSTITUTION: fL,L, - ful. X (LL, - Ful,

Numb {ndividual
kavalsi in Matings a Winter Egg Production of Daughters
of this Type
Total Adult
do | vy Class Over 30 | Under 30 Zero | Q Progeny
4 | 17 Observed 21 30 8 59
Expected 22.1 29.5 74
Mean winter egg production of all os ERS A en eas
C O in indicated class... 0.0.5. 3 48.85 eggs 16.34 eggs) 0 eggs

8. High fecundity may be inherited by daughters from
their sire, independent of the dam. This is proved by the
numerous cases presented in the detailed evidence where
the same proportion of daughters of high fecundity are

702 THE AMERICAN NATURALIST [Von XLVI

produced by the same sire, whether he is mated with
dams of low or of high fecundity.

9. High fecundity is not inherited by daughters from
their dam. This is proved by a number of distinct and
independent lines of evidence, of which the most im-
portant are: (a) continued selection of highly fecund
dams does not alter in any way the mean egg production
of the daughters; (b) the proportion of highly fecund
daughters is the same whether the dam is of high or of
low fecundity, provided both are mated to the same
male; (c) the daughters of a fecund dam may show
either high fecundity or low fecundity, depending upon
their sire; (d) the proportion of daughters of low fe-
cundity is the same whether the dam is of high or of low

fecundity, provided both are mated to the same male.

10. A low degree of fecundity may be inherited by the
daughters from either sire or dam or both.

11. The results respecting fecundity and its inherit-
ance stated in paragraphs 3 to 10 inclusive are equally

° Pearl, R., ‘‘The Relation of the Results Obtained in Breeding Poultry
for Increased Egg Production to the Problem of Selection,’’ Rpt. 30th
Meeting Soc. Proc. Agr. Sci., pp. (of reprint) 1-8, 1910; ‘‘Inheritance in
‘Blood Lines’ in Breeding Animals for Perf. ormance, with Special Refer-
ence to the ‘200-egg’ Hen,’’ Ann. Rpt. Amer. Breeders’ Assoc., Vol. 6,
pp. 317-326, 1911; ‘‘Inheritance of Fecundity in the Domestic Fowl,’’
AMER. Nar., Vol. 45, pp. 321-345, 1911; ‘‘Breeding Poultry for Egg
Production,’’? Me. Agr, Expt. Sta. Ann. Rpt. for 1911, pp. 118-176. Pearl,

.„ and Surface, F. M., ‘‘Data on the Inheritance of Fecundity Obtained
from the Records of Egg Production in the Daughters of ‘200-egg’ Hens,’’
Me. Agr. Expt. Sta. Ann. Rpt. for 1909, pp. 49-84 (Bulletin 166), 1909;
**Studies on the Physiology of Reproduction in the Domestic Fowl.

Data on Certain Factors Influencing the Fertility and Hatching of Eggs,’’
Me. Agr. Expt. Sta. Ann. Rpt. for 1909, pp. 105-164, 1909; ‘ʻA Biometrical
Study of Egg Production in the Domestic Fowl. I. Taraka in Ann
Egg Production,’’ U. S. Dept. Agr., Bur. Animal Ind. Bulletin 110, Part I,
pp. apse 1909; ‘‘A Biometrical Study of Egg Production in the Domestic
Fowl. ensonad Distribution of Egg Production,’’ Ibid., Part II, PP.
81-170, Da

‘This is true, of course, only for certain gametic types of low fecundity
females, as will be clear to any one who has studied the detailed evidence.
This limitation, however, in nowise diminishes the force of this particular
evidence in titer of the conclusion standing at the beginning of paragraph 9.

a

No. 552] INHERITANCE OF FECUNDITY 703

true for Barred Plymouth Rocks, Cornish Indian Games,
and all cross-bred combinations of these breeds in F,
and Fs

The above statements are of definite facts, supported
by a mass of evidence. Their truth is objective and de-
pends in no way upon any theory of inheritance whatso-
ever. With this clearly in mind we may undertake their
interpretation.

It is believed that these general facts, and the detailed
results on which they are based, are completely accounted
for and find their correct interpretation in a simple Men-
delian hypothesis respecting the inheritance of fecund-
ity in the fowl. This hypothesis involves the following
points, each of which is supported by direct and perti-
nent evidence derived either from physiological and
statistical studies of fecundity, or from the detailed data
respecting the mode of inheritance of this character.

It is assumed in this hypothesis that:

1. There are three distinct and separately inherited
factors upon which fecundity in the female fowl depends.

2. The first of these factors (which may be called the
anatomical) determines the presence of an ovary, the
primary organ of the female sex. The letter F is used
throughout to denote the presence of this factor.

3. There are two physiological factors. The first of
these (denoted by L,) is the basic physiological factor,
- which when present alone in a zygote with F brings about
a low degree of fecundity (winter record under 30 eggs).
This factor is under no limitations in gametogenesis, but
may be carried in any gamete, regardless of what other
factors may be also present.

4. The second physiological factor (denoted by L,)
when present in a zygote together with F and L,, leads to
a high degree of fecundity (winter record over 30 eggs).

$ And F,. It was thought wise to delay publication any longer in order
to include the data for F, It may be said, however, that they are in full
accord with those which have been obtained from earlier cross-bred genera-
tions and the parent forms.

704 ‚THE AMERICAN NATURALIST __[Vou. XLVI

When L, is absent, however, and L, is present the zygote
exhibits the same general degree of fecundity (under 30)
which it would if L, were present alone. These two inde-
pendent factors L, and L, must be present together to
cause high fecundity, either of them alone, whether
present in one or two ‘‘ doses,’’ causing the same degree
of low fecundity.

5. The second physiological factor L, behaves as a
sex-limited (sex-correlated or sex-linked) character, in
gametogenesis, according to the following rule: The
factor L, is never borne in any gamete which also carries
F. That is to say, all females which bear L, are hetero-
zygous with reference to it. Any female may be either
homozygous or heterozygous with respect to L, Any
male may be either homozygous or heterozygous with
reference to either L,, L, or both.

TABLE III
CONSTITUTION OF BARRED PLYMOUTH ROCK MALES IN RESPECT TO FECUNDITY
Class Zygote Gametes Produced
ee | LiL , flrla iL
2 fink: . flalz fInLe2, fInle
3 ibe . file JLiLa, flLe
4 JLiz2 . fle fInLe, fLil, f LL, f hl
5 fLi . flale Tals
6 JLi . fhie fLil, fLl
rA ShLle . f hla lIa
8 fhLz . fli fle, fhl
9 fh .fhk fhh
TABLE IV

CONSTITUTION OF BARRED PLYMOUTH Rock FEMALES IN RESPECT
TO FECUNDITY

T ` T Probable Winter Egg
Class Zygote PS irk a ( G Produckag) Pennin ote And
Gametes Gametes Constitution

1 fInLe . Fhie LiLo, f hL? Fle, FLil

2 JLiLl . Flik 5 4 ala m fe 43 eggs

3 Jaiz . Fhe | fLil, fla Fhl, FLil Under 30 eggs

4 flak . Flat || flak FLilə Under 30 eggs

5 Shie . Fuh | Shh Fhill Zero eggs

6 Shi: . Fhe | Sule Fhlz Under 30 eggs

° The reason that gametes of the type fL,l, and fil, are not formed here
will be evident on consideration, Since no gametes of type FL, can, by

No. 552] INHERITANCE OF FECUNDITY 705

The different gametic constitutions in respect to fe-
cundity which are to be expected in Barred Plymouth
Rock males and females are shown in Tables III and IV.

Of these expected types six (1, 2, 3, 4, 7 and 8) were
found and used in the experiments in the case of the
males. In the case of the female class 5 birds were the
only ones not actually tested out in the breeding experi-
ments. Birds undoubtedly belonging to each of the
omitted classes have been reared in the course of the ex-
periments, but not yet submitted tô continued breeding
test.

The gametic constitutions of pure Cornish Indian
Games in respect to fecundity are given in Tables V
and VI.

TABLE V
CONSTITUTION OF CORNISH INDIAN GAME MALES IN RESPECT TO FECUNDITY
Class Zygote Gametes Produced
i fLil . fIale flile
2 fLi . fh flak, fll
3 fl . fll fhk

TABLE VI
CONSTITUTION OF CORNISH INDIAN GAME FEMALES IN RESPECT
“to FECUNDITY

i Probable Winter Egg
f-bearing F-bearing Production of Q
Class Zygote (ð Producing) | (Q Producing) Indicated Zygotic
` Gametes Gametes Constitution

1 JLi . Flak flail Flilk Under 30 eggs

2 Shh . Flak Shia, fLl | Flalz, Flies Under 30 eggs

3 Flak . Fhe flak, f hlk Fh, F Lil Under 30 eggs

4 fhlk . Fhe fli Flilz Zero

It will be noted that C.I.G. 9 classes 2 and 3 are gametically identical.
Both are left in the table, however, since the whole table is so short that no
confusion can be caused, and this example may make clear to some readers
the nature of the compression (by omission of duplicate classes) which was
practised in Tables III and IV.

How well this Mendelian hypothesis agrees with the
facts has been shown in detail in the complete paper. By
hypothesis, be formed this implies that an interchange of the factors E, and

between F and f gametes can not occur. The experimental proof of the
truth of this conviction has been furnished in the case of the inheritance of
the barred color pattern.

706 THE AMERICAN NATURALIST [Vou. XLVI

way of summary the following table shows the accord
between observation and expectation for all matings of
each general type taken together. For reasons set forth
below, the lumped figures do not give an altogether fair
estimate of the matter, but some sort of a summary is
necessary.
TABLE VII
SHOWING THE OBSERVED AND EXPECTED DISTRIBUTIONS OF WINTER EGG
PRODUCTION FOR ALL MATINGS TAKEN TOGETHER

A PENETRAN ROEN PERTRA ENA Pe
Winter Production of Daughters

Mating Class Over 30 | Under 30 | Zero r
A AT A A ee a l a
AN OLG. X O10... 14... FO Oe cicero = io
h a J 1 Gerad 25 | sers | 0.75
All F: and back-crosses.......... { eae a as sn nn

Considering the nature of the material and the char-
acter dealt with it can only be concluded that the agree-
ment between observation and hypothesis is as. close
as could reasonably be expected. The chief point in
regard to which there is a discrepancy is in the tendency,
particularly noticeable in the B. P. R. X B. P. R. and the
F, matings, for the observations to be in defect in the
‘* Over 30 ”’ class and in excess inthe ‘‘ Zero. ”’ class. The
explanation of this is undoubtedly, as has been pointed
out in the detailed paper, to be found in disturbing
physiological factors. The high producing hen, some-
what like the race horse, is a rather finely strung, delicate —
mechanism, which can be easily upset, and prevented
from giving full normal expression to its inherited
capacity in respect to fecundity.

The writer has no desire to generalize more widely
from the facts set forth in this paper than the actual
material experimentally studied warrants. It must be
recognized as possible, if not indeed probable, that other

” With exception of one set of matings discussed in full in the complete
paper.

No. 552] INHERITANCE OF FECUNDITY 707

races or breeds of poultry than those used in the present
experiments may show a somewhat different scheme of
inheritance of fecundity. The directions in which devia-
tions from the plan here found to obtain may, at least
a priori, most probably be expected are two. These are:
(a) differences in different breeds in respect to the abso-
lute fecundity value of the factors which determine the
expression of this character, and (b) gametic schemes
which differ from those here found either in the direction
of more or fewer distinct factors being concerned in the °
determination of fecundity, or in following a totally
different type of germinal reactions.

Regarding the first point, it seems probable from the
evidence in hand that the absolute fecundity value (i. e.,
the degree of actual fecundity determined by the presence

of the gametic factor) may differ for the factor L, in the

case of the Barred Rock as compared with the Cornish
Indian Game breed. It is hoped later to take up a detailed
study of this point, on the basis of the material here pre-
sented, and additional data now in process of collection.
Whenever there is a difference in the absolute fecundity
value of the L, factor, it means that the division point
for the classification of winter productions should be
taken at a point to correspond with the physiological
facts. Similarly, the absolute fecundity value of the
excess production factor L, may be different in different
breeds. In applying the results of this paper to the pro-
duction statistics of other breeds of poultry the possi-
bility of differences of the kind here suggested must
always be kept in mind.

The second point (the possibility of gametic schemes
for fecundity differing qualitatively from that found in
the present study) is one on which it is idle to speculate
in advance of definite investigations. I wish only to
emphasize that nothing is further from my desire or in-
tention than to assert before such investigations have
been made that the results of the present study apply
unmodified to all races of domestic poultry.

708 THE AMERICAN NATURALIST [ Vou. XLVI

It can not justly be urged against the conclusions of
this study that the Mendelian hypothesis advanced to
account for the results is so complicated, and involves
the assumption of so many factors or such complex inter-
actions and limitations of factors, as to lose all signifi-
cance. As a matter of fact the whole Mendelian inter-
pretation here set forth is an extremely simple one,
involving essentially but two factors. This surely does
not indicate excessive complication. To speak in mathe-
matical terms, by way of illustration merely, it may
fairly be said that the formula here used to ‘‘fit’’ the
data has essentially the character of a true graduation
formula. The number of constants (here factors) in the
formula is certainly much less than the number of ordi-
nates to be graduated.

There is no assumption made in the present Mendelian
interpretation which has not been fully demonstrated by
experimental work to hold in other cases. That the ex-
pression of a character may be caused by the coincident
presence of two (or more) separate factors, either of
which alone is unable to bring it about, has been shown
for both plants!! and animals by a whole series of studies
in this field of biology during the last decade. To find
examples one has only to turn to the standard hand-
books summarizing Mendelian work, as for example those
of Bateson and Baur. Again sex-linkage or correlation
of characters in inheritance has been conclusively demon-
strated for several characters in fowls by the careful and
thorough experiments of a number of independent inves-
tigators. Finally it is to be noted that Bateson and
Punnett! have recently shown that the inheritance of the
peculiar pigmentation characteristic of the silky fowl
follows a scheme which in its essentials is very similar to
that here worked out for fecundity.

“ Particularly important here are the brilliant researches of Nilsson-Ehle
on cereals, and of Baur on Antirrhinum.

_ "Bateson, W., and Punnett, R. C., ‘‘The Inheritance of the Peculiar
Pigmentation of the Silky Fowl,’’ Journal of Genetics, Vol. 1, pp. 185-203.

No. 552] INHERITANCE OF FECUNDITY 709

THE SELECTION PROBLEM

The results of the present investigation have an inter-
esting and significant bearing on the earlier selection ex-
periments on fecundity at this station. It is now quite
plain that continued selection of highly fecund females
alone could not even be expected to produce a definite
and steady increase in average flock production. The
gametic constitution of the male (in respect especially to
the L, factor) plays so important a part in determining
the fecundity of the daughters that any scheme of selec-
tion which left this out of account was really not ‘‘ sys-
tematic ’’ at all, but rather almost altogether haphazard.
It is repeatedly shown in the detailed account of these
experiments that the same proportion of daughters of
high fecundity may be obtained from certain mothers of
low fecundity as can be obtained from those of high fe-
cundity provided that both sets of mothers are mated to
males of the same gametic constitution. What gain is to
be expected to accrue from selecting high laying mothers
under such circumstances, at least so far as concerns the
daughters?

‘ Selection ”’ to the breeder means really a system of
breeding. ‘‘ Like produces like,’’ and “‘ breed the best
to get the best ’’; these epitomize the selection doctrine
of breeding. It is the simplest system conceivable. But
its success as a system depends upon the existence of an
equal simplicity of the phenomena of inheritance. If the
mating of two animals somatically a little larger than the
average always got offspring somatically a little larger
than the average, breeding would certainly offer the
royal road to riches. But if, as a matter of fact, as 1m
the present case, a character is not inherited in accord-
ance with this beautiful and childishly simple scheme,
but instead is inherited in accordance with an absolutely
different plan, which is of such a nature that the appli-
cation of the simple selection system of breeding could
not possibly have any direct effect, it would seem idle to

710 THE AMERICAN NATURALIST [ Vou. XLVI

continue to insist that the prolonged application of that
system is bound to result in improvement.

It seems to me that it must be recognized frankly that
whether or not continued selection of somatic variations
can be expected to produce an effect on the race depends
entirely on the mode of inheritance of the character
selected, In other words, any systematic plan for the
improvement of a race by breeding must be based and
operated on a knowledge of the gametic condition and be-
havior of the character in which improvement is sought,
rather than the somatic. Continued mass selection of
somatic variations as a system of breeding, in contrast
to an intelligent plan based on a knowledge of the gametic
basis of a character and how it is inherited, seems to me
to be very much in the same case as a man who, finding
himself imprisoned in a dungeon with a securely locked
and very heavy and strong door with the key on the
inside, proceeded to attempt to get out by beating and
kicking against the door in blind fury, rather than to
take the trouble to find the location of the key and unlock
the door. There is just a possibility that he could finally
get out in a very few instances by the first method, but
even in those cases he would be regarded by sensible men
as rather a fool for his pains.

Of course what has been said is not meant to imply that
selection, on the basis of somatic conditions may not have
a part in a well considered system of breeding for a par-
ticular end. In many cases it certainly will have. Thus
in the case of fecundity in the fowls, selection of mothers
on the basis of fecundity records is essential in getting
male birds homozygous with respect tò L, and L,. But
the point which seems particularly clear in the light of
the present results is that blind mass selection, on the
basis of somatic characters only, is essentially a hap-
hazard system of breeding which may or may not be
successful in changing the type in a particular case.
There is'‘nothing in the method per se which insures such
success, though that there is inherent potency in the

No. 552] INHERITANCE OF FECUNDITY 711

method per se is precisely the burden of a very great
proportion of the teaching of breeding (in whatever form
that teaching is done) at the present time.

It seems to me that it has never been demonstrated, up
to the present time, that continued selection can do any-
thing more than:

1. Isolate pure biotypes from a mixed population,
which contains individuals of different heredity constitu-
tion in respect to the character or characters considered.

2. Bring about and perpetuate as a part of a logical
system of breeding for a particular end, certain combina-
tions of hereditary factors which would never (or very
rarely) have occurred and would have been lost in the
absence of such systematic selection; which combinations
give rise to somatic types which may be quite different
from the original types. In this way a real evolutionary
change (i. e., the formation of a race of qualitatively
different hereditary constitution from anything existing
before) may be brought about. This can unquestionably
be done for fecundity in the domestic fowl. But here
“ selection ” is simply one part of a system of breeding,
which to be successful must be based on a definite knowl-
edge of gametic as well as somatic conditions. It is very
far removed from a blind “ breeding of the best to the
best to get the best.’? The latter plan alone may, as in
the case of fecundity, fail absolutely to bring about any
progressive change whatever.

It has never yet been demonstrated, so far as I know,
that the absolute somatic value of a particular hereditary
factor or determinant (i. e., its power to cause a quanti-
tatively definite degree of somatic development of a char-
acter) can be changed by selection on a somatic basis,
however long continued. To determine, by critical ex-
periments which shall exclude beyond doubt or question
such effects of selection as those noted under 1 and 2
above, whether the absolute somatic value of factors may
be changed by selection, or in any other way, 1s one of
the fundamental problems of genetics.

REFLECTIONS ON THE AUTONOMY OF BIOLOG-
ICAL SCIENCE

PROFESSOR OTTO GLASER
UNIVERSITY OF MICHIGAN

INTRODUCTORY

Ir the knowledge of facts and comprehension of prin-
ciples by certain writers had been adequate, and others
had. freed their minds from the survivals of animism,
the taxonomic position of biology in the scheme of knowl-
edge would appear uncertain to no one. Prolonged and
extensive inkshed however have surrounded this ques-
tion with much unnecessary difficulty and confusion.
Some claim that biology can not properly find a place
among the sciences at all; others, that if our science is
nothing more than physics and chemistry, it can have no
right to independent existence; and finally, the vitalists
postulate an absolute autonomy based on a specific prin-
ciple.

BIoLoGICAL PREDICTION

Merz, Enriques,’ and other present-day writers on the
systematics of biology dwell at length on the fact that
within the realm of the living, very strange and unex-
pected events take place. From the protozoans, human
beings can hardly be inferred: the chromosomal com-
plex, on account of the variations and surprising simi-
larities of its constituent elements, fails to tell us whether
we are dealing with sister species or with forms as re-
mote as snails, frogs, ferns and mice. Because one
crustacean is positively heliotropic, it does not follow
that the next one, even if the species be identical, will re-
spond in like manner, nor because one child in a family
has blue eyes can we conclude that its parents or broth-
ers and sisters have eyes of the same color. A dog or a

* Merz, John Theodore, ‘‘A History of European Thought in the Nine-
teenth Century’’; Enriques, Federigo, ‘‘ Probleme der Wissenschaft.’

712 ;

No. 552] AUTONOMY OF BIOLOGICAL SCIENCE Tic

man may be friendly to-day and vicious to-morrow under
similar external circumstances. Irregularities such as
these, our informants tell us, should quench the ardor of
the dullest, and to convince us still further of the inade-
quacy of our materials for science they point to rational
mechanics, a domain free from ambush, and pervaded by
an order in which only the foreseen and predictable find a
place.

The juxtaposition of these two disciplines is not only
unparliamentary, but unfortunate. Inasmuch as ra-
tional mechanics deals with abstractions and has only
the slightest objective basis, it can have no materials
comparable with the contents of any natural science.
On the contrary, it is a method of thinking. Thinking is
a phenomenon of consciousness; and consciousness, a
biological event. If, therefore, the mechanic produces an
orderly and coherent system in which one thing follows
with certainty from another, this shows nothing else
shan that certain biological events, to wit, mental proc-
esses, are among the most reliable phenomena in nature.

The biologist readily concedes that he is not as weather-
wise as the rational mechanic, but he does not concede
that this is due either to the fundamental disorderliness
of his section of nature, or because his colleague’s oracu-

lar powers differ in origin from his own where he happens
to possess them. As a whole man can not as yet be in-
- ferred from the protozoa, yet from the study of oxida-
tion, secretion and digestion in unicellular organisms we
could readily foresee the existence of these processes
in higher forms. The conditions of the heliotropic re-
sponse are such that an organism must be neither neu-
tral nor alkaline to react positively, and one at variance
with expectation can be made to do the expected by
acidulation. Although half the children of brown-eyed
parents may have blue eyes, this, instead of being a
symptom of disorder, is in strict conformity with a law
which enables us to say that two grandparents, one ma-
ternal, the other paternal, had this eye-color. The same

714 THE AMERICAN NATURALIST [Vou. XLVI

law makes it possible to predict the proportion and sex
of color-blind persons in a family in which this defect is
present. The change from friendliness to viciousness in
dogs and men has been traced to definite chemical and
structural changes so often that it could undoubtedly be
foreseen if these were known. The ‘embryologist fore-
tells the hour of ovulation, the obstetrician the birth of
a child, the entomologist the reappearance of a brood of
locusts, the ornithologist of a flock of birds, and the ichthy-
ologist of a school of fish, with the same reasonable cer-
tainty with which the celestial mechanic predicts the re-
turn of a comet. By the behavior of Convoluta roscoff en-
sis, even though far from its native haunts, it is possible
-= to tell the state of the tides. It should never be forgot-
ten that Weismann predicted the phenomena of matura-
tion in the germ cells.

Because the chromosomes have at present no taxo-
nomic importance, Merz concludes that they never can
have, and that biological events are therefore disorderly.
It so happens that the particular facts which Merz would
like to predict from these bodies are not related to what
we know about them in a manner so intimate that in the
present state of science prediction here would be any
more reasonable than in the absence of wind to judge the
weather from a bonfire. There are no reasons to doubt
that if we knew accurately the chemical structure of the
chromosomes, instead of merely their general composi-
sition, number, size and shape, we could tell the species,
and perhaps predict their composition in related spe-
cies, much as the organic chemist predicts the make-up
of one compound from another. Even now, the physio-
- logical state of the cell, and in numerous instances its
kind, as well as the sex of the individual from which it
was taken, can be determined from the chromosomal
complex.

How the rational mechanic acquired his prophetic
powers can be answered by considering the development
of geometry. Are we expected to believe that from the

No. 552] AUTONOMY OF BIOLOGICAL SCIENCE 715

qualities of a line, the geometrician could predict the
properties of the angle between two lines, if he had yet
to discover the possibility of angles? Knowing angles,
he could probably tell in advance not a few of the proper-
ties of triangles, but can any one imagine, on the basis of
this information alone, the relations which enable us to
measure the heights of trees we have never climbed, or
the distances of sun and moon? On the contrary, the
history of the subject shows that the mechanist is now
able to predict the motions of bodies, and the properties
of configurations, not because he deals exclusively with
prediction, but because he has made certain valid as-
sumptions concerning space, and by deduction has dis-
covered their consequences. He deals with controlled
materials, but the trick of augury has no other secret
than knowledge.

THE SPECIFICALLY BIOLOGICAL PROBLEM

If we reject the classification of biology necessitated
by a belief in the fundamental disorderliness of its phe-
nomena, two mutually exclusive views remain to be con-
sidered. Fortunately for the biologist the discord be-
tween them is quite unnecessary, for biology may be
physics and chemistry and autonomous at the same time.

Some of the most fruitful and illuminating discussions
in recent years have emanated from biological chemists
and physicists, and it is hard to follow the literature on
these subjects without sensing the enormous possibili-
ties with which it is freighted. It must not be supposed, -
however, that proof of the purely physical-chemical na-
ture of vital processes will show that living things are in
any way different than they really are. Whether analy-
sis can subtract qualities from things certainly seems
an idle question, yet we are constantly being told that the
reduction of the phenomena of life to a chemical-physical
basis will demonstrate that living things are, after all,
not alive! ee

Anatomical and histological analysis of a horse is 1m-

716 THE AMERICAN NATURALIST [ Von. XLVI

capable of showing that this animal is a cow. Even if we
reduce its tissues to their constituent chemical elements,
and, not content with this, continue until we have shown
that a horse is entirely composed of electrons, and their
activities, how could this show that a horse is not a horse?
If therefore resolution can detract nothing from the
things analyzed, it is clear that if these are in any way
unique, they will be no less so after this proeess than be-
fore. The only question which can be at issue is whether
living things are, or are not, unique. To this only an
affirmative answer is possible.

To reason with defectives is unprofitable for they have
no organ with which to perceive the qualities by which
we differentiate between the organic and the inorganic.
If we ask ourselves how we make this distinction we
naturally think of the fact that living things are ma-
chines with the power, as Loeb puts it, of automatic self-
preservation and reproduction. All the wonderful proc-
esses for which in the aggregate this simple formula
stands divide animals and plants sharply from matter
not alive and constitute the specific basis for the auton-
omy of our science. This autonomy is nothing meta-
physical, or absolute, but practical, like the autonomy of
physics, chemistry, astronomy and geology.

HistorrcaL BACKGROUND OF THE POSTULATED ABSOLUTE
AUTONOMY

In their analyses of living things, modern biologists
make use of only one practical method, but they apply it
from two distinct points of view, and since the signifi-
cance of phenomena in general depends on the point of
view, the whole meaning of the science hangs in the bal-
ance. The validity of these theoretical standpoints,
therefore, should be tested as carefully as the proposed
site of an observatory.

Unfortunately the issues at stake can not be properly
apprehended without some knowledge of their history.
To begin with Aristotle, and the few Ionians and Eleat-

No. 552] AUTONOMY OF BIOLOGICAL SCIENCE {LT

ics who preceded him, however, does not give us the
needed historical background, for the impression that
Aristotle was a primitive man, or that science was born
in Greece, is surely wrong. Scientific knowledge began
with the human race.

Although the thoughts of early men are for the most
part unrecorded, study of the primitive men living to-day
shows conclusively that the problem of the origin and
nature of life is realized by the savage. In the lore of
medicine men, magicians and seers, scientific knowledge,
theories and beliefs, fuse into an alloy which, despite the
varied conditions of its genesis and growth, presents
remarkable homogeneity. In this cultural amalgam the
attempt is made to explain the difference between a dead
man and a live one, by means of ‘‘a thin unsubstantial
human image, in its nature a sort of vapor, film or
shadow; the cause of life and thought in the individual
it animates; independently possessing the personal con-
sciousness and volition of its corporeal owner, past or
present; capable of leaving the body far behind to flash
swiftly from place to place; mostly impalpable and in-
visible, yet also manifesting physical power, and espe-
cially appearing to men waking or asleep as a phantom
separate from the body of which it bears the likeness;
continuing to exist and appear to men after the death of
that body; able to enter into, possess and act in the bodies
of other men, of animals and even of things.’’ *

These conclusions, drawn from the experience of
dreaming, are not much more primitive than the opin-
ions prevalent during the middle ages and surviving in
the shadows of church spires to-day. Now and again,
however, revolutionary teachings arose, and the most
significant of these for our immediate purposes are the
doctrines of René Descartes.

In his splendid history of biological theories, Rádl?
has traced with considerable detail the fortunes of the

2 Tylor, Edward B., ‘Primitive Culture.’’
? Rádl, Emil, ‘‘Geschichte der Biologischen Theorien.’’

718 THE AMERICAN NATURALIST [ Vou. XLVI

controversy set going in 1644 by the ‘‘Principes de la
Philosophie ’’ at a time when practically all men were
vitalists. During the seventeenth and eighteenth cen-
turies this contest engaged the ablest minds, yet mechan-
ism achieved no decisive victory, but only an increase in
the number of its followers, and the substitution of the
original soul in vitalism by the life force of Müller,
itself destined to elimination in the nineteenth century
by supersession, largely by neglect, and by direct experi-
ments on vital energetics.

Emil du Bois Reymond stands out as the champion of
mechanism during this period, although the limitations
of his materialism led him to classify the problem of life
with six other insoluble riddles. Lotze overthrew the
life force with arguments, substituted a purposeful pre-
formation in the germ, and protected it from further
harm by asserting that to inquire into its origin was un-
scientific. Fechner and Preyer attempted to clear the at-
mosphere by insisting that life is fundamental and the
real problem the origin of the inorganic. Virchow con-
tributed the idea of a mechanism superimposed upon
that already known, and this in the hands of his successor
Rindfleisch became a theory of atomic consciousness. In
the seventies, however, ghostly voices fell upon deaf
ears, for under the leadership of Darwin a seemingly
satisfactory natural explanation of adaptation forced
the mechanistic pendulum to its highest point.

While this period of scientific development proved
fatal to naturalistic vitalism, metaphysical not only sur-
vived, but during the latter-day Darwinian decadence
and reconstruction has again emerged, leaving behind
some of the erudities of its forerunners, and apparently
purged of ghosts. A change of names, however, does
not constitute a change of nature. The ghosts, more
rarefied than ever, are with us still, only to-day we call
them Entelechies, Dominants, Psychoids and Elan Vital.

No. 552] AUTONOMY OF BIOLOGICAL SCIENCE 719

Awnatysts oF Nro-VITALISM

Plate* finds in neo-vitalism four fundamental postu-
lates about which discussion must necessarily center.
These propositions are as follows:

I. Neither now nor in the future can the organism be
explained by chemistry and physics without a remainder.

II. There is an absolute distinction between dead and
living matter; in the inorganic world the law of causa-
tion holds, but in the organic causation holds together
with a unique law.

III. The uniqueness expresses itself in this, that every
organic process is final (teleological), that is, governed
by immanent purposefulness. |

IV. The cause of this finality, in so far as the vitalists
are not agnostic, is (a) a psychical factor; (b) a meta-
physical factor.

PosTULATE I |

Neither now nor in the future can the organism be
explained by chemistry and physics without a remainder.

Nothing could be more physical and chemical than the
analysis of the whole universe into a system of electrons.
When such resolution has been accomplished and every
known chemical element has been shown to be a special
case of corpuscular movement, the organic world and all
that characterizes it will be expressible in terms of elec-
trons if this mode of expression should appear service-
able. Would it not remain true, however, that hydrogen
is hydrogen, and oxygen, oxygen? Even if these gases
were proved to be configurations of essentially similar
corpuscles, they would nevertheless continue to be indi-
vidually different, and those so inclined would find it
possible to found separate sciences of hydrogenology
and of oxygenology, and these subjects would be auton-
omous. Does any one conclude from this that the me-
chanist is not fit to deal with these matters? Or that his
methods are fundamentally inadequate? Yet the argu-

*Plate, Ludwig, tt Darwinsches Selektionsprincip,’’ 3d ed.

720 THE AMERICAN NATURALIST [Vou. XLVI

ment of those who would cast mechanism out of biology
is identical. Resolution leaves intact uniqueness wher-
ever found, and the declaration that this is true of the
organism is a platitude.

PostūLATE IT

There is an absolute distinction between dead and liv-
ing matter; in the inorganic world, the law of causation
holds, but in the organic, causation holds together with
a unique law.

The second part of this proposition will be considered
in connection with postulate III. To the first part the
mechanist subscribes heartily, but adds that in his ex-
perience the distinction between hydrogen and oxygen
is equally absolute.

Postunate IIT

The uniqueness expresses itself in this, that every or-
ganic process is final (teleological) ; that is, governed by
immanent purposefulness.

In discussing postulate ITI, all that is needed is (a) to
sound its logical consequences; (b) to inquire how it
agrees with observations on individual and racial final-
ity; and lastly, (c) to expose the psychology of the teleo-
logical idea itself.

(a) From the harmony between the organic and the
inorganic, Driesch concludes that ‘‘nature is nature for
a purpose.’’ If the whole universe, however, is governed
by immanent purposefulness what becomes of the dis-
tinction between the organic and the inorganic? Ina
purposive system the teleological nature of any particu-
lar event or group of events can not be inferred, for pur-
posefulness can only be recognized by comparison with
purposelessness. Thus general teleology denies the ex-
istence of half the materials for the inference of the very
thing on which it bases itself, and with the best inten-
tions in the world, and without in any way seeming to
sense it, vitalists themselves have not only disarmed

No. 552] AUTONOMY OF BIOLOGICAL SCIENCE 121

teleology in the realm of the living, but have made the
principle scientifically impossible.

(b) Were every organic event final or purposeful,
functional adjustment, training and education would be
unnecessary and impossible. Jennings? tells us:

How the relations that impress us as teleological were brought about,
constitutes undoubtedly a set of most difficult problems. But to keep
us from despairing, we find this process taking place in the lives of
individuals in a manner that can readily be studied. This is in the for-
mation of habits. In the formation of habits, we see that the organism
at first does not react in a way that impresses us as teleological, while
later it does, and we can watch the process change from one condition
to the other, and discover how it is causally determined. Since then a
method of action that appears to us teleological is produced in an
intelligible way under our very eyes, in the lifetime of the individual,
there is no reason why we may not expect to find out how teleological
relations have been brought about in the life of the race when we have
actually made a start in the study of the physiology of racial processes.

past ” reappears again in the future.

The ability to make functional adjustments of this
character is only a special case of automatic self-preser-
vation, and is found in all organisms because those de-
void of it are for this very reason eliminated and conse-
quently remain largely unknown. Paleontology 1s the
science that deals chiefly with these failures. How many
organisms have been unable to make the necessary ad-
justments is attested by the great number of extinct ani-
mals and plants; how many are failing to-day is shown
by every rapidly vanishing species, as well as by many
experiments and special observations. Several of the
mutants of de Vries have for one reason or another

‘c Diverse Ideals and Divergent Conclusions in

i :
Jennings, Herbert S., 1» American Journal Psychol-

the Study of Behavior in Lower Organisms,
ogy, Vol. XXI. :

€ Loeb, Jacques, ‘‘The Mechanistic Conception of Life,
Vol. LXXX.

”? Pop. Sci. Mo.,

722 THE AMERICAN NATURALIST [ Vou. XLVI

proved indurable, whereas Loeb* has pointed out that
faulty organisms must frequently arise, although we only
become aware of them under exceptional conditions.

Moenkhaus found ten years ago that it is possible to fertilize the
egg of each marine bony fish with sperm of practically any other marine
bony fish. His embtyos apparently lived only a very short time. This
year I succeeded in keeping such hybrid embryos between distantly
related bony fish alive for over a month. It is therefore clear that it
is possible to cross practically any marine teleost with any other.

The number of teleosts at present in existence is about 10,000. If
we accomplish all possible hybridization 100,000,000 different crosses
will result. Of these teleosts only a very small proportion, namely,
about one one-hundredth of one per cent., can live. It turned out in
my experiments that the heterogeneous hybrids between bony fishes
formed eyes, brains, ears, fins and pulsating hearts, blood and blood
vessels, but could live only a limited time because no blood circulation
was established at all—in spite of the fact that the heart beat for
weeks—or that the circulation, if it was established at all, did not last
long.

The possibility of hybridization goes much further than we have
thus far assumed. We can cause the eggs of echinoderms to develop
with the sperm of very distant forms, even mollusks and worms
(Kupelwieser) : but such hybridizations never lead to the formation of
durable organisms.

It is therefore no exaggeration to state that the number of species
existing to-day is only an infinitely small fraction of those which can
and possibly occasionally do originate, but which escape our notice
because they can not live and reproduce. Only that limited fraction
of species can exist which possesses no coarse disharmonies in its auto-
matie mechanism of preservation and reproduction. Disharmonies and
faulty attempts in nature are the rule, the harmonically developed
systems the rare exception. But since we only perceive the latter we
gain the erroneous impression that the “ adaptation of the parts to the
plan of the whole” is a general and specifie characteristic of animate
nature, whereby the latter differs from inanimate nature.

the structure and the meclianism of the atoms were known to us
we should probably also get an insight into a world of wonderful
j harmonies and apparent adaptations of the parts to the whole. But in
this case we should quickly understand that the chemical elements are
only the few durable systems among a large number of possible but not
durable combinations.

(c) Overlooking for the moment the obvious difficul-
ties of the assumption, we can be certain that the idea of

No. 552] AUTONOMY OF BIOLOGICAL SCIENCE 123

teleology would never have entered the biologist’s head `

were he not himself a living thing. Since this is the case,
however, his interest in life exceeds all others, and he
attends to the processes that make life possible only be-
cause of their resultant. Inasmuch as the latter occu-
pies the focus of his mind, he wrongfully reasons back-
ward from results to processes, and finding in these none
that might have rendered the cherished product im-
possible, concludes that the processes were all along
aiming at what, from his standpoint, is the end. Clearly
the conclusion has only an anthropocentric basis.

PostuLaTE IV

The cause of this finality, in so far as the vitalists are
not agnostic, is (a) a psychical factor; (b) a metaphys-
ical factor.

Since biological finality is an anthropomorphism, a
discussion of the supposed teleological factors is futile.
Inasmuch, however, as psycho-vitalism has its counter-
part in psycho-mechanism, the fallacy common to both
may be pointed out.

(a) To reflect mind oi the cell, and so reflected to
use it as an explanation of what the cell does, is the
method of primitive animism. Quite apart from the fact
that the existence of mind, so far, at least, has been dem-
onstrated only in the case of certain higher animals, but

not at all for the lower, or the developmental stages of _

the higher, as an explanation it can have no title to
serious consideration since it is itself one of the elements
of the automatic self-preservation which it is the aim of
biology to analyze. To interpret something we do not
understand in terms of something else which at present
we understand even less, may give temporary comfort to
some minds, but the ideals of scientific explanation call
for the reverse process.

(b) The difficulties of Driesch’s style are > such that
many biologists refuse to read his books. For this rea-

kd

d

ii. Slee THE AMERICAN NATURALIST [ Vou. XLVI

son I have made from one of them’ a series of extracts
to serve as illustrative material. The italics are not
mine.
DriescH’s ENTELECHY
Entelechy or the psychoid has nothing of a “ psychical” nature.

We indeed are in a rather desperate condition with regard to the
real analysis of the fundamental properties of morphogenetic, adaptive,
and instinctive entelechies: for there must be a something in them
that has an analogy, not to knowing and willing in general, as it
may be supposed to exist in the primary faculties of psychoids, but to
the willing of specific unexperienced realities, and to knowing the
specific means of attaining them. (P. 142.

To build up the organism as a combined body of a typical style is
the task of entelechy; entelechy means the faculty of achieving a “ forma
essentialis ”; being and becoming are here united in a most remarkable
manner; time enters into the Timeless, i. e., into the “idea” in the
sense of Plato. (P. 149.) _

There is first the entelechia morphogenetica, and after that the
entelechia psychoidea and the latter may be discriminated as governing
instinets and actions separately. Furthermore the different parts of the
brain, such as the hemispheres and the cerebellum in vertebrates, may
be said to possess their different kinds of entelechy.

In fact we may speak of an order concerning the rank or dignity of
entelechies, comparable with the order of ranks or dignities in an army
or administration. But all entelechies have originated from the pri-
mordial one and in this respect may be said to be one altogether.

Now the primordial entelechy of the egg not only creates derived
entelechies, but also builds up all sorts of arrangements of a truly
mechanical character; the eye, in a great part of its functioning is
nothing but a camera obscura, and the skeleton obeys the laws of inor-
ganie statics. Every part of these organie systems has been placed by
entelechy where it must be placed to act well in the service of the whole,
but the part itself acts like a part of a machine.

So we see finally that the different forms of harmony in the origin
and function of parts that are not immediately dependent on one
another, are in the last resort the consequence of entelechian acts. The
entelechy that created them all was harmonious in its intensive mani-

oldness; the extensive structures which are produced by it are therefore
hiornoniits too. In other words there are many processes in the

organism which are of the statical- teleological type, which go on
ee or purposefully on a fixed machine-like basis, but entelechy
esch, Hans, ‘‘The Science and Philosophy of the Organism,’’ Gif-

ford Teia, 1908.

No. 552] AUTONOMY OF BIOLOGICAL SCIENCE 725

has created this basis, and so statieal teleology has its source in dynam-
ical teleology.

e now see the full meaning of the statement that entelechy is an
“intensive manifoldness” realizing itself extensively; in other words,
we know what it means to say that a body in nature is a living organ-
ism; we have given a full descriptive definition of this concept. (Pp.
150-151.)

Any single spatial occurrence induced or modified by entelechy has its
previous single correlate in a certain single feature of entelechy as far
as it is an intensive manifoldness. (P. 154

Entelechy may be aroused to sipnifeckation by a change in bodily
nature, such as is effected by fertilization, or by some operation, or by
some motor stimulus; on the other hand, entelechy may on its own part
lead to changes in bodily nature. (P. 156.)

It is the essence of an entelechy to manifest itself in an extensive mani-
foldness: all the details of this extensive manifoldness depend upon the
intensive manifoldness of the entelechy, but not upon different spatial
“ causes.” 15

Entelechy lacks all the characteristics of quantity; entelechy is order
of relation and absolutely nothing else; all the quantities concerned in
its manifestations in every case being due to means which are used by
entelechy, or to conditions which ean not be avoided. (P. 169.)
Entelechy, as far as we know, at least, is limited in its acting by many
specificities of inorganie nature, among which are the specificities
included under the phrase “ chemical element.” (P. 179

Entelechy is also unable to cause reactions between chemical compounds
which never are known to react in the inorganic world. In short
entelechy is altogether unable to create differences of intensity of any
kind.

But entelechy is able, so far as we know from the facts concerned in
restitution and adaptation, to suspend for as long a period as it wants
any one of all the reactions which are possible with such compounds as
are present, and which would happen without entelechy. (P. 180.)
Entelechy though not capable of enlarging the amount of diversity of
composition of a given system, is capable of augmenting its diversity
of distribution in a regulatory manner, and it does so by transforming
a system of equally distributed ener into a system of actualities
which are unequally distributed. (P. 1
Entelechy . . . is a factor in nature ia acts teleologically. It is an
intensive manifoldness and on account of its inherent diversities it is
able to augment the amount of diversity in the inorganie world as far
as distribution is concerned. It acts by suspending and setting free
reactions based upon potential differences asi There is noth-
ing like it in inorganic nature. (P. 205

726 THE AMERICAN NATURALIST [ Vou. XLVI

Entelechy is an elemental factor of nature conceived to explain a certain
class of natural phenomena. (P. 206.
You may say if you like that entelechy, when turning a mass particle,
acts upon it at right angles to its path—this kind of action requiring
no energy, but even thus there would be only a pseudo-obedience to the
laws of real mechanics, since entelechy must be regarded here as non-
energetical and as interfering with inertia at the same time. (P. 223.)
Entelechy is affected by the accomplishment of its own performance, in
acting as well as in morphogenesis. (P. 228.
In order that adaptation may happen, the fundamental state of the
organism must be disturbed in its normality; this fact affects or calls
forth entelechy. (P. 229.)
Entelechy is affected and thus called into activity by changes of any
normality governed by it which are due to external causes and these
changes do not affect entelechy as a mere sum of changed singularities,
but as changes of normality as a whole. (P. 232.)
Entelechy is affected by and acts upon spatial causality as if it came
out of an Br dimension; it does not act in space, but it acts
into space. (P. 235.)
Entelechy is an agent acting manifoldly without being itself manifold
in space or extensity. Entelechy then is only an agent that arranges,
but not an agent that possesses quantity. (P. 250.)
Entelechy is something different from matter and altogether opposed to
the causality of matter. (P. 255.)
May not entelechy be called a “substance” in the most general philo-
sophical sense of the word, that is, in the sense of a something irredu-
cible, which remains the always unchangeable bearer of its changeable
qualities. (P. 256.)
Entelechy has the power of preserving its specifie intensive manifold-
ness in spite of being divided into two or more parts. (P. 257.)
Entelechy therefore can not possess a “ seat.”
At present the question whether entelechy is a “ substance ” must remain
as open as the previous question about the relation of entelechy to
causality. . . . Entelechy was a kind of “ quasi” causality, and now
may be said to be an enduring “ quasi-substance.” (P. 260.
Entelechies, though transcending the realm of the Imaginable, do not by
reason of their logical character as such form constituents of meta-
physies in the sense of something absolute and independent of a sub-
ject. (P. 320.)
Entelechy is alien not only to matter but also to its own material
purposes. (P. 336.)

Mir wird vor alle Dem so dumm

Als ging mir ein Mithlrad im Kopf herum!

No. 552] AUTONOMY OF BIOLOGICAL SCIENCE 727

CONCLUSION

I have tried to show that biological events are orderly;
that a distinct problem guarantees the autonomy of the
science; that the application of physical and chemical
methods has no shortcomings specifically different from
those met with when applied to the inorganic, and finally
that vitalism in addition to being unnecessary is absurd.
The question whether the modern outburst of metaphys-
ical biology, a movement which finds favor among phi-
losophers and psychologists, and has no small following
among zoologists and botanists, is not, despite its obvious
faults, sound in motive, remains to be answered. Me-
chanical methods, even if applicable to vital events no
less than to any others, might nevertheless possess an
inherent weakness discoverable only when enlisted in
biological service. The only reply possible to this ques-
tion is that they are the best methods which human
beings can devise, for their excellencies are grounded in
our structure, their deficiencies in that of the world out-
side.

It has been pointed out over and over again that the
explanations of science never amount to more than the
enumeration of the conditions under which the events
in nature take place. With ultimate explanation science
does not deal, not because men of science do not want to,
but because in their experience nature contains nothing
ultimate. The failure, therefore, of science to give us
more than it does can be held up as a fault only by those
who are dissatisfied with the structure of the universe.
For this feeling intellectual hygiene is the only cure.

If the limitations of scientific methods are to be found
in the limitations of a limitless universe, their excellen-
cies, as instruments for the automatic preservation of
life, are to be found in ourselves, for the mechanical
symbols by the aid of which natural phenomena are in-
terpreted are the easiest for us to use. The value of
these symbols depends on our power to visualize, and

728 THE AMERICAN NATURALIST [ Vou. XLVI

visualization depends on sight. Is it without signifi-
cance in this connection that the eye begins in the embryo
earlier than any other receptor of special sense, or that
sight, except perhaps by a few poets and musicians, is
acclaimed the most priceless of all our senses?

If we lived in a world of phantasms, the value of sight
would largely disappear, for, as Berkeley® has pointed
out, it is an organ of anticipatory touch upon which de-
pends our ability to avoid harmful collisions, and to
bring about desirable ones. From the very beginning
of our lives we see and deal with visible objects. Is it
strange then that we should attempt to express all our
experience in terms of the language which by our very
structure and history is the most used and hence the most
efficient medium of interpretation we possess?

Modern energetics has indeed discarded solid mole-
cules and atoms, and has replaced these by constellations
of electrons, yet even if the electrons are nothing more
than electrical charges, they are believed to possess mass,
and to have certain properties in common with visible
things. Does not the physicist still draw pictures on the
wall to make clear what he means? Is nota picture a vis-
ual symbol by the aid of which we understand a less fa-
miliar one? Escape is impossible, for mechanistic sym-
bolism is grounded in our very nature, and for this rea-
son its employment rises to the dignity of amoral act, for
it involves neither more nor less than the application of
our best capabilities to the best of all purposes—the in-
terpretation of nature.

* Berkeley, George, ‘‘An Essay Towards a New Theory of Vision.’’

THE SPAWNING HABITS OF THE SEA LAMPREY,
PETROMYZON MARINUS!

DR. L. HUSSAKOF
AMERICAN Museum or NATURAL HISTORY

THE spawning habits of several species of lamprey are
known from observations which have been made in both
Europe and America.?_ Those of the sea lamprey, Petro-
myzon marinus, however, have not been studied, not-
withstanding that this is the largest of the lampreys and
is common to both sides of the Atlantic. It is merely
known that this species ascends rivers for the purpose of
spawning; and that the ‘‘fish’’ transport stones in build-
ing their nest much like other lampreys (Burroughs,
’83; Holder, ’85). In 1883 a French observer, L. Ferry
(83), noted the development of sea lamprey eggs taken
directly from a female specimen. He concluded that the
eggs must already have been fertilized, and hence that
fertilization in the lampreys is internal. This conclusion,
in the light of the careful observations on the spawning
of various lampreys, especially Petromyzon planeri and
Lampetra wilderi, is undoubtedly erroneous. Moreover,
the discovery that lamprey eggs can develop partheno-
genetically (Bataillon, ’03), affords a simple explana-
tion of the facts recorded by Ferry. None the less the
observation of the breeding habits of the sea lamprey ©
was very desirable.

The observations recorded in this paper were made by
the writer on Long Island, June 1 and 2, 1911, while col-
lecting material for a group to represent the nesting
habits of the sea lamprey in the American Museum of
Natural History. The locality, Smithtown, on the Nis-
sequogue River, Long Island, was suggested to me by

*Read before the American Society of Zoologists, at Princeton, N. J.,
Dee. 27, 1911.

3 See annotated bibliography at end of paper.
- 129

730 THE AMERICAN NATURALIST [ Vou. XLVI

Professor Bashford Dean, who had learnt of it through
Dr. Tarleton H. Bean.

Locality and Date of Observation.—At the date of
these observations, June 1 and 2, 1911, lampreys had been
seen in the Nissequogue River for several days. A num-
ber of abandoned, partly scattered nests were also to be
found; hence June 1, appears to be toward the end of the
spawning season, which for Long Island must be put
down as the latter half of May.

Fic. 1. SEA LAMPREYS, Petromyzon marinus, on Nest. An exhibition
group, 4 by 5 feet, in the American Museum of Natural itary: prepared under
the piip of the writer.

The Nissequogue is a small stream which empties into
Long Island Sound. At the village of Smithtown, three
and a half miles from its mouth, it is shallow (a foot
or two deep), perfectly clear, and flows over a bed of
large, water-stained pebbles. Here and there are patches
of ‘‘river grass.’? The water is perfectly fresh here,
although still affected by the tide. A quarter of a mile
above and below the village bridge, the river grows

No. 552] THE SEA LAMPREY 731

deeper and muddier; my observations were therefore
confined to the pebbly portion, a stretch of about half a
mile. Here a dozen nests were found, four with lam-
preys on them, the others deserted and partly scattered
by the tide.

Nests——The nest of the sea lamprey is similar to
that of other species, but much larger. It is a circu-
lar depression in the river bed, two to three feet in
diameter. One that was measured was two feet three
inches across, and six inches deep in the center. The

Fie THREE SPECIMENS OF Petromyzon marinus ON A Nest. Instantaneous

photograph taken without special apparatus A bright sunlight, at low tide; with

y three or four inches of water above the “fish.” One male and two females.
(Net EE

nests are easily recognized, eyen at a distance of
several feet, by the large number of whitish quartz peb-
bles which have been uprooted and turned with their
clean faces up. They are built at random anywhere in
the river: near the bank, in the shade of overhanging
trees; in the middle of the stream, exposed to the glare
of the sun; or even, as with the nest shown in the figures

732 THE AMERICAN NATURALIST [Vou. XLVI

(Figs. 1 and 4), partly under a log. Occasionally two
nests adjoin so that their peripheries overlap.

Standing in the water close to a nest, one may observe
minutely every movement the lampreys make. One may
even stroke them or lift them by the tail without disturb-
ing them. A ‘‘fish’’? must be raised to a considerable
angle before it will loosen its hold on the stone to which
it clings, and dart away; and then it will go only a short
distance, fifty or a hundred feet, and seek refuge under
the ‘‘river grass.”

The manner of building the nest is quite like that of the
brook lamprey (Lampetra wilderi), as de-
scribed by Gage (’93), and by Dean and
Sumner (’97). But owing to the large
size of the species all the processes are
writ large, as it were, so that one can see
the purpose of every movement. Build-
ing the nest consists in carrying the
pebbles and stones out of a circular area
until a basin-like depression is formed.
As the work proceeds the finer material
in the interstices between the pebbles
gradually accumulates, so that the bottom
of the nest becomes covered with sand and
fine gravel. The stones are seized with
the circular mouth to which they cling en-
tirely by suction. The ‘‘teeth’’? play no

Fic. 3. Freshly part in this work, as may be proved by ex-
He amprey cdas perimenting with the freshly dead ‘‘fish.’’
the vacuum pro- By pressing the mouth of such a ‘‘fish’’

a agian against a stone, it may be made to hold on
pated i "so tenaciously, that by lifting the stone

one lifts the fish (Fig. 3). A vacuum is
produced inside the buccal funnel, and this is the imme-
diate cause of the hold. In carrying stones out of the
nest, the procedure varies with the size of the stone.
Small stones, an inch or two across, are picked up in the

p.
et
°
f
a
S
=]
®©
>
“

No. 552] THE SEA LAMPREY 733

mouth and carried out. Larger stones, firmly rooted in
the bottom, require considerable effort to be dislodged;
the stone is tugged upward, the lamprey receding back-
ward in a straight line. Sometimes instead of pulling
backward, the lamprey charges head-on and pushes the
stone in front of it up the incline, the body remaining
rigid and acting as a lever, while the tail is lashed vio-

Fie, 4. Sa THREE SPECIMENS AS SHOWN IN FIG. 2, PHOTOGRAPHED
SHORTLY AFTER THEY WERE TAKEN FROM THE Nest. Upper two, females;
lowest one, a male. The stones in lower half of photograph were picked up
just as they were carried by the lampreys out of the nest. The half brick
shown in the picture weighs 840 grams.

lently to gain a firmer support. Some of the larger
stones carried by the lampreys out of the nest (Fig. 4),
were picked up just as they were released from the
mouth; they were found to weigh (in air) from 145 to
840 grams. `
The building and improving of the nest go on continu-
ously between intervals of mating. Both the male and
the female take part in this work. On one nest there
were observed a male and a female; they were joined now
and then, for some minutes, by a second female which

734 THE AMERICAN NATURALIST [ Vou. XLVI

had been for some time by herself on an adjoining nest.
During the time the three ‘‘fish’’ were on the one nest
(Figs. 2 and 5) they all took part in repairing it, in
the intervals between mating—the male apparently not
distinguishing between the female of his own nest and
the intruder. In two other cases there was one indi-
vidual to a nest. After carrying a stone out the lamprey
immediately returns for another. This is repeated a
number of times and then the lamprey clings to a stone
apparently exhausted (Fig. 4). Now and then the tail is
lashed against the sides of the nest to pad it down. When
on a nest by itself, a lamprey occasionally wanders a
distance of some feet—even several hundred feet—but
invariably returns to continue its nest-building. These
wanderings are perhaps for the purpose of finding a
mate.

Mating.—The method of copulation is similar to that
of the brook lamprey, Lampetra wilderi, as described by
Dean and Summer (’97); and it is unnecessary to re-
describe it here. I will merely comment on a few de-
tails. The female must cling to a large stone in the
nest in order that copulation take place. The male
seizes her by the top of the head. In copulo, the two are
arranged so as to form an ellipse. The caudal portion
of the male is applied immediately back of the first dor-
sal of the female, and curved in a loop around her body.
Several authors have referred, in the case of both Amer-
ican and European lampreys, to the vibration of the
posterior portions of the ‘‘fish,’’ in copulo. In the sea
lamprey this vibration may be observed very closely. It
lasts two or three seconds. It begins slowly, gradually
increases in frequency until it reaches an exceeding ra-
pidity of vibration, then subsides by a few slow beats.
The motion strongly suggests the vibration of a rattle-
snake’s tail in the warning pose. Indeed while watching
the lampreys one can hardly keep from imagining the
sound which ought to accompany the lampreys’ vibra-

No. 552] THE SEA LAMPREY 735

tions, so similar is the movement to that of the rattle-
snake’s tail.

As to the length of time ‘‘fish’’ on a nest continue to
spawn, I was able to make some observations. The
nest shown in Figs. 1 and 4 was observed continuously
for over four hours, from 10 a.m. until after 2 p.m.; and
during that time copulation took place at intervals
of from a few to ten minutes; and in all probability
would have continued several hours longer had the

THE SAME NEST AS IN FIG. 2, SHOWING TWO OF THE SPECIMENS in
y
the female. The second female is seen clinging to a stone. (Not r AE EERE

‘“‘fish”’ been left on the nest. Both became gradually
more and more scarred from seizing each other with
their mouths: round pale wounds stood out clearly
against the blue-black of the head of the female where
the male had repeatedly seized her; and large whitish
wounds could be seen on her back, especially posterior
to the first dorsal fin. The male likewise was scarred in
several places on the head and back. These scars are

736 THE AMERICAN NATURALIST [ Vou. XLVI

greatly augmented through continual rubbing against
stones, padding down the nest, ete.

Fate of Sea Lampreys After Spawning—From the
facts at hand, it appears that lampreys that go up the
river to spawn do not again return to the sea, but die
shortly after spawning. I found two dead, badly
scarred, spent lampreys in the river not far from de-
serted nests. One was in the shade of tall grass near
the bank, the other tangled in weeds and twigs in the
middle of the river. Both had been rasped, apparently
for food, by other lampreys. Burroughs (’83), also re-
cords that it is not unusual to find dead lampreys in
June.

The causes of the death of these lampreys—and in-
deed of all anadromous fishes—are still rather obscure.
Death is probably chiefly due to the cycle of katabolic
processes initiated on the maturing of the gonadial
products. Besides this at least two other causes must
be regarded as contributary: first the greatly lessened
vitality due to the constant exertion in uprooting and
transporting stones. Lampreys thus weakened become
the prey of other lampreys. Secondly, the numerous
sears or wounds which they inflict on one another in
mating allow ‘‘fungus’’ to invade, and ultimately to de-
stroy, their tissues. All three of these causes, probably,
play a part in causing the death of the sea lampreys after
spawning.

Remarks on the Senses and Mentality of the Sea Lam-
prey.—The behavior of the lampreys was carefully noted
to see in how far their senses come into play in building
the nest and in other activities. The general impres-
sion was that the sea lamprey is guided by touch more
than by any other sense. Sound does not disturb them
while on the nest: one may carry on a conversation right
over a nest without in the least affecting them. Indeed
one lamprey that was building a nest under a wooden
bridge was not disturbed by the clattering of automo-
biles over it. This insensitiveness to sound may mean,

No. 552] THE SEA LAMPREY Tal

however, merely absorption in the work of nest-build-
ing, and not that the lampreys are insensible to this
stimulus.

The eyes of lampreys in the water shine like black
beads; but they are not very sensitive. If one meets a
lamprey swimming toward him in the river, it will come
almost right up before it will discover the person and
turn aside.

Many of the movements of the sea lamprey on the
nest are purposeless—as was noted also for the brook
lamprey by Dean and Summer (’97). Thus a lamprey
will sometimes pick up a stone outside the nest, carry
and drop it into the nest; or while carrying out a stone,
will drop it half way up the side of the nest. It will tug
at a large stone which it cannot possibly dislodge, or at
a log, in an effort to drag it out of the nest, and will
repeat this again and again, without profiting in the least
by previous failures. On the whole one has a feeling
that the lamprey possesses a very low mentality even
as compared with fishes.

ANNOTATED BIBLIOGRAPHY
The following list includes all the papers I have been able to find that
deal with the spawning habits of lampreys. Those having to do with the
cytology of fertilization are listed in Ziegler’s text-book referred to below.

Bataillon, E.

1903. La segmentation parthénogénétique expérimentale chez les œufs

de Petromyzon planeri. Comp. Rend. Acad. Sci., Vol. 137, pp.
79-80.

Carried embryos as far as the blastula stage—‘‘en plongeant
et maintenant les œufs dans des solutions de saccharose à 5 ou
6 pour 100 ou dans des solutions isotoniques de NaCl.’’

Burroughs, John.
1888. A lamprey’s nest. The Century Magazine, XXV, p. 457.

Observed Petromyzon marinus spawning in a creek. Describes
their mode of transporting stones; their indifference to an ob-
server at close range; but does not describe the nest, nor appar-
ently recognizes the purpose for which the stones are transported.
The female is the larger of the two. ‘‘In June it is not unusual
to find their dead bodies in the streams they inhabit.’’

738 THE AMERICAN NATURALIST [Von XLVI

Dean, Bashford, and Sumner, Francis B.
1897. otes on the Meroe habits: of the brook lamprey (Petro-
n wilderi).. Trans. N. Y. Acad. Sci., XVI, pp. 321-324,
PL VII.
Detailed observations on manner of building nest, on copula-
ion, and general behavior. Contains the best figure extant o
lampreys on nest.
Ferry, L.
1883. Sur la lamproie marine. Compt. Rend. Acad. Sci., pp. 721-723.
Translated, in part, in Ann. Mag. Nat. Hist., (5), II, p. 388
(May, 1883
A female sea lamprey, taken by a fisherman while i to
his boat, was opened, and its ova put into a large bas It
rained at the time and the basin became partly filled with eae
After twenty days the eggs hatched into perfect larve. Ferry
era ‘‘Il resort de ce fait que les œufs pris dans le ventre
a Lamproie étaient déja ng ndés et avaient dû l’étre dans
‘Sateen de l’animal’’ (p. 722).
Gage, Simon H., and Meek, Seth E.
1886. The lampreys of Cayuga Lake (abstract). Proc. Amer. Assoc.
Adv. Sci, XXXV, p. 269.
Brief tafermnes to nests and to length of breeding season,
which is said to last ally eae months (May and June).
Gage, Simon H.
1893. The lake and brook lampreys of New York, especially those
of Cayuga and Seneca: Lakes. The Wilder Quarter-Century
Book. Ithaca. Pp. 421-493, Pls. I-VII.
Gives extended account of nest building and spawning of the
brook lamprey renee wilderi) and of the lake lamprey
(Petromyzon natar a color); discusses fate of lampreys

after spawning (pp. 4 Man). Figures two lake ATESA on a
nest, carrying aie vii, Fig. 39.
Herfort, Karl.
01.

Die Reifung und Befruchtung des Eies von Petromyzon fluvia-
tilis. Arch. ae mikrosk. Anat. u. Entwickl., Vol. 57, pp. 54-
95, Pis.
ape review ca some European papers on spawning habits
p. 55-57). For his own studies he fertilized the eggs arti-
fici aiy = 57-58).
Holder, Charles Frederick.
1885. The Abi aie eel and nest. In Marvels of Animal Life; 8°; New
York; pp. 5-8
Describes how Petromyzon marinus transports stones, but mis-
takes the mass of stones accumulated outside the nest, for the
nest itself. Quotes from an account of over fifty lampreys build-
ing near a dam in the Saco River, Maine. They dropped the
stones on the dam until it became covered over. This was errone-
ously interpreted to mean that the lampreys had built the dam.
A stone is sometimes carried by two lampreys.

No. 552] THE SEA LAMPREY 739

Kupper, Carl v., and Benecke, B.
1878, Der Vorgang der Befruchtung am Ei der Neunaugen. Fest-
schrift fiir Theodor Schwann. Kénigsberg.
Loman, J. C.
1910. De copulatie van Petromyzon planeri. Tijdschr. nederl. dierk.
ereen, (2), XII, p. vii.

A careful puniki of the spawning of Petromyzon planeri,

observed in a brook near Königsberg.
McClure, Charles F.
1893. Notes on the mad stages of segmentation in Petromyzon

marinus L. (americanus Le S.). Zool. Anz., XVI, pp. 367-

368; 373-376.

Collected thirty specimens from a river near Princeton, New
Jersey, between May 20 and June 1.

Müller, August.
1856. Ueber die Entwickelung der Neunaugen. Ein vorläufiger
ericht. Arch. f. Anatomie, Physiol. u. wissen. Medicin, Jahrg.

1856, pp. 323-339.

This is the earliest account of the breeding habits of the lam-
prey. The observations were made on the brook lamprey; some-
times ten or more were seen together tugging at, and transporting
stones. The object for which the stones were carried was, how-
ever, not made out. He made the discovery, by rearing Ammo-

cœtes, that they are the larve of lampreys.
F, Jaco
1903. An experimental study of the spawning behavior of Lampetra
Science, N. S., XVII, p. 529 (abstract).
“a constant relation between individual fish and indivdual
is determined not by character of

ir natural ee iat

: S. Bureau Fisheries, XXVII, pp. 43-68, Pls. iii-v.

Photograph of four Lampetra wilderi on a nest—P1. iii, Fig. 2.

Vejdovsky, F.
1893. O trenf mihule (Petromyzon planeri) [Die äussere Befruchtung
des Neunauges]. Sitzber. & königl. bohm. Gesell. Wiss. in

Prag. Math.-naturw., 1893, article XLIX, PL xviii.

Observed spawning process In an aquarium containing four
males and one female. Figures two specimens (Pl. xviii) in
ade but this figure does not represent them in their charac-
teristic attitudes.

ay Vieira, Lopes

= 1831. Remarks on the eggs and spawning-season of Petromyzon

| fluviatilis Linn. Ann. de Sciencias Natures, I, pp. ag !
Sea lamprey enters rivers of Portugal at end of December an

beginning of January.

740 THE AMERICAN NATURALIST [ Vou. XLVI

Yarrell, W.
1831. PRE on the Eggs and Spawning-season of Petromyzon
fluvia and P. marinus. Proc. agp ay of Sct. and Cor-

a Zon Soc. London, Part I, pp.

Examined ee es eto aac from March to
middle a ay. U April 19, there were more females than
males; thereafter aa cnbatinionsa females, two to one.
Specimens taken April 26, appeared ready to spawn. By May
10, nearly all examined had spawned. Seven specimens of P.
marinus were taken in the Severn on May 3—‘‘ about Pah time
they ascended that river for the purpose of spawning.’

Young, Robert T., and Cole, Leon J.
1900. On the nesting habits of the eae lamprey (Lampetra wilderi).

Amer, Naturalist, XXXIV, pp. —620.

Notes on nesting observed in pa small tributaries of the
Huron River near Ann Arbor, Michigan. Males precede females
pa beginning the nest. Nests are 74 inches in diameter and

uated anywhere in river.
sae nea Ernst
1902. Lehrbuch dee on Entwickelungsgeschichte der
niederen Wirbelti
Résumé (pp. 14-78) , and bibliography (pp. 74, 89-91).

SHORTER ARTICLES AND DISCUSSION

A SIMPLE TEST OF THE GOODNESS OF FIT OF
MENDELIAN RATIOS

In actual experimentation the so-called Mendelian ratios,
2k, 9:3:3, 9:3:4, 9:7, 15:1, 37:9: 9:9:3:3:3:1, otd, are
never exactly realized because of the errors of sampling TASER
in all statistical work. Notwithstanding this fact, the best the-
oretical formulæ must be selected on the basis of these mislead-
ing experimental results.

Now the test of the validity of any Mendelian formula is two-
fold: the number of individuals found should agree with the
number expected within the limits of experimental error,’ the
assumed germinal composition of the several groups of individ-
uals should be capable of substantiation from a study of the
soma of their offspring.

For the most part, Mondaini have been satisfied to judge
the goodness of fit of the theoretical frequency to the empirical
by inspection merely. More recently, however, attempts have
been made to apply scientific tests to this problem. The first
was that of Weldon,? but Professor Johannsen doubtless de-
serves the credit of having interested the few Mendelian workers
who have taken the pains to calculate probable errors in this
indispensable part of their work.

The test used by Professor Weldon and recommended in a
much extended form by Professor Johannsen? is essentially the
determination of the probable error of the number of individ-
uals in one of the subgroups by the formula

=Va XP X4,
where p is the chance of occurrence of an individual of any
class, g= 1 — p, and n is the number of individuals. Thus the

1 In some cases, valid reasons for discrepancy between calculated and
observed frequencies may be shown. These factors should then be taken
aF account in calculating the theoretical numbers.

n, W. F. R., ‘‘ Mendel’s Laws of Alternative Inheritance in Peas,’?
Sabet J: 228-254, 1902, especially pp. 233-234.

3 Johannsen, W., ‘‘Elemente der Exakten Erblichkeitslehre,’’ pp. 402-
410, 1909.

741

742 THE AMERICAN NATURALIST [Vou. XLVI

‘probable error’’ of the number of individuals of any class,
say p, is
Ep = .67449 \/npgq.

Now while Professor Weldon’s use of this formula for the
simple 3:1 ratios seems quite proper, the same can not be said
for Professor Johannsen’s generalization. This is true for three
reasons:

(a) The formula is valid only when neither n, p nor q is
small. In polyhybrid ratios p or q may be relatively small.*
It is then quite idle to use the probable error suggested, unless
n be large, which unfortunately is generally not the case.

(b) Even when p is not so low as to render the use of the
conventional formula for the probable error open to question,
it is very laborious to calculate the probable errors for the fre-
quency of each class.’

(c) It is not only cumbersome and laborious, but theoretically
unjustified to test the validity of a given ratio by the determina-
tion of the probable error of one or of all of its individual com-
ponent groups. The random deviations of the class frequencies
are not independent, but correlated. We must have a usable
criterion of the goodness of fit of the theory to the data as a
whole.

Such a criterion was furnished several years ago by Pearson.’
Its applicability to the problem of testing the goodness of fit of
Mendelian ratios seems obvious, but since, as far as I can ascer-
tain, it has nowhere been applied to this problem, it seems
worth while to call the attention of students of genetics to its
usefulness.

xX = S{(0—c)?/c},
where o is observed frequency of any class, c is calculated fre-
quency on the basis of Mendelian theory and S indicates a
_ Summation for the several classes distinguishable in the ratio
under consideration.

P, a measure on the scale of 0 to 1 of the probability that

‘For example, Pap (loc. cit., p. 405) tables values for p = 3/4,
q= 1/4 to p = 63/64, q= 1/6

5 See, for instance, Ka pió given by Johannsen, loc. cit., p. 396

“Pearson, K., ‘On the Criterion that a Given System of Deviations
from the Probable in the Case of a Correlated System of Variables is Such
that it Can be Reasonably Supposed to have Arisen from Random Sampling,’’
Phil. Mag., 50: 157-175, 1900.

No. 552] SHORTER ARTICLES AND DISCUSSION 743 °

the deviations from the theoretical frequencies may be reason-
ably supposed to be due to the errors of sampling, may be cal-
culated from x? by formule which need not concern us here,
since its values for systems of frequency of 3-30 classes have
been tabled.” Hence the Mendelian has only the simple task of
calculating x? and looking up the value of P in Elderton’s tables.

Illustrations will make method of computation and usefulness
most clear.

ILLUSTRATION I, DOUBLENESS AND PLASTID COLOR IN STOCKS
Saunders, Journ. Gen., 1: 349-350, 1911

Obs. Cale. | o-c | (o-c)? (0-c)2/¢
|
Singles; White .4c...'. 1,666 1,615 51 2,601 1.611
Doubles, White......... 773 807 —34 1,156 1.433
Singles, Cream......... 790 807 = SOE 289 .358
3,229 | 3,229 | rE =3.402

Whence, from the tables in Biometrika, and by interpolation,
m= A P= eee
W238, Cam Pee ig5eso
Diff. = .087795
P = .223130 — .087795 X .402 —.1878.

Thus only in about one case in five would the errors of sampling
lead to divergences from theory as bad as this. The theory is,
as far as this evidence goes, possible, but certainly not demon-
strated.

ILLUSTRATION IJ. SEED FORM AND COLOR IN Pisum
Bateson and Killby, Report Evol. Com., 2: 77, 1905

Obs. | Calc.8 o-¢ | (0-¢)2/e

Round: Yalow- ei eee 4,926 4,883 +43 3787
Tinkled, Yellow. -e 1,656 1,628 +28 4816
Round, Gron: Pare ee 1,621 1,628 —7 .0301
Wimkled, Green.. ... 3.3503. 478 542 —64 7.5572

x? = 8.4476, P= .0384. Thus taking the data as they stand,
it is impossible to regard the 9:3:3:1 ratio as satisfactorily

* Pearson, loc. cit., gives a small table. A much more comprehensive one
is given by W. Palin Elderton, Bogie for Testing the Goodness of Fit of
Theory to Observation,’’ Biometrika, 1: 155-163, 1901.

3 These are not the calculated as given by Bateson and Killby,
but have been recalculated as closely as possible on the 9:3:3:1 ratio.
Theirs are nine seeds short.

744 THE AMERICAN NATURALIST [Vou. XLVI

describing the facts. But the great factor in the magnitude
of x’ is the deficiency in the wrinkled green seeds, and the
authors have suggested a reasonable biological explanation for
this deficiency.

ILLUSTRATION III. COLOR IN OATS
Nillson-Ehle, fide Baur. Einf. Exp. Vererbungsl., pp. 66-67

| Obs. | Cale. | 0-c! | (o-c)?2/e
Schwarsspelsig.................. 418 wo | 0095
Deepa a a 106 ee | 0095
Vepa, o an La 36 Sob ek | 0286

Thus x?=.0476 only. P is not tabled for x? < 1, since the
probabilities of such deviations being due simply to errors of
sampling are so enormously high. Theory and observation
could hardly agree more perfectly.

ILLUSTRATION IV. Bopy COLOR IN Drosophila
Morgan, Journ. Exp. Zool., 13: 35, 1912

obs. Cale. o-c (o-c)2/e

Me E gia a 525 529 —4 .030
Oe oe re 340 265 +75 21.226 -

MN Be a 194 265 -7i 19.023

x? =40.279

Here x? is over 40, the odds against the deviations, being due
to errors of sampling, are so enormously great that it is idle to
express them in figures. In short, the facts do not substantiate
the hypothesis, and Professor Morgan has himself suggested
possible reasons for the disagreement.

ILLUSTRATION V. PARTIAL GAMETIC COUPLING IN SwEET PEAS
Bateson, Saunders and Punnett, Rep. Evol. Com., 4: 11

Observed Calculated on | Calculated on
Number of (OG Bd Ot g ye Be ie d
Cases Basis Bas

Fore boae. o a 493 471 490
Furla, romad a 25 40 20
Red ee 25 40

Red. romid.. a 138 130 151

For the 7:1:1:7 basis, x’ = 12.7699, P=.0053. For the
15:1:1:15 hypothesis, x = 3.6375, P=.3086. Thus the

-

No. 552] SHORTER ARTICLES AND DISCUSSION 745

chances are about 995:5 or 199:1 against the validity of the
first hypothesis and only 69:31, or about 2:1, against the
second.

ILLUSTRATION VI. COLOR INHERITANCE IN Antirrhinum
Wheldale, Marryat and Sollas, Rep. Evol. Com., 15: 15

| Obs. | Cale. | o-¢ (0-¢)2/e

a CPCS ren ae E E R | 399 361 +38 4.000
E Maisa Genk: oe 122 120 + 2 .033
L RN iG ee ree 131 120 +1 .008
i Cri cola: a a 38 — 2 .100
Toa a cee ee a 88 120 —32 9,075
T ivory DE aa 35 40 — 5 .625
ELS ARES AE a rag 33 40 — 7 1.255

T. aopla aa T rE A us O T 19 14 +5 1.786
| 855 855 x? =16.852

Hence, P==.0185 or the chances are about 980 :20 against such
discrepancies being chance deviations from the theory. Thus
either the theory must be discarded or reasons for the discrep-
ancies found.

A conspicuous advantage of this method of Pearson is that in
its application the deviation of observation from theory for
each class and the amount which this discrepancy contributes
to x’ are under the worker’s eye.

If used with the caution that should be exercised in the draw-
ing of any conclusion from probable errors,® I believe that this
well-known criterion of pen of fit will prove most useful
to Mendelians.

i J. ARTHUR HARRIS

Some biologists apparently seem to feel that the calculation of a statis-
tical ‘‘probable error’’ covers all the biological sins which may be com-
mitted in the collection or manipulation of their data.

NOTES AND LITERATURE
NOTES ON ICHTHYOLOGY

In the Annals of the Carnegie Museum, Vol. VII, 1911, Pro-
fessor Edwin C. Starks gives the result of the survey of San
Juan Island in Puget Sound. Professor Starks regards Hez-
anchus corinus from this region as identical with Hexanchus
griseus, a view already suggested by Mr. Regan. He regards
Raja stellulata as a valid species. Raja.kincaidi is identical
with F. stellulata. New species as follows are described and
figured: Sebastodes deani, S. clavilatus, X. empheus. Xystes
axinophrys is the young of Averruncus emmelane. Xiphistes
ulve is identical with X. chirus. One hundred and fifty-eight
species of fishes are enumerated as known to occur in Puget `

und.

In the Publications of the University of California, Vol. VIII,
1911, Edwin C. Starks and William M. Mann discuss a collec-
tion of fishes from San Diego. A new genus, Orthonopias, based
on O. triacis, a new species of sculpin with a scaly back allied
to Astrolytes, is described.. Another new genus is Rusulus,
related to Clinocottus and based on a new species, R. saburre.
Maynea californica Gilbert is a new species described from Gil-
bert’s manuscript. Valuable notes are given on other rare
species.

In Science, Vol. XXXI, p. 346, Mr. Henry W. Fowler notes
that Coccogenia Cockerell and Callaway (Proc. Biol. Soc. Wash.,
1902, p. 1, 90) is a synonym of Coccotis Jordan (Rept. Geol.
Surv. Ohio, IV, 1882, p. 852) both being based on Hypsilepis
coccogenis Cope.

In the Proc. Ac. Nat. Sci. Phila. for 1910, Mr. Henry W.
Fowler gives a list of little known fishes of New J ersey. He has
also notes on Chimæroid and Ganoid fishes. He recognizes a
number of Gar pikes, instead of the three usually recorded as
valid. The number is certainly greater than three, but such
studies as we have been able to make would not indicate that all
of those noted and figured by Mr. Fowler are really distinct
species. Mr. Fowler describes as new, Cylindrosteus scabriceps
from Leavenworth, Kansas, and C. megalops from Bay Port,
Florida.

In the same proceedings for 1910, Mr. Fowler describes

746

No. 552] NOTES AND LITERATURE 747

Dixonina nemoptera, a remarkable new fish of the family of
Albulide from Santo Domingo.

In the same proceedings Mr. Fowler gives notes on Salmonoid
fishes, describing as new, Stomias bonapartei from Bonaparte’s
collection from Sicily, Synodus dominicensis from Santo
Domingo, and Synodus dermatogenys from Hawaii.

In the same proceedings Mr. Fowler lists the fishes of
Delaware.

In the same proceedings Mr. Fowler describes a new flat
fish from New Jersey under the name Citharichthys micros.

In the same proceedings Mr. Fowler gives notes on the Clupe-
oid fishes. The new genus Heringia is established for Clupea
amazonica Steindachner. Ilisha narrangansette is described
from Narragansett Bay. The subgenus Anchoviella is proposed
for Engraulis perfasciatus. This group includes nearly all the
species of Anchovia, and it is perhaps of generic value as dis-
tinct from Anchovia and from Engraulis. Anchovia scitula is
described from San Diego, Anchovia lepidentostole from Suri-
nam, and Anchovia platyargyrea from St. Martins.

In the same proceedings for 1911, Mr. Fowler describes new
species from Venezuela and Ecuador.

In the Proc. U. S. Nat. Mus. Dr. Jordan and W. F. Thompson
discuss the gold-eye of the northwest, Amphiodon alosoides.

In the Proc. Biol. Soc. of Washington Barton A. Bean and
Alfred C. Weed discuss recent additions to the fish fauna of the
District of Columbia.

In the Bull. Wisconsin Nat. Hist. Xoc., Vol. IX, 1911, George
Wagner describes a new species of cisco from Green Lake, Wis-
consin, under the name of Leucichthys birgei.

In Sctence, Vol. XXXIV, No. 879, Mr. T. D. A. Cockerell
describes a new minnow from Julesburg, Colorado, under the
name of Notropis horatii.

In the Proc. Biol. Soc. Wash. Mr. T. D. A. Cockerell discusses
the scales of various fishes, concluding that the soles are not
degraded flounders but degenerate descendants from some flat
fish from which both have been derived. This conclusion has
also been reached by Professor G. H. Parker from a study of
the optic nerves of the two types. :

In the Bull. Amer. Mus. Nat. Hist., Vol. XXX, 1911, John
Treadwell Nichols has notes on Teleostean fishes. He describes
as new Moxostoma alleghaniense from Marshall, North Carolina.
Menidia audens Hay from Moon Lake, Mississippi, he thinks

748 THE AMERICAN NATURALIST [Von XLVI

identical with Menidia gracilis from Long Island. Blennius
fabbri Nichols, lately described as new from Florida, is the
young of Chasmodes bosquianus. In this paper the curious fish
called Stathmonotus teckla Nichols is figured.

In Science, December 3, 1905, p. 815, the Smooth Hound,
Mustelus mustelus, is recorded from New J ersey. This Euro-
pean species has not been previously known from our coast.

In the Ann. of Mag. Nat. Hist., Vol. VIII, 1911, Mr. C. Tate
Regan publishes a detailed classification of the Siluroidea or cat

hes.

Tn the same annals, Vol. IX, 1912, Mr. Regan gives a classifi-
cation of the Pediculate fishes.

In the same annals Mr. Regan describes the structure of the
Symbranchoid eels.

In the same annals, Vol. XI, 1912, Mr. Regan gives a study of
the Opisthomi.

In the same annals, Vol. VIII, 1911, Mr. Regan gives an
analysis and classification of the Gobioid fishes.

In the same annals, Vol. VIII, 1911, Mr. Regan gives a classi-
fication of the Cyprinoid fishes and their allies.

In the Sitz. Acad. Wiss. Wien, 1911, Dr. Franz Steindachner
describes a number of new fishes from South America.

In the Proc. Biol. Soc. Wash. Dr. R. W. Shufeldt gives a
valuable and interesting account of the rare pelagic fish Ptery-
combus brama. The singular Caristius lately described from
Japan by Dr. Smith, is an ally of Pterycombus, and belongs to
the same family.

n the Memoirs of the Museum of Comparative Zoology at
Harvard Samuel Garman gives a classification of the Chis-
mopnea or Chimeroid fishes. He describes Chimera gilberti
from Hawaii, with valuable notes on all the known species.

In the Mus. Nat. Hist. of Paris Dr. Pellegrin describes nu-
merous fishes from Ecuador, South America.

In the Ann. Carn. Mus., Vol. VII, 1911, John D. Haseman
gives an elaborate catalogue of the Cichlid fishes collected by
the Carnegie Expedition to South America.

In the same annals Mr. Haseman describes and figures numer-
ous new species from South America.

In the same annals; Vol. VII, 1911, Mr. Haseman describes
new species from the Rio Iguassu, an isolated tributary of the
Rio Trabernath, with its peculiar fauna.

In the Sitz. Acad. Wiss. Wien, 1911, Dr. Steindachner dis-

No. 552] NOTES AND LITERATURE 749

cusses the fish fauna of Lake Tanganyika with several new
species and excellent plates.

In the Bull. Soc. Zool. of France Dr. Pellégrin describes a
new Barbus from South Africa, and in the Bull. Soc. Philom.,
Paris, he describes a new Tilapia.

In Arch. Zool. Exper. of Paris Louis Fage discusses the small
codfish of the Mediterranean, showing that capelanus is distinct
from luscus and from minutus.

In the Publ. Dept. Agric. E. W. L. Holt and L. W. Byrne
describe the fishes of the genus Scopelus (earlier and therefore
preferably known as Myctophum).

In the Publ. Zool. Inst. of Lund University Nils Rosen gives
an account of the reptiles and fishes of the Bahamas, an excel-
lent piece of work. New species as follows are described:
Nannocampus nanus from Andros; Garmannia rubra from An-
dros; Gobiesox androsiensis from Andros; Anchenopterus gran-
dicomis from Andros. Mr. Rosen regards Holocentrus siccifer
Cope and Holocentrus puncticulatus Barbour as identical with
H. coruscus. He also suggests the possible identity of the genus
Gymneleotris and Pycnomma with Garmannia. The supposi-
tion is that in the first named genus the ventrals being described
as separated have been simply split apart, the membrane being
very thin.

In the Proc. Roy Soc. of Queensland J. Douglas Ogilby de-
scribes an interesting series of new species.

David G. Stead in the Publications of the Department of
Fisheries of New South Wales gives a valuable account of the
fisheries of that region.

In the Kongl. Sven. Vet. Handl., XLVII, 1911, Professor
Einar Lönnberg gives an account of the reptiles and fishes of
British East Africa.

In a considerable volume published by E. J. Brill, of Leyden,
1911, Dr. Max Weber, of the University of Amsterdam, and
Dr. L. F. de Beaufort give a complete index to the genera or
species described and mentioned by Dr. Pieter Bleeker, the
most voluminous of all writers of ichthyology. In view of the
exceedingly great difficulty in getting exact references to Dr.
Bleeker’s works, this volume of 410 pages of names and refer-
ences is exceedingly useful.

In the Proc. of the New Zealand Institute, 1910, Mr. Edgar
R. Waite gives a record of additions to the fish fauna of that
country with several new genera and species.

750 THE AMERICAN NATURALIST [Vou. XLVI

In the Report of the Scientific Investigations of Shackleton’s
British Antarctic Expedition, Edgar R. Waite describes the new
species obtained.

In the Records of the Canterbury Museum Edgar R. Waite
gives the scientific results of the trawling expedition of the New
Zealand government. Numerous interesting discoveries are
recorded.

In the Publications of the Department of Trade and Com-
merce of Australia, 1911, are given the results of the investiga-
tions of the steamer Endeavour by Mr. Allan R. McCulloch.
Many interesting discoveries are recorded.

In the Proc. U. 8S. Nat. Mus. for 1912 Mr. Radcliffe gives a
most interesting account of new Pediculate fishes taken by the
Albatross in the Philippines.

In the same proceedings Dr. Hugh M. Smith describes the
three Chimeroid fishes taken in the Philippines.

In the same proceedings Dr. Smith describes a new family of
Notidanoid sharks. The genus Pentanchus differs from the
others in having five branchial openings only, like the ordinary
shark. In a note in Science, July 19, 1912, p. 81, Mr. Regan
claims that this shark is merely a Scylliorhinus which has been
deprived of a dorsal fin. é

In the same proceedings Dr. Smith describes numerous Squal-
oid sharks from the Philippines. As to these, Mr. Regan claims
that Nasisqualus is identical with Acanthidium and with Deania.
Squaliolus is a valid genus.

In the same proceedings Mr. Radcliffe describes 15 new
species of Amia (Apogon) and related genera from the Philip-
pines.

In the Abhandl. Senckenberg. Naturf. Ges. Frankfurt, Vol.
XXXIV, 1911, Professor Max Weber gives an account of the
fishes taken in the Aru and Kei Islands with a series of excellent
figures.

In Science, May 12, 1911, Dr. Theodore Gill gives a valuable
review of Professor Thompson’s translation of Aristotle’s ‘‘ His-
tory of Animals.’’

In a volume entitled ‘‘The Freshwater Fishes of the British
Isles” Mr. C. Tate Regan gives a most valuable popular account
of the different river fishes of Great Britain and Ireland. This
is written in such a way that no person of intelligence need have
any difficulty in recognizing the different species found in the
British streams.

No. 552] NOTES AND LITERATURE 751

In the Abhandl. Bayer. Akad. Wiss., Munich, Dr. Victor Franz
publishes a most valuable paper on the bony fishes collected in
Japan by Haberer and Déflein. Many new species are described,
particularly from that richest of all collecting grounds, Sagami
Bay, and contains many notes of value in our study of the
Japanese fauna.

In the Publ. Imp. Univ. Tokyo Mr. Shigeho Tanaka, lecturer
in the Science College, has begun a series of figures and descrip-
tions of the fishes of Japan. This work is extremely well done
and each species is illustrated by excellent plates. There is no
attempt at classification, each of the five parts now issued from
April 15, 1911, to March 10, 1912, containing species valuable
for his purpose, without attempt at orderly arrangement.

In the Sitz. Acad. at Vienna, 1909, Dr. Victor Pietsechmann
describes a new species, Hemilepidotus megapterygius from
Japan.

In the Proc. U. S. Nat. Mus. Professor J. O. Sardor describes
many new species and genera from Japan and from the Riu Kiu
Islands.

In the Ann. Nat. Mus. Wien. Dr. Victor Pietschmann de-
scribes the variations of a frog fish in Japan, and also describes
two species of fishes from Formosa.

In the Journal of the College of Agriculture Dr. K. Kish-
inouye describes new herring from the Bonin Islands, and also
gives an account of prehistoric fishing in Japan.

In the Publ. Roy. Mus. Belgium at Brussels Professor Louis
Dollo has a very interesting discussion of what he calls Ethologie
Paleontology.

In the American Journal of Science Dr. Charles R. Eastman
describes several new sharks from Solenhofen, in the Carnegie
Museum.

In the Bull. Soc. Geol. France Dr. Maurice Leriche describes
eretaceous fishes from the basin of Paris.

In another publication at Lille M. Leriche describes Stampian
fishes of the basin of Paris.

In the Memoirs of the Carnegie Museum, 1911, Dr. Eastman
gives a catalogue of the fossil fishes contained in that museum,
with a description and figure of many species.

In the Mem. Mus. Roy. Belgique Professor Ramsey Traquair
gives an elaborate and valuable account of the fossil fishes of
the Weald from the Bernissart.

In the Proc. Acad. Sci. of Naples Francesco Bassani gives an

752 : THE AMERICAN NATURALIST [Vou. XLVI

elaborate account of the fossil Berycoid fishes (Myripristis
melitensis) from the Miocene at Malta.

In the Annals de Paleontologie M. F. Prien gives a valuable
study of the fossil fishes of the basin of Paris.

In the Conn. Geol. Surv., 1911, Dr. Eastman gives a descrip-
tion of the Triassic fishes known from Connecticut.

In the Geol. Mag. Dr. Louis Hussakof describes several
Arthrodira from Ohio.

In the Publ. Carn. Inst. Wash. Mr. Hussakof describes the
amphibian fishes known from Permian rocks from North
America.

In the Publications of the Fish Commission of Pennsylvania
Dr. David Marine and Dr. C. H. Lenhart describe their observa-
tions on the thyroid carcinoma or goitre of the brook trout, and
the possibility of the relation of this disease to cancer. These
studies are continued in the Journal of Experimental Medicine,
Vol. XIII, 1911. It is concluded that there is no evidence that
goitre is either infectious or contagious, its cause probably
depending on lack of or a disproportion of elements necessary
for proper nutrition. This is also discussed by the authors in
the Johns Hopkins Medical Bulletin, Vol. XXI, 1910.

In the Science Bulletin of the University of Kansas, Professor
Ida H. Hyde gives experiments on the effects of salt injections
on the blood pressure of the skate.

In the Amer. Jour. Anat. William F. Allen describes the
lymphaties in the tail of a large sculpin in California.

In the Trans. Canad. Inst. Professor J. P. MeMurrich gives
an interesting account of the life history of the Pacifice salmon.

In the Proc. Roy. Soc. Canada Professor MeMurrich has an
elaborate study of the marks on the scales of fishes by which the
age of salmon may be known.

In the Publications of Stanford University, 1911, Professor
E. C. Starks gives a detailed account of the osteology of the
mackerel-like fishes. He shows that Leiognathus is a true
Scombroid, and not in any way related to the Percoid family,
Gerride, with which Regan has placed it. In general the rela-
tions of the families on the Scombroid group are fairly deter-
minable by their external appearance.

In the Journal of Comparative N eurology Dr. R. E. Sheldon
discusses the relation of the dogfish to chemical stimuli, and also
the sense of smell among sharks. He shows that the dogfish

No. 552] NOTES AND LITERATURE 753

obtains its food chiefly through the sense of smell, which is com-
parable to that of the higher vertebrates.

In the Internat. Revue Hydrobiol. Leipzig, 1909, Dr. Victor
Franz discusses the effect of light on the movements of Indian
fishes.

In the Journal of Morphology, Dr. J. F. Gudernatsch de-
scribes the thyroid glands of fishes.

In the University of California publications Mr. Asa C.
Chandler describes the lymphoid structure on the brain of the
gar pike.

In the Bull. Bur. Fish. Profesor G. H. Parker describes the
influence of sense organs on the movements of the dogfish.

In the Zool. Jahrb. Wiss. Mr. J. C. Loman describes the nat-
ural history of the European lampreys.

In the Arkiv. fér Zool. Stockholm Nils Rosen describes the
blood-vascular system of the Plectognath fishes.

In the American Journal of Physiology Professor Parker
describes the integumentary nerves of fishes, their reception of
light and their significance in relation to the origin of eyes of
vertebrates.

In the Bull. Bur. Fish. Professor Parker describes the rela-
tion of fishes to sound.

In the Amer. NATUR., 1908, he discusses the origin of the lat-
eral eyes of vertebrates.

In the same bulletin, 1908, he discusses the structure and
function of the ear of the squeteague.

In the Century Magazine, 1910, Mr. Charles H. Townsend
discusses under the head of ‘‘Chameleons of the Sea,’’ the
changes of color among fishes.

In the Bur. of Fish. documents Professor Parker discusses ths
effect of explosive sounds on fishes. These noises are faint under
water and may startle Taia for the moment, but they have no
. permanent effect.

In the Journ. Exper. Zool. Dr. Francis B. Sumner discusses
the color changes of flat fishes with respect to their adaptation
to various backgrounds.

In the Journ. Coll. Sci. Imp. Univ. Tokyo Mr. H. Ohshima
gives an interesting and valuable study of the luminous organs
of different species of fishes.

In the Transactions of the American Fisheries Society Mr.
John P. Babcock describes his experiments in burying the eggs
of salmon and trout in gravel, the result of this being that a

754 THE AMERICAN NATURALIST [Vou. XLVI

much larger number of eggs hatch, and the young are more
vigorous than when hatched in the ordinary way.

If the eggs of salmon are buried beneath five or six inches of sand
and gravel, such eggs will hatch and the young will work their way up
through the sand and gravel to the surface, and by the time they
emerge they have absorbed their sacs and are then exempt from the
attacks of vegetable moulds.

Mr. Babcock thinks that to follow more closely the method of
nature will give more value to artificial fishery hatching.

In the Report of the Fishery Board of Scotland, 1910, Dr. H.
C. Williamson gives a valuable report of the reproductive organs
of different species of Scottish fishes.

In the Bull. Zool. Soc. New York Dr. F. B. Sumner continues
his study on the changes of color of fishes on different bottoms.
The purpose of these changes seems to be simply concealment
from their enemies as well as from those fishes on which they

prey.

In Knowledge, Vol. XXXIII, 1910, Rev. T. R. R. Stebbing
discusses genders in zoology, sharply criticizing the carelessness
with which scientific men have made what he calls ‘“‘ Homeric
blunders,’’ Homer being accustomed to nod when questions of
classical refinement were brought before him. Mr. Stebbing
proposes that in zoology every generic name, whatever its ter-
‘mination, should be recorded as masculine.

In the Philippine Journal of Science Alvin Seale gives a val-
uable account of the fishery resources of the Philippine Islands.

In the Biological Bulletin Victor E. Shelford gives an inter-
esting account of experiments on stream fishes and pond fishes.

In the Bureaus of Fisheries Document 733 William C. Ken-
dall discusses the American fishes, their habits and value.

In Science, October 11, 1911, Professor E. C. Starks discusses
the structure of the air bladder in Ophicephalus.

In Science, Vol. XXXIII, 1911, Mr. Starks discusses the
origin of the gobies. He regards them as somewhat allied to
the sculpins.

In the First Annual Report of the Laguna Marine Laboratory
of Pomona College, Claremont, California, are numerous excel-
lent papers on the local fauna of Laguna Beach in southern
California. Among these papers is an elaborate and well-
planned study of the fishes of the tide pools, by Charles W.
Metz. The report is accompanied by excellent plates.

INDEX

NAMES OF CONTRIBUTORS ARE PRINTED IN SMALL CAPITALS

ABBOTT, JAMES An Unusual

F.
iter tue Relation

Wat and T 5
peters a Seon, Factors in Mice,
A. STURTEVANT, 368; C. C:
LITTLE, 491
Alcohol, Effects not Inherited in
Hydatina senta, D. D. WHITNEY,

fas sit Colors in F, of a Cross
betw n Non-colored Varieties of
. EMERSON, 612; Cells
ize, E. AST, 363
sao in cage
rations, L. R. WALDRON, 463
Allelomorphist, Spurious, and ’Gen-
e Cor p E,

nae Distribution and Origin of
Le fe in, R. F. Scharff, T. Bar-
BOUR, ‘500
American Per vere
Samuel W. Williston, O. P Hay

Asplanchna, Case of ion E
simulating a Mutation, J. H.
Poses 441, 526

Asymmetric Color Resemblance in
the Guinea-pig, JOSEPH H. mash
and G. D. BUCKNER, 505
P eT = Tone Science,
OTTO GLASER, 712

Banta, A. M., The Influence i
Cave Conditions upon
Development in Larvæ of Tabiy
stoma tigrinum, 244

, T., Distribution and Or-
igin of Life in America, R. F.
Scharff, 500

ergs ‘and Punnett ; paise
ir eatin xt W. J. SPIL

Biological, Processes, Physical Anal-
F. CooK, 493; Sci-
ence, Reflections on pp es of,

TO GLASER,
— of the Crayfish, F. E. CHI-
ESTER, 279

sae La and Bie baad HUBERT
LyM. s 139
Bos indic Ev idence of Alterna-
tive Tihotitasen in the F, Genera-
tion cage Crosses = Bos taurus,
wE. r
ns of Paleobotany
oa °F. H. KNOWLTON,
207; Phylogeny and Taxonomy,
JOHN M. COULTER, 215; Morphol-
ogy, EDWAR :
Ecology, ARTHUR
Breeding of Mice, J. FRANK DANIEL,

Bryophyta,
— nk Cav
ON Taon, 684
oleg G. D., Asymmetrie Color
Resemblance in the Guinea- -pig,

eat -relationships of,
D s Hoveu-

5

Bursa dela Defective Inherit-
ance-Rati , G. H. Shull, wW. J.
EALES, "309

CAMPBELL, DOUGLAS HOUGHTON,
Distribution of Plants
; The Classifi-
84

ruct
tions in Xenojarasitim, 6
CasTLE, W. E., The Inconstancy of
Uni ohne. Ta On the In-
horitaneo of the Tricolor Coat in
ea-pigs, and its Relation to
Galton s Law of Ancestral Hered-
, 437

Cattle, a -horn, prenien of
LAUG

Col
Cave Conditions and Pigment Devel-
opment in Larve of Amblystoma

. S. JENNINGS, 366
STER, F. E., The Biology of
the Crayfish, 279
Chromosomes, and Particular Char-
acters in Hybrid Echinoid Larve,
NNENT, 68; Sex, E

avp H. E.
B. Wilson, W. J. SPILLMAN, 164

755

Do o

CLARK, HUBERT LYMAN, Biotypes

and Phylogeny, 139

perm of the Liverworts,
AS HOUGHTON ,CAMPBELL,

N., Gametie Coupling
a Cause of Correlations, 569
Color, Celettaate | in Short-horn Cat-

CoLLINS, G.

VELL,

83; Inheritance in the Meatere

Cells of Maize, E. M. East, 363;

mblance, Asymmetric, in the
Guinea-pig, JosepH H. Kast

UCKNER, 505

Colors, Aleurone, in F, of a Cross

between Non- colored Varieties of
Maize, R. A. EMERSON, 612

Characters, HENRY FAIRFIELD
vag 185, 249

C . F., Physical Analogies of
Bi olo ogi cal Processes, 493

ahr genes between piakup osomes
and Particular Speen in Hy-
brid Echinoid Larv Dives H.
TENN paie. 68; Tables, Condensed,

R HARRIS
Correlations, Gametic Coupling as a
of N. COLLINS, 569

DESTER, 279; and Waterbug, Da

usual Symbiotic epen between,

JAMES F. ABBOTT, 5

Cytology, Some kavera of, in rela-
tion to the Study of ‘Genetics,
EpMuND B. Witson, 57

DANIEL, J. FRANK, Mice: Breeding
for § Scientific p aispa , 591
Darwin’s Theory of uh by
the Bolari of Minor Saltations,
AIRFIELD OSBORN, 76
Davenport, C. B., Heredit ity and Fac
ha and Method s of Evolution, 129
a RE, Genetic.
ties o s E daat thera
efective Teheritance- Kation, 0. B
Shull, W. J. Spr 309

THE AMERICAN NATURALIST

G., Problems of :

[Vou. XLVI

Descendants, Starvation of the As-
cendants an Aiono a TIA of,
RTHUR HARRIS, 313, 6
Monee oo Canadian Oyster,

J. Sta x
Differential Mortality with Respect
eed Weight, J. ARTHUR HAR-
of Plants in North
LAS HOUGHTON
and Origin of
in America, R. F. Scharff,
T. BARBOUR, 500
TER r akie, CHARLES
OFOID, 308
ral aud Guinea- -pigs, ilk r
ey in, AREND L. HAGEDOORN,

512
Distr ihotion,

B
@
Ps
a 2
J

Dominance, Law of, W. W. STOCK-
ER,
ris ng Begs CHARLES W. HARGITT,

Daali PN F. E. Lutz,
W. J. SPILLMAN

East, E. M., Inheritance of Color
x the Aleurone Cells of Maize,
363; The Mendelian Notation and

Physiological ne 633

East, E. M., and H. K. Hayes, In-
heritaneo | in Maize, Wed, DP

y it

Echinold ot Hybrid, Correla-
tion bet Chromosomes and
Particular Characters in, Davip

TENNENT, 68

Eggs, Double, CHARLES W. HAR-
GITT, 556

Em MERSON, R. A., Aleurone Colors in

Maize , 61 2
Senses ; enetic al bana
and Spurious Allelomorphism
Maize, J. W. LMAN, 119
Evolution, yede Theory o
the Selection of Minor oM ipd
ENRY Pa OSBORN, 76;
Trion Epwin G. CONKLIN,
en Study of

erimental
Heredity, C. B. DAVENPORT, 129;
xperiments with Drosophila Am-
pelophila concerning, F. utz,
W. J. SPILLMAN, 163; of the Ver-
tebrates, William Patten, Wo. E.
RITTER, 623

Fairness in sesh Reviewing, J.
ART 49
Tean in the Domestic Fowl,
OND s 697
Fodevtay's Breeding Experiments

No. 552]

with stig a Pygerd, A. H.
STURTEV

Fertilization, Partial Is the ae
e Sex

HENRY W. 470
Pow, ——, Mendelian coxa
of Fecundity, Ray
Paa ge das
Hra Henry W., Ornamentation
in Fr esh- ici Fishes , 470
Frog, Sex-Ratio and Partial Fertil-
ization? T. H. Morean, 108

Galton’s Law of Ancestral Hered-
ity and Tricolor Coat in Guinea-
pigs, W. E.

Gametic Coupling as a Cause
Correlations, . CoLLINS, 569

Genetic Correlation “and Spurious
PIRE on WS in Maize, A.

J.

©
Fh

Em
daa ” Studies el rial
As er of us

Study of, Ep-

‘ ; oT
app, Relations of Paleobotany
F n papeki a 207
Paaka WALTER B.,
species of rii pi ays
LASER, OTTO, The Autonomy of
Biological Science, 712
Gortner, R. A., Studies in Melanin,
. J. SP PILLMAN, 1
Growth, Nuclear, during Persil De-
S. JEN

A New Sub-
L., 616

velop! oe Nes, 366
Guinea-pigs, Inheritance of the Tri-
1 e STLE, 437;
AREND RN, 682;
Pom a oe EEE p A in,
JOSEPH and G. D.
BUCKNER, oa
HAGEDOORN, AREND. On Tricolor
Coat in Dogs and Guinea-pigs,
Hardi Suce lfalfa
er ae L R N, 463

HARLES W., Double Eggs,

J. A , A First Study
Influence of Starvation of
) a

656; The

Correlation Tables when the Num-
ber of Combinations is large, 477;
On Fairness and Accuracy in Sei-

INDEX

757

entific Reviewing, 498; On Differ-

; Mendelian Ratios,
Harshberger, John Wo, a
graphic Survey of North America,
aE HOUGHTON pa oiea 166
O. merican Permian Ver-
tebaa, “Samuel W. Williston, 561
Hayes, H. K., and E. M. East, In-
heritance in Maize, W. J. SPILL-
MAN,
HENN, ARTHUR W., e Range of
Size in the aie 157

Heredity, W. J. SPILLMAN, 110, stay
309; and Evolution, ; _ DAVE
PORT 29; An cestral, dika’ s

Law, ' W. E. Cas , 437
Heterozygotes, PR of Pure

Homozygotie Organisms by Self-

Ee con, H. S. JENNINGS, 487
HOLLICK, ARTHUR, The Relations of
Paleobotany to ' Botany—Ecology,
23

Poea porie Organisms, Pure, Pro-
duction from Heterozygotes
Self- Ea aak, H. S. JENNINGS,

487
Honey-bee, Color Sense of, JoHN H.
LOVELL, 83
HUSSAKOF, L., e Spawning Hab-
its of the a Lamprey, 729
Hybrid Echinoid Larve, Chromo-
somes and Partic ular Characters
in, aren H. TENNENT, 68
Hybrids, rnp Defective Inherit-
, G. H. Shull, W. J.

ny 309
Hydra und die Hydroiden, Otto
ar CHARLES Atwoop KOFOID,

Tan pipet Cope, Distribution,
Habits and Variation, C. H. Ric
pe p 5

Ichthyology, Notes on, DAVID STARR
ORDAN,

Inconstancy of betas Characters, W.
E. Cas

gee at og Selection, RAY-

D PEARL,
nina — Studies in, H. M.
ke, W. J. SPILLMAN, 309
nia nla of Color in Shorthorn
AU

ttle, H GHLIN,
Maize, E. an

ayes, W. PILLMAN, 1113; of
Physiognomy, R. N. Sala Ww

SPILLMAN, 116; of Pigmenta-

758

tion, Bateson and Punnett, W. J.
SPILLMAN, 117; Ratios, Defective,
in Bursa Hybrids, G. H. Shull,
. J. SPILLMAN, 311; of ces in
the Aleurone Cells of Mai E.
M. East, 363; re arg om
R ’ NABOURS, grat bed the
Coat in a-pigs,
E. CASTLF, 437; Mansauan,
RAYMOND PEARL, 697
Invertebrates, CHARLES ATWooD
, Korom, 695
JEFFREY, EDWARD C., The Relations
of Paicobotany to Botany—Mor-
pholo 225

JENNINGS, H. S., Nuclear Growth
during "Barly Developmen nt, 366;
Production of Pure Homozygo otie

Organisms fro wi ao aas by

Self- fertilisation, 487

JOHNSON, RoswELL H., The Mal-
thusian Prineiple and Natural
bass 372

AVID STARR, Notes on
Tonor, 756
KASTLE, , Asymmetrie Color

J.
ii i ra "in the Guinea-pig,

Know ton, F. H., The Relations of
Paleobotany Ba Geolo ogy, 20
OFOID, CHAR ATW Proto-
zoa, 308; Tavertebeatie. 695

Lamprey, Sea, Petromyzon marinus
Spawning Habits, L. Husssaxor,
72

H., The Inheritance
of Color in Short-horn Cattle, 5
ke, 3 :

tu
Cotton, W. J. SPILLMAN, 309
Literary Note on Mendel’s ’ Law, W.
a TOCKBERGER, 15
G: G, Yellow and Agouti
Pate in Mice,
Liverworts Classification of, Doug-
N CAMPBELL, 684
LIVINGSTON Pook E., "Pr t
Problems in Soi Physics as re-
lated to Plant Activities, 294
j e Color Sense

of the Honey-bee, 83

Lutz, E., riments with Dro
sophila Po Sr concerning
Evolution, W. J. SPILLM AN, 163

Maize, Inheritance m, E. M. East
and H, hk. Hayes, W. J. Senz-

THE AMERICAN NATURALIST

[VoL. XLVI

MAN, 111; Genetic S

and ’ Spurious Allelomorphism in

R. A. Emerson, W ILLMAN,

119; Aleurone = of, E
East, 363; R. EMERSON, 612

Malthusian ee and Natural
Selection, ROSWELL H. JOHNSON,
372

Meisenheimer, J poset Die Wein-
bergse hnecke, CHARL ATWOOD
Koro, 695

Melanin, Studies in, R. A. Gortner,
W. J. MAN, 1

Mendel’s Taw, Literary Note on,

W. CKBERGER, 151

Mendelian, Proportions and the In-
siv FRANCIS

L,
697; Ratios, a ARTHUR Higa
741

Mice, the Yellow and Agouti Fac-
tors in, A. H. STURTEVANT, 36
C E. LITTLE, 491; their Poolt
and Rearing for Scientific Pur-
pos J. FRANK x Daxter,

Morean, T. H., Is Change
the Sox. Ratio of tn Fog that is is
affected by External Agents due

8

pary Relations of Paleobot-
o Botany, EDWARD C. JEF-

oi ly

Mortality, Differential, with Re-
svect to Seed Weig ht, J. ARTHUR
Har 12

IS

Mutation, Case of Polymorphism in

Asplanchna fe a B
Powers, 441, 526

Nasours, Ropert K., Evidence of
Alternative Inheritance in the F,
i from reg

sian seen

North America, Phytogeographie
Survey of, John Ha patina
— E HOUGHTON CAMPBEL

Nota and Literature, 110, 163, 308,
500, 561, 623, 684, 756

Nuclear Size and Cen Size, E. G.
Conklin, H. S, JENNINGS, 366

No. 552]
@nothera, Genetical Studies, BRAD-

LE ORE S, 377
Fresh-water
470

W. ihe

HENRY FAIRFIE "Tar

Baia of Minor Saltationsy 76;

5, 249
Supplementary

Oyster, Canadian,
Observations on Development, J.
STAFFORD, 29
Paleobotany, Modern Aspects t%
elations to G F.
KNOWLTON, 207; AN apie
logeny and Taxonomy, JOH
COULTER, 15; rphology,
WA

K,
Paleontologist, Continuous Orii of
Certain Uni aracters as ob-
serv re by x HENRY FAIRFIELD
OSBORN, 185
Patten, William, The Evolution of
she ‘Vertebrates and their Kin
. E. RITTE
Du. ba Selection Index
Numbers, 302; The endelian
itance of a ay in the
mestie Fowl, 697
Phenolic Substances, Inhibitory ae
tion upon Tyrosinase, . Gort
, W. J. SPILLMAN 117
Phylogeny, and Biotypes, HUBERT
LYMAN CLARK ; and Taxon-

3

ganie Beings,
mee paaie of Biological
es, O. F. CooK, 493
_Physiognomy, Inheritance = R. A.
Salaman, W. PILL 116
Physiological ce and " Mendelian
E

Pigment Development in Larvæ o
Amblystoma tigrinum, Influence
of Cave Conditions upon, A.
BANTA, 244

Pigmentation, Inheritance of, Bate-
son and Punnett, W. J. Seri ILL-

117
Plant Activities, Present Problems
hysies as gob to,
veal E. LIVINGSTO , 294

INDEX

759

Pollination of Green Flowers, JOHN
Powers, J. H., A Case of Poly-
F
ArwooD Ko-
308
Soca, eara Federley’ s sp saat
Experim with, A. H. STUR

VANT, ae

RAMALEY, FRANCIS, Mendelian Pro-

portions and the Increase of Re-
cessives, 34
Recombination, Law of, W. W.

RICHARDSON, H., J2, The Dis-
tribution of Hyla 'arenicolor Cope,
with Notes on its Habits and
Vinnkios, 605
RITTER, ; Patten on the
Origin of Vertebrates, 623

alaman, R. M., Inheritance of
Physiognomy, W. J. SPILLMAN,

Saltations, Minor, Darwin’s Theory
of Evolution by the Selection of,

Origin of Life in America, T.
BARBO 00

sore ine Reviewing, o_o in,

ARTHUR HARRI
Sed Weight, Differential Mortality
occurring in Field Cultures of
Phaseolus vulgaris, J.

HARRIS
Segregation, Law of, W. W. poo
BERGER, 152
— Index Numbers, RAYMOND
PEARL, 302

Self- fertilization,

Production of
ure Homozygotic —
from ee by, S.
JENN 487
= Ratio a the Frog, T. H. Mor-
08; Chromosomes, E. B
Wilo , W. J. SPILLMAN 4.
Shave hora Cattle, Sapien may of
Color, H. H. Lau , 9

Short rter Articles ana gg ee

108, 151, 244, 302, 363, 437, 487,
6 2, 741

. H., Defective Inheritance-

Ratios in Bursa Hybrids, W. J.

AN, 311

Range in De Vertebrates,
ARTHUR W. HENN, 157

760

Soil Physics, ea Problems, as
diman ted to Plant Activities, BUR-
N E. Liv VINGSTON, 294
A ao Habits of the Sea Lam-
etromyzon marinus, L.

om aeth J. ARTHUR HAR-

656
Shakes Otto, Hydra und die H

roiden, CHARLES Atwoop Ko- .

FOID,
STOCKBERGER, W., A Literary
ote on M endel ’s Law, 151
Structural Relations in Xenopara-

sitism, W. A. CANNON,
nA cHe The Yellow and
Factors in 368 ;
eeding Wivethnent,
oth Pygera, 565
i between
ater and a Crayfish,
JAMES F. pi a 553

Tables, Condensed Correlation, For-

mation when the number of Com-
es is large, J. ARTHUR
Harris, 47

Pieni and elas ae argv
of Paleobotany to, M.
COULTER, 215

TENNENT, Pary
tion betwee Chromosom
Particular Characters in Hybrid
Echinoid 68

Tricolor Coat in n Guinea: e pigs, W. E.

ASTLE, A L. HAGE-
DCORN, 682

H. The peirat
and

THE AMERICAN NATURALIST

[Vou. XLVI

Unit Characters, Continuous Origin
of Certain, as 'obse rved by a Pale-
adi 13 HENRY yp pineal Os-
9: Ponnan W.

E. as 352”

Variation, Habits and Distribution
Cope, U. B.

ON, JR., 6
Vertebrates, Range of Size, ARTHUR
NN, 157; Evolution, Will-
iam Patten, Wu. E. RITTER, 623
oN, L. R., Hardiness in Suc-
63
, Unusual
between,

Weinbergschnecke, y ies Meis-
enheimer, CHARLES ATWoop Ko-
por,
WHITNEY . D., The Effects of Al-
ste not pea in Hydatina

Witiston,
Perm

Samuel W., American
n Vertebrates, O. P. Hay

Witson, EDMUND B., Some Aspects
of Cytology in relation to the
Stu tedy of Genetics, 57

Wilson, E. B., Sex Chromo-
somes, W. J. ’ SPILLMAN, 164

Seer hep Structural Rela-

, W. A. Cannon, 675

Yellow and sin. ot Factors in Mice,
A R RTEVANT, 368; C. C.
LITTLE, 491

Zea mays L., New A of,
WALTER B. ’ GERNERT, 616

The American Journal of Science

Established by Benjamin Silliman in 1818

The Leading Scientific Journal in the United States
Devoted to the Physical and Natural Sciences, with special reference to
Physics, and Chemistry on the one hand, and to Geology and Mineralogy
on the other.

Editor: EDWARD 8. DANA.
Associate Editors: Professor GEORGE L. GOODALE, JOHN TROWBRIDGE, W. G. FARLOW and WM. M. DAVIS
ambridge; Professors A. E. VERRILL, HENRY S. WILLIAMS and L. V. PIRSSON, of New
Haven; Professor JOSEPH S. AIMES, of Baltimore ; MR. J. S. DILLER, of Washin,
Two volumes annually, in monthly numbers of about 80 pages each.

This Journal ended its first series of 50 volumes as a quarterly in 1845; its second
series of 50 volumes as a two-monthly in 1870; its third series as a monthly ended Dec-
ember, 1895. A Fourth Series commenced in 1896.

Subscription price, $6 per year or 50 cents a number, postage prepaid in the United States

$6.25 to Canada ; $6,40 to Countries in the Postal Union. Back numbers at reduced prices.

s&@ Ten-Volume Indexes, Vols. I-X XI-XX, fourth series, price one dollar. Address

The American Journal of Science
New Haven, Conn.

SECOND EDITION ee
AMERICAN MEN OF SCIENCE
A BIOGRAPHICAL DIRECTORY | :

EDI TED BY J. McKEEN CATTELL

ee nen he rei Soa pecan Serine thin of he names in ie

ee 5 eae ihe Dirsttory extends tor than 600 pages Moki
is we print onal oroa ne bound in buckram with eather label. raatna the work ba eee ;
increased in size size by more than 30 percent it ia sold atthe same as the first

The American Naturalist

A Monthly — established i in wae Devoted to the Advancement of the —— Sciences
the Factors

with Special Reference

of Organic Evolution and Her

CONTENTS not aug NUMBER
A First Study of the Ir of Starvation of the

ants. Dr. J. Arthur Harris,
and the Increase of Recess-

les and Dis : Inheritance of
Color in the Aleurone Cells of Maize: Professor
E. M. East, Nuclear Gro Early De-
nt: Professor H. 8. Jennings. Is

tors in Mice? A. H. Sturtevant, The Malthusian
Principle and Natural Selection. Dr. Roswell H.

SOENT OF JULY NUMBE
tuđi thers, HI. te Bradley

Bridence of nee R F Genera-
from Crosses of = hana ig on Tiao
eog Robert K. Nabo
Shorter Articles and SORE : On the Inheritance

Consens: Correlation Tabl
yten tho Sumber af Combinations pe Dr.
__J, Arthur Harris,

CONTENTS S SEPTEMBER NUMBER

Mortality th Bes t to Seed
woe r Dr.J. Art Field A Culfures of Phaseolus

between
Frofemor James E. Abbott. bonb e Eggs. Pro-
"Grate, De, Or E- Hayy sa Federley s Ere D ae
fe a a
o miaa a Se e ee

CONTENTS OF NOVEMBER NUMBER

E “The Mendelian Notation Description
“"Physiolo naan
ark ti