Caps > A

THE .

AMERICAN NATURALIST

A MONTHLY JOURNAL
DEVOTED TO THE ADVANCEMENT OF THE BIOLOGICAL SCIENCES

WITH SPECIAL REFERENCE TO THE Factors OF EVOLUTION

VOLUME LII

NEW YORK
THE SCIENCE PRESS
1918

THE |
AMERICAN NATURALIST

VOL: LIT. January, 1918 No. 613

INHERITANCE OF NUMBER OF FEATHERS OF
THE FANTAIL PIGEON

PROFESSOR T. H. MORGAN

CoLUMBIA UNIVERSITY

SEVERAL years ago I began to study the inheritance of
the number of the tail feathers in fantail pigeons, partly
because of a challenge that I would not recover the fantail
in the F, generation, the implication being that the in-
heritance was not Mendelian. The race of fantails is a
very old one and the pigeons have been very intensively
selected by fanciers for many years. It was therefore to
be expected that several modifications had in time been
accumulated in the direction of selection. Nevertheless,
it was to be expected that if a sufficient number of indi-
viduals were bred, the original type would reappear. If
two factors in homozygous condition are essential for the
reappearance in F, of the original fantail, then such an
individual is expected once in sixteen cases; if three fac-
tors, once in sixty-four cases; if four, once in two hundred
and fifty-six; if five, only once in 1,048 cases, etc. This
relation holds if the fantail factors are all recessive, but
fewer factors are called for if one or more of them is
dominant, and the question will be still more complicated
if the highest reaches of the variation are due to modify-
ing factors acting only in the presence of other factors.

It seemed unlikely, however, that the situation would
be found to be as simple as this; for, in the first place,
there is no fixed number of tail feathers characteristic of

5 ee

6 THE AMERICAN NATURALIST [ Vou. LIT

the fantail; selection of these birds has not been made ex-
clusively in regard to number of feathers, but in regard
also to their size and shape, their regularity of distribu-
tion, their method of spreading, ete. It was a priori un-
likely that the race itself is homozygous for all of the fac-
tors that influence the number of feathers. How far the
results would depend on whether the maximum effects
are produced by a homozygous condition in several of
the factors, with heterozygous condition in others, would
be a point not easy to ascertain in a race that produces as
few offspring as does the pigeon. Nevertheless, the re-
sults give, I believe, pretty clear indications that the effects
are due to several factors, and they indicate, moreover,
that the failure to recover the extreme type of the fantail
in F, is probably only a question of insufficient numbers—
in fact, the fantail type has probably reappeared in F,
though not in its most extreme form, even with the rela-
tively few F, pigeons that I have been able to get.

The work has extended over several years, owing to
lack of suitable quarters in which to keep the birds and of
assistance to take care of them. They had to be removed
to and from Woods Hole each year, with the consequent
loss of young and disturbance of the regularity of habits
essential to a bird as conventional as the pigeon.

The original stock was obtained from Dr. F. D. Solley,
of New York City, a well-known breeder of high-grade
fantails. Dr. Solley has also supplied me with informa-
tion as to the number of tail feathers in birds of his strain.
Unfortunately these numbers were not obtained until a
year after these particular birds had passed out of his
hands. He assures me they are typical, and the birds of
his stock that I saw when my parent birds were ob-
tained were closely similar in tail number, ete., to those
here recorded.

The birds with which the original fantails were bred
to get F, stock were ordinary birds. As they were not
pedigreed stock there is a small chance that they might
have contained factors of the fantail type, but this is

No. 613] INHERITANCE IN FANTAIL PIGEON 7

highly improbable, since they had the number of feathers
characteristic of nearly all other strains of pigeons, and
especially of the more common ones.! Three P, pairs
were used (two male fantail and one female) but the F, in-
dividuals were not kept apart (for want of space) and, as
no marked difference appeared amongst the F, progeny
when the fantail parent was female or male,.the F,’s from
the reciprocal crosses were mixed together. This is un-
fortunate, for fuller and more accurate observations
might have revealed significant differences indicative of
sex-linked factors. I can only state that if such are here
involved their effect is slight, and was not observed at
the time when the two kinds of F, offspring were reared.

History oF THE FANTAIL Race
In his book on ‘‘ Animals and Plants under Domestica-
tion’’ Darwin has given a great deal of important infor-
mation about the origin and characteristics of the fantail.

“The normal number of tail feathers in the genus Columba is 12; but
fantails have from only 12 (as has been asserted) up to, according to
MM. Boitard and Corbie, 42. I have counted in one of my own birds
33, and at Caleutta Mr. Blyth has counted in an imperfect tail 34
Pee tliat: In Madras, as I am informed by Sir W. Elliot, 32 is the
standard number; but in England number is much less valued than the
position and expansion of the tail. The feathers are arran m an
irregular double row; their permanent fan-like rings and their
upward direction are more remarkable characters than their increased
number. The tail is capable of the same movements as in fae pigeons
and can be depressed so as to sweep the ground. It arises from a more
expanded basis than in other pigeons; and in three skeletons there were
one or two extra coccygeal vertebrae. I have examined many specimens
of various colors from different countries, and there was no trace of
the oil gland; this is a curious case of abortion.2 The neck is thin and
bowed backwards. The breast is broad and protuberant. The feet are

1 At least one other of the Gomantientod, races may have more than twelve
feathers in the tail.

2 ** This gland occurs in most birds; but Nitzsch (in his” Prerylógraphio, *
1840, p. 55) states that it is absent in two species of Col: veral
species of Psittacus, in some species of Otis, and in most or all birds of the
Ostrich family. It can hardly be an accidental occurrence that the two
species of Columba which are destitute of an oil gland have an unusual
number of tail a namely 16, and in this respect resemble fantails.’’

8 THE AMERICAN NATURALIST [ Vor. LIT

small. The carriage of the bird is very different from that of other
pigeons; in good birds the head touches the tail feathers, which conse-
quently often become crumpled. They habitually tremble much; and
their necks have an extraordinarily, apparently convulsive, backward
and forward movement. Good birds walk in a singular manner, as if
page small feet were stiff. Owing to their large tails, they fly badly on

windy day. The dark-colored varieties are generally larger than
shies fantails.”

“Mr. Swinhoe sent me from Amoy, in China, the skin of a fantail
belonging to a breed known to have Tai imported from Java. It was
colored in a peculiar manner, unlike any European fantail; and, for
a fantail, had a remarkably short banca oie a good bird of the
kind, it had only 14 tail feathers; but Mr. Swinhoe has counted in
others of this breed from 18 to 24 tail feathers. From a rough sketch
sent to me, it is evident that the tail is not so much expanded or so
much upraised as in even second-rate European fantails. The bird
shakes its neck like our fantails. It had a well-developed oil gland.
Fantails were known in India, as we shall hereafter see, before the year
1600; and we may suspect that in the Java fantail we see the breed in
its earlier and less improved condition.” Vol. I, Chap. V, p. 153.

“ The first notice of the existence of this breed is in India, before the
year 1600, as given in the “ Ayeen Akabery”; at this date, judging
from Aldrovandi, the breed was unknown in Europe. In 1677, Wil-
lighby speaks of a fantail with 26 tail feathers; in 1735, Moore saw one
with 36 tail feathers; and in 1824, MM. Boitard and Corbie assert that
in France birds can easily be found with 42 tail feathers. In England,
the number of the tail feathers is not at present so much regarded as
their upward direction and expansion. The general carriage of the
bird is likewise now much valued. The old descriptions do not suffice
to show whether in these latter respects there has been much improve-
ment; but if fantails with their heads and tails touching had formerly
existed, as at the present time, the fact would almost certainly have
been noticed. The fantails which are now found in India probably
show the state of the race, as far as carriage is concerned, at the date of
their introduetion into Europe; and some, said to have been brought
from Caleutta, which I kept alive, were in a marked manner inferior to
our exhibition birds. The Java fantail shows the same difference in
carriage; and although Mr. Swinhoe has counted 18 and 24 tail feathers
in his birds, a first-rate specimen sent to me had only 14 tail feathers.2

A later statement in regard to fantails from Fulton’s
Book of Pigeons gives some additional bras
2 Darwin, ‘‘ Animals and Plants, ”” Chapter VII, p.

4‘*The Illustrated Book of Pigeons with cad for Judging,’’ by
Robert Fulton, edited by L. Wright. Cassell & Co., Ltd., New York.

No. 613] INHERITANCE IN FANTAIL PIGEON 9

The tail also is peculiar, and quite uncommon. It is long and com-
posed of 14 to 22 feathers, 16 being about the average number in these
birds; these are arranged equally on either side, one above another, and
the two top ones, diverging a little outwards, show a slight division in
the tail, but there is not the slightest affinity or resemblance to a “ fan’
tail, as some might suppose by the excessive number of feathers, but
it is a distinet peculiarity of this breed (12 being the normal number
of tail-quills in most pigeons). The greater the number of quills in
“Oriental Rollers” the more the specimens are valued. A further
singular feature noticeable in the tails of these birds is that occasionally
two feathers may be found growing from one quill, separating at its
pithy junction as a twin feather, each rather narrower than ordinarily,

ut of the usual length, and not outgrown, or causing a disordered
formation of the tail (p. 195).

- The tail is the other chief point in the English breed. The
fest should lie flat and evenly over one another (none of them
being set edgeways), so as to form a neat double row. In number they
should not be less than 28, but as many more as the bird can carry
nicely. The Birmingham Columbarian Society, in an article published
by them some years ago, laid down 40, arran in 3 rows, as the
proper number; but though I have heard of such birds I have never
seen one. I once had a hen with 38 tail-feathers. I purchased her -
from Mr. Fulton, and I believe she had been imported from India;
and I have often bred birds with tails of 36 or 37 feathers carried in
most orthodox fashion. In an exhibition pen the number is of no conse-
quence, provided that the tail is well spread and cireular, and well
filled up all around; but in the breeding pen a thickly-feathered tail is
of great value. In dee breeding of any animal for any faney point, if
you can get that point in excess in either of the parents so much the
easier is your task. You have then something to spare, instead of
something to breed up to, whieh is a very different matter (p. 329).

THe P, GENERATION
The three original fantails had 29, 30 and 32 tail
feathers, respectively (Fig. 1). From Dr. F. D. Solley I
got the records of other fantails of the same stock given
in Fig. 2. The other parents were ordinary homers pur-
chased from a breeder of these birds.
Tue F, OFFSPRING
The numbers of tail feathers shown by the 41 pe ri

uals of the F, generation are recorded in Fig. 3. The
range of variation is from 12 to 20, with the highest fre-

10 THE AMERICAN NATURALIST [ Vou. LII

Fic, 1. One of fantail pigeons used in the experiments.

quency in the 14-tail-feather class. Evidently one or
more of the factors of the fantail act as partial domi-
nants, producing tails that have for the most part more
tail feathers than has the common pigeon but less than
the fantail. In appearance these F, birds are more like
the common pigeon, having lost the peculiar carriage

4
2
| a Me sea T Je
12 28 29 30 31 32 33 34 35 36 37 38
Fig. 2. Frequency distribution of tail feathers in parent “ homers ” (left),
and fantails (right).

12

8

8

4

2

12 15 14 15 16 17 18 1
Fic. 3. Freqtency distribution of tail feathers in F;.

No. 613] INHERITANCE IN FANTAIL PIGEON 11

of the fantail and its peculiar shape. The tail is, how-
ever, often wedge-shaped instead of flat as in ordinary
birds. There were 28 birds with an even number of

de

12 15 14 15 16 17 19 19 2 21 22 23 2 25 26

Fic. 4. Frequency distribution of tail feathers in Fo.

Aiet and 13 with an odd TETEE considerable pre-
ponderance of even number of feathers. Of the 41 in-
dividuals, 30 are inodd i in the classes with 14, 15, 16 tail
feathers.

12 THE AMERICAN NATURALIST [ Von. LII

Tue F, GENERATION?

A glance at Fig. 4 shows that the range of variation of
the F, group is greater than that of the F,; that the 12-
feathered tail has reappeared in considerable numbers;
that the ‘‘curve’’ is at least bimodal with one apex in the
14, 15, 16 rows, and the other in the 12 row; that there
are a few individuals that approach the lower range of
variation of the fantail, viz., those with 24, 25 and 26
tail feathers.

There is a distinct return of one of the grandparental
types, viz., the 12 class. The 13-16 groups clearly corre-
spond to á large part of the heterozygous group seen in
F,. Whether the range to the right of this middle group
in the F,'s is significantly different from that in the F,
can not be determined by inspection, as the number of in-
dividuals is too small. If the F, and the F, groups are
made into curves the results show that it is doubtful if
the wider range in F, is significant, although the large
12-feathered class in F, makes the F, variability much
more marked than the variability in

Back Cross
Some of the F, birds, both males and females, were
back-crossed to fantails. Twenty-three offspring were
obtained which differed strikingly as a group from the
F, and F, lots. The number of tail feathers (Fig. 5) was
greater; no 12-feathered birds appeared (the lowest num-

ey | +
12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Fie. 5. Frequency distribution of tail feathers in back-cross.

ber was 14) ; while the highest number included birds with
30 and 31 tail feathers. The latter would undoubtedly
pass-for-fantail, so far as the number of tail feathers was

5 There are some diserepancies between the F, and back-cross tables
given here and the records of the groups given in ‘‘The Mechanism of
M an Heredity.*? The present account is more accurate, as some of
- the former data was obtained from the birds while still aliv

No. 613] INHERITANCE IN FANTAIL PIGEON 13

concerned. . The carriage of most of the birds was notice-
ably much more like the fantail than that of the F, and F,
birds.

NUMBER or Factors [NVOLVED

The recovery of a certain number of the normal 12-
feathered tail in the F, might seem to furnish a basis on
which to calculate the number of factors involved; but
the fact that a few 12-feathered birds appear in F, shows
that some, at least, of the heterozygous combination are
included in the F, troie feather group. ‘It is also pos-
sible, even probable, that other F, combinations may also
fall within this group. It is impossible, therefore, to ar-
rive at anything more than a possible conclusion from
the F, data because the relative value of the DAÑE
classes can only be guessed at.

Two factors will obviously not fit the results, pa
there would be expected more of the higher numbers of
tail feathers both in the back eross and in the F, count.
Three factors fit fairly well. Let.A, B, C represent par-
tially dominant factors for fantails, and a, b, e their nor-
mal allelomorphs (aabbee being the normal 12-feathered
tail). In the F, there will be expected only one pure
fantail out of 64 (viz., AABBCC) and one pure 12-feath-
ered type (viz., aabbec). There will be six F, classes with
only one dominant factor heterozygous for A or B or C.
These, theoretically at least, if all the factors have equal
efficiency, would be the most likely ones to fall within the
12-feathered group. If these include all of the expected
12-feathered tails in F, there should be seven 12-feath-
ered in 64. There were 278 F, individuals. On the same
calculation this would give. an expectation of only 10.5
twelve-feathered tails. But the F, records actually gave
46 normal tails. Obviously still other combinations
realized in F, must come under this class. It would be
mere guesswork to try to state wiih are the more prob-
able combinations.

: The back cross furnishes dais that permit a better means |
of calculation. Here eight kinds of germ cells and eight

14 THE AMERICAN NATURALIST [Von. LIT

zygotes are expected on the assumption of a three-factor
cross, Viz.,

ABC ABc Abe aBC abC aBe abe

abe abe abe abe abe abe abe

Of these eight kinds of individuals, some of only one class,
might be expected to be wild type (viz., of class abe ABC)
in the sense that individuals of this class correspond in
formula to the F, offspring, and, of these F, offspring, 2
out of 41 have tails with 12 feathers, or 1 in 20. Amongst
the 24 back crossed individuals, there were none with 12
feathers only and at most one is expected. If we assign
to the group of abeABC also the four individuals of the
back cross in the 14 and 16 groups, and assign the 3 indi-
viduals of the 30 and 31 groups to the pure fantails, there
remain 17 individuals in the middle range that belong to
the six intermediate groups that are homozygous in one
or in two fantail modifiers. There are six intermediate
classes between the end classes just spoken of. If we
are right in the limits assigned to the end classes, the ex-
pectation would be 18 individuals for the intermediate
classes, where 17 are so classified, which is also not a bad
fit.

Four factors fit the data about as well as three,* but if
three will suffice the smaller number is perhaps preferable.
It is evident that the data do not allow close analysis, but
only because they are not sufficiently large, especially in
the back cross. Nevertheless, it is important to find out
that, so far as the results go, they are not unconformable
with the Mendelian es of segregation of a few
pairs of factors.

LINKAGE

When all F, tails that are blue are classified they fall
into the groups shown in Fig. 6; similarly, the white tails

6 On this assumption TE S fewer fantails are expected in F., which
is a better fit, but fewer also in the back cross, which apparently is not so
good a fit. The proportion would also depend, however, on the relative _
efficiency and the completeness of the dominance of each factor. The
above evidence proves that there must be at least three factors. o

No. 613] INHERITANCE IN FANTAIL PIGEON 15

give the groups shown in Fig. 7. A comparison of these
groups shows that there is a relatively large number of
high-feathered tails amongst the whites, while among the

18,195 T4- 15 16" 17. 10 19 20 21. 28 25- 94 85 y

Fig. 6. Frequency distribution of blue tail feathers in Fə.

12 153 M4 15 16-17: 18 29 2 D21 98 95 2 25
Fic. 7. Frequency distribution of white tail feathers in Fo.

blues, the 12-feathered tails are relatively more frequent.
A not improbable interpretation of this relation is that
the principal factor for white is linked to one or more of
the factors for increased number of feathers,

16 THE AMERICAN NATURALIST { Von. LII

Since these results occur in the F, count, it is unfortu-
nately not possible to deduce from them whether crossing
over takes place in one or both or neither sex.

Amongst the tails were some that had both blue feathers
and white feathers. These give the group shown in Fig.
8, which closely corresponds to the blue-tail group (Fig.
6). There were other tails with white feathers having

Í
dd Bee 7. 18 19020 21 22 (Os... 24. 96
Fig. 8. Frequency distribution of blue-and-white tail feathers in Fe.

pigment along the margins as in Fig. 15. These, when
classified, gave the group shown in Fig. 9, which ap-
parently is the same as the group of white tails (Fig. 7). .

The number of birds in the F, and in the back cross
are too few to give significant results when broken up
into the two groups of white or blue. `

12 "13.14 15 EA ee ee.
Fro. 9. Frequency distribution of “edged white” tail feathers in Fs.
The tails are not a complete index of the bird from
which they came, for a bird with a pure white tail might
have color patches elsewhere on its body; but as no rec-

No. 613] INHERITANCE IN FANTAIL PIGEON 17

ords were kept of the entire color of each bird, it is not
now possible to find out how closely the complete pattern
would correspond with the tail color. In general, how-
ever, in these birds the tail is a partial index, at least—a
fair sample, perhaps—of the entire color.

‘í CORRELATION ?? BETWEEN THE OIL GLAND AND THE NUMBER
or Tar. FEATHERS

Darwin suggests a ‘‘correlation’’ between the absence
of the oil gland and the increased number of tail feathers.
Such a relation might be a direct correlation in the sense
that the overdevelopment of the tail feathers suppresses
or tends to suppress the development of the oil gland that
is situated on the uropygium just above the base of the
tail feathers. If this were the true interpretation of the
condition in the fantail, one would expect to find in F,
and F, all degrees of development of the oil gland. If,
on the other hand, the absence of the oil gland is an in-
herited peculiarity having nothing directly to do with the
number of feathers, then in the F, series we might expect
to find a numerical relation indicating its mode of inherit-
ance. Unfortunately the pedigrees of the normal tailed
pigeons that had been mated to the fantails were un-

own. While the oil glands may be occasionally absent
in domesticated pigeons, it is highly improbable that any
of the homers used in the experiment carried such a fac-
tor. In classifying the F, and F, birds according to the
condition of the oil gland three dassos were recognized.
First, ‘‘double’’ glands, those with the right and left sides
almost separate, each with a separate opening; second,
““single”” glands, those with the halves united more closely
and with but one external outlet;” third, those with no oil
glands. The results are given in graphs of Figs. 10-11.

_ The few F, birds available when the oil gland was
studied show a wide range of variability; all but one were
double, Fig. 10 (above). This doubling might be due to

7 An intermediate stage was also tad viz., one with two closely fused

we. fo mae lost a single. sland,

[VoL. LIT

THE AMERICAN NATURALIST

18

“gq pue "W ul spurj jjo Jo aoualimodo Jo seAano Aouenboay

*(M0[94) TT “DIM *(9A0qU) OT “DIM

SUON etus etanog
S3 8 SS 383 12 OF Gl ƏT ¿il ƏT Gi HL ST Zl 32 13 03 61 81 4t Ot 91 FL ET Bt sz ve čs 828 183 OS 6l Bt ¿1 91 SI pl SI St
Y
suoN e12u1S etqnog
et úl Of 91 Sf ‘St. et
a
b
Y

No. 613] INHERITANCE IN FANTAIL PIGEON 19

partial dominance of a gene for doubling, or to a ‘‘corre-
lation”” such as Darwin spoke of, for the number of tail

b w bw ew Pe ew
38 Double Single None

1

Fic. 12 (above). o 13 (below). Frequency distribution of double, single
o oil glands in differently colored tails.

feathers in this particular lot was high. That the latter
is probably not the explanation is shown in the F, birds.

20 THE AMERICAN NATURALIST [Vou. LII

The three groups of F, tails (Fig. 11) show 126 doubles,
36 singles, 46 none. The expectation for two factors
(9:3:4), on the assumption that the doubles differ from
the singles by one factor, and both from none by an-
other factor, is 117 doubles, 39 singles, 52 none. This is
not a very bad fit.

There is one striking result brought out by these curves,
There are no 12- or 13-feathered birds without an oil gland.
This is an expression of the relation that Darwin sug-
gested as due to ““correlation”” in the sense that more tail
feathers suppressed the development of the oil gland.
But there is no such obvious solution as is shown by the
F, group, for there may be a large number of tail feathers
present and the oil glands be well developed, single or
double. The curves suggest, rather, linkage between
a gene for extra feathers, and a gene for absence of oil
glands.

The tails with double, single and no oil glands were also
classified according to the four color groups already re-
ferred to, viz., blue, white, blue and white, edged white
(Figs. 12 and 13). The double and single curves appear
to be the same, the no oil gland curve seems significantly
different. If so, it means that there is some linkage be-
tween white color and absence of oil glands.

The foregoing evidence makes probable the view that a
gene for more than 12 feathers, and the gene for no oil
gland, and a gene for white color are linked, 1. e., are
carried by the same chromosome. The genes for the oil
gland and for the number of tail feathers are closer to
each other than either is to the gene for white. More
data, especially from back-crosses, will be necessary to
establish this conclusion.

Spurr FEATHERS |
Dr. Solley tells me that the split and double feather
that occurs at times in the fantails is selected against.
It is of not infrequent occurrence in the F, and back-
crossed birds that I have obtained. In the records these

No. 613] INHERITANCE IN FANTAIL PIGEON 21

split feathers have been counted as one feather, without,
however, intending to prejudice the question of the single
or double nature of these feathers.

The most striking cases are like those represented in
Fig. 14 (top row) and Fig. 15, where what appears to be

Fic. 14. Some types of split feathers.

a single feather is split in two throughout its length.
While there may be a complete shaft in each half, yet the
two vanes that lie on the ““inner”” side are not so broad as
the outer half vanes, and their edges are generally frayed
out and imperfectly formed. Often the vanes run across
and unite the two halves.

22 THE AMERICAN NATURALIST [Von LIL

In some of the split feathers, the division is obviously
into right and left halves (Figs, 14, 15, 16) ; in other cases
the halves make an angle with each other (Fig. 14, Nos. 37
and 80), while in still others one larger part may lie above
a smaller part (Fig. 14, lowest row). Whether in the
last cases the division has been in a horizontal plane, and
in the first cases in a vertical one, is not certain, although
the shape of the feathers even in the last case, with the
imperfect edge and a narrower margin, would seem to
make most probable the view that in all cases the division

Fie. 15. An F, tail (“edged white”) with one split feather.

has been into morphological right and left halves. The
final position of the feather halves may be due to a later
twisting in the sheath, or to crowding of the feathers at
the base. This interpretation is further substantiated by
cases in which the center of all the feathers has a white
area (Fig. 14, No. 55, and Fig. 15); this is found on the
imperfect side of the split feathers even when they lie one
above the other. In all there were 24 F, tails with split
feathers. Five of these had each two split feathers.
These cases grade into those in which only the distal
end of the feather is split, as shown in Fig. 14, middle
figure, and Fig. 16. The impression produced by feathers

No. 613] INHERITANCE IN FANTAIL PIGEON 23

of this kind is strikingly in favor of a single feather split
into right and left parts at the distal end. In all, three
eases of this sort were found.

To the same group are to be referred two cases, one of
which is shown in Fig. 14, No. 5, b. Here there is a single
feather, but the midrib is split near the end. The vane

wan

Fic. 16, Split feathers with normal feathers that lie next to them.

lying between the two midribs is continuous, yet the bend-
ing inwards of this part is indicative of its dual nature.

More extreme are the eight cases of which three are
shown in Fig. 16, lowest row. In all such cases there is
a large, almost fully formed feather with a smaller, less
perfect piece underneath the larger part. The first im-
pression is that a piece has been split off the ventral side
of the feather by a division in the horizontal plane. A

24 THE AMERICAN NATURALIST [ Vou. LII

closer scrutiny shows, however, that the large feather is
_ ragged along one edge only (or on a part of one edge),
while the smaller piece has also on the same side (as can
be seen in some cases at least) a ragged edge with the
other vane more nearly complete and with a not-rough
edge. It seems, therefore, more reasonable to interpret
even these cases as extremes of the split-feather type in
which one piece has fared worse than the other (or in
which the original division was into unequal pieces).

THE Size or THE DOUBLE FEATHERS

There is a graded difference between the outer and
inner vanes of the feather from the edge to the middle
of the tail Fig. 15. The outer half of the vane is rela-
tively smaller in the outermost feather, right or left, and
equality of the two sides is more and more reached, the
two middle feathers of the 12-feathered tails are about
symmetrical. In the multiple feather tail these relations
still hold, but are more difficult to trace than when the
tail is simpler. It would not be profitable to attempt to
analyze in detail these relations as applied to the double
feathers further than to compare their surface relations
with that of the feathers nearest to them or with their
symmetrical mates, In all cases of split feathers the
outer halves of the vanes are not so wide as is expected
from the nearest feathers (or their symmetrical mates as
seen in Figs. 14-16). The middle part is, as a rule, very
much less than a right or left vane. The total width of
the split feather is, as nearly as I can judge, about the
same as the expected feathers for that position. The im-
pression indicates that the sum of the four vanes is a little
greater than the sum of the two normal vanes, but there
can not be much difference as measurements show. The
looseness of the frayed inner edge makes it difficult to get
a very close estimate of the actual relations.

The general conclusion is that we are dealing with a
single rudiment that has split at a very early stage into
two parts that have completed themselves as whole

No. 613] INHERITANCE IN FANTAIL PIGEON 25

feathers, so far as this intimate union in the middle line of
the bud permitted. There are no indications that the
split feather is due to the union of two separate rudiments
that have been pressed together so closely as to interfere
with the full development of each when they came in
contact.

THE Location or THE SPLIT FEATHERS

The location of the split feathers (and modified types) is
given in the next table.

Feather ‘‘split’’ ... 9 near middle, 1 one quarter from side
End only split ..... 2 near middle, 1 one third from side
Double vein at tip. . 2 near middle

1 two thirds from side,

Very unequal parts. 7 near middle, Piiterniost Heather

In the great majority of cases the doubling occurs near
the middle of the tail. The meaning of this is not at all
apparent. We know so little about the cause of duplica-
tion in general and about the embryological mechanism
that is involved in laying down the feathers in the tail,
that it is useless to speculate about the result. The evi-
dence from experimental embryology shows unmistakably
that doubling may result from a mechanical interference
with the relation of the blastomeres after they have as-
sumed a definite position in regard to each other, but there
are also many other cases known where, in normal devel-
opment, a part is repeated several or many times. In
these cases we can as yet only surmise that the rudiments
of the structure—simple cells or groups of cells—become
mechanically drawn apart by the more rapid growth of
surrounding parts and separated so that each gives rise
to a separate organ. Split feathers, from this point of
view, would be looked upon as an incomplete separation
of certain of the rudiments. However this may be, one
can imagine other ways by which a specialized group of
cells could become broken up into islands.

26 THE AMERICAN NATURALIST [ Von. LIT

OTHER CHARACTERS IN THE Cross

Three other characters are conspicuously present in
the fantails besides the tail, viz., the white plumage, the
carriage of the bird, and the shaking of the head and
neck. The dominance—incomplete—of the white of the
fantail was noted,’ but the mixtures that appeared both

in F, and F, make it probable that the results are not

due to a single factor. The extraordinary position of the
fantail pigeon with its head thrown back until it touches
the tail feathers appears also to be due to at least as
many factors as is the number of feathers in the tail;
for it was not recovered in any of the F, birds, although
in the back cross there were birds that showed some ap-
proach to the fantail posture. The shaking of the head
disappeared in F, and indications of it were seen occa-
sionally in F, and especially in back crosses. The char-
acter is of such a kind that its study is difficult, and it may
well be an expression of some structural modification of
the body rather than any direct psychological factor.

CASTRATION OF MALE

The absence of marked. secondary sexual characters in
the male, characters that are so conspicuous in many
other birds, suggested the possibility that here, as in the
Sebright male fowl, the suppression of the male plum-
age might be due to substances developing in the testes.
Unlikely as this seemed (because pigeons with diseased
testes would probably have occurred and any change re-
corded), nevertheless I tried the effect of castration on
one young F, male that was just weaned. Some feathers
were removed at the same time. The bird was kept
for about five months and did not show any change in
its plumage. It appears probable, then, that there are
no genetic factors in pigeons, like those js the Sebright,
which, acting through the testes, suppress the develop-
ment of the plumage in the male.

8 See Cole et al.

TEI ESS ES CIN mes
=f Rg pac E OF re, CESS PL e a a a Rep ny ore rg i

No. 613] INHERITANCE IN FANTAIL PIGEON

REFERENCES
Browne, R. Staples.—On the rias: of verga in Domesticated Pigeons
with Special aparine ce to Reversion. Proc. Zool. Soc. London, 1908.
PR os dde mes o Taertance = the DAS Character in Pigans

p. 36.

Morgan, T. HL Notei on Two Crosses Betwesk Ditea Races of Pigeons.

Biol. Bull, XXI. 1911.

Morgan, T. H., Sturtevant, A. H., Muller, H. J., and aan: C. B. The
cabana of Mendelian Par New You. 1915, p.

Cole, J. A Case of Sex-linked Inheritance in the ds Pigeon. Science,

1912
, + 3
Darwin, Chas. Animals and Plants under Domestication.

A PRELIMINARY REPORT ON SOME GENETIC
EXPERIMENTS CONCERNING EVOLUTION

RICHARD GOLDSCHMIDT

Tue nature of the gene, the variability of factors and
the effects of selection are favorite topics of recent dis-
cussion, which is well known to geneticists. The latest
publications of Jennings and Castle will stir up anew the
uncompromising parties and lead to new discussions. We
think it advisable, therefore, to give a brief account of
certain parts of a very large body of work on fundamental
questions of evolution which we have carried on during
- the past nine years, with the collaboration of Dr. Seiler
and Dr. Poppelbaum. Although some parts of the work
have been finished for some years, we do not intend to
publish a full account until all the details are worked out.
But as certain results have already allowed us to form
definite views in regard to some fundamental questions
of evolution, we may present them, together with ex-
amples of the experiments in question.

The majority of the experimental work in regard to
the fundamental problems of evolution has been done
with domesticated animals and their mutations (rats,
Drosophila) or with Protozoa, which present the compli-
cation of asexual reproduction. We have directed our
attention to experimental analysis of such phenomena in
nature, which must give basic information about evolu-
tion, and we have studied the following phenomena:

1. The Geographic Variation of the Gypsy-moth.—This
well-defined species is spread over a great part of the
globe. In different habitats, however, different races are
found. How many of these exist can not be stated, but
the number must be extraordinarily large, as we know
but two localities where the same race is found. We
have found all the races to be perfectly fertile with each
other, with the exception of one combination which has
never been successful. We have studied and are still
studying the genetics of a large number of these races.

28

No. 613] EXPERIMENTS CONCERNING EVOLUTION 29

2. The Melanism of the Nun-moth, Lymantria mo-
nacha.—The nun is one of the moths which have devel-
oped melanic varieties within recent times; and these
melanic varieties, which were extreme rarities not many
decades ago, have almost supplanted the original white

orm. We have worked out the genetics of this case and

shall publish the details when conditions permit. Some
of the results were read before the German Zoological
Society in 1911 but no abstract was published.

3. The Genetics of Alpine Varieties, especially of
Parasemia plantaginis and the Italian Races of Calli-
morpha dominula.—This work has been broken off by the
war, but some of the first results are available.

We shall begin with a few facts concerning the geo-
graphic variation of the gypsy-moth. We have here a
form that is spread all over Europe, through Siberia into
China, and all over Japan, infesting; furthermore, part

Fic. 1. Types of Caterpillars from different races after the first moult.
Drawin, by Mr. Yokoyama, Tokyo, 1914.

of the Atlantic coast of the United States. We have
studied races from different parts of Europe and Japan
and the Massachusetts form and we have found different
forms in comparatively near-lying regions. Thus the
races from the Rhineland, Silesia and Hungary are dif-

30 THE AMERICAN NATURALIST [Vou. LIT

ferent from each other and from the Massachusetts race
(probably imported from France). All of them are dif-
ferent from the Japanese races and these again differ
in the different parts of Japan. The characters of dif-
ference are manifold; we shall confine ourselves here to
a single character, more interesting and more character-
istic than the others—the markings of the caterpillars.
Fig. 1 shows caterpillars of a few races after the first
moult. We see here some of the transitional stages from
a very light to an almost black caterpillar. The genetic
study of this character of marking shows that we are
dealing here with a primary type of marking which be-
longs to the entire group of moths in a similar form, that
is, the light pattern. All the darker forms have the same
genetic basis of marking, on which, however, dark pig-
ment encroaches increasingly until the markings prac-
tically disappear. We may now divide this increasing
melanism into ten classes and place the lightest individ-
uals in Class X and the ones without marking in Class I.
It must be added that the dark series extends beyond Class
I, but the difficulty of classifying them is such that no
darker classes have been adopted.
The young caterpillars of the different races show
markings which fluctuate around a mean at a certain point
of the series and this behavior is remarkably constant
for the different races. The following Table I gives a
few polygons for different European and Japanese races.

TABLE. I
CLASS FREQUENCIES IN PER CENT.

Breed | Race I Ir Tit Iv T VI VUI | YDI; X x
WAR as H .. | 38.6| 54.5| 6.9
UAM: y K 15 | 64.7 | 20.3
WAIT 2... F 1 38.5] ..
Aeros... O 8:2| 48.8 | 31.4| 11.6
w Eo po ey ars AS 1 1.2
Mia ua A EAn OR a e i
WAla...... S+ |.. 154! 19.1| 44 | 30.2] 13] ..

WA56...... M, S 100

No. 613] EXPERIMENTS CONCERNING EVOLUTION 31

In crosses of these different types F, is about inter-
mediate, as some curves in Table II show.’

TABLE. II
F,. CLASS FREQUENCIES IN PER CENT.

Breed | Cross EL | ar} rere Yo Pe VER we e | x
AO BRE TI 17.3 | 29.6| 39.9 | 13.2
¿AAN Moa oes 9.2| 33.3| 14.4} 42.2 9
MAGE Gs. vis SXxA | ES nee ae 45.4| 28.7 | 25.9
MAST Li. AXS | 9 120}29/121 115 | 13 3 i
VAL. ORKO FE E 4.7! 67.4 | 27 de we
Vie catas Md ox Ss” | 9 9| 22.4 0. 35
WASTE ce bre cw ko chk ES ba 49.6; 49 | 0.7
WAB... | GXM a) ae | ae loa 8 4
WAS8...... [EMX AR] 10 | 70 | 20 Ei

And F, gives a 1:2:1 ratio, or 3 light + medium: 1 dark,
whatever races are involved. (This statement should be
taken only on its face value. As a matter of fact, we find
here, within the invariably present ratio of 3:1, very
strange details of the kind described as ‘‘gametic con-
tamination,’’ and, furthermore, an obscuring of the ratio
in earlier stages followed later by the right ratio, ap-
parent lack of segregation, etc. From a purely genetic
point of view, the analysis of these phenomena consti-
tutes the most interesting part of this work, but it has no
special relation to the problems here under discussion.)
Back crosses, however, give a 1:1 ratio. The following
table gives a few data of this kind.

TABLE IT
Breed Fa From Dark, Per Cent. mar ee

12. at SXK 23.8 76.2
WAISE SR ees KXS 25.5 74.5
b.. SXKy 25.9 74.1
D EEA E a S 25.8 74.7
Bot LAA OXS 26.2 73.8
WAI... cores SsxH 30.2 69.8
A OXM 22.1 77.9
DAD oa MXH 25.8 74.2

TATA. eee. GXM 25 75

ene give here a few random examples. The amount of actual material

is very large, as more than 100,000 caterpillars have been bred and studied.
We Pee
crosses, ote.

32 THE AMERICAN NATURALIST [Vor. LIT

The actual curves look like the example in Table IV of
an F, cross.
TABLE IV
ZA9 F, From M X H

1 | II | mr | rv | v | ve | var | vom | xx | x
| 8 Jo, | ao [i5 [3 |.
| Mie Ceca O) INEA lo.

The sum of all the hundreds of curves shows that we
are dealing here with a case of multiple allelomorphism :
The pigment factor, producing the gradual covering of
the markings, is present in the different races in different
degrees, all being allelomorphic to each other. :

Thus far we have dealt only with the very young cater-
pillars. Their further history in regard to the effect of
these factors leads us one important step further. We
find mainly the following types of behavior within the
pure races: (1) Light marked caterpillars, which remain
practically the same throughout the entire larval life.
(2) Light-marked caterpillars which grow darker with
every moult and finally are about medium or more than
medium dark. (3) Light-marked caterpillars which
change during the larval stage, so that they finally are all
dark. (4) Medium light caterpillars of different degrees
changing to dark during larval life. In the following
tables we give a few examples of these races, showing the
shifting of the type of marking during the stages of larval
life. The large range of variation after the third and
fourth moult visible in these tables is due more to a dif-
ferent speed of shifting in different individuals than to
the initial variability. This is shown in Table VI, which
gives an example of the shifting of types of pigmentation
during the larval stages for a series of isolated individ-
uals of some of the pure races.

The genetic analysis of this phenomenon seems to re-
veal the real nature of the multiple allelomorphs, which
cause these different types of pigmentation and their

behavior during development of the caterpillars. With-

No. 613] EXPERIMENTS CONCERNING EVOLUTION 33

TABLE V
EXAMPLES OF SHIFTING RACES
Race H
Class Frequencies in Per Cent.
Stage of Caterpillars I Il TI IV Vv VI | VII VIII Ix ed
3 ae Sap A ol ae ae eae 38.6} 54.5) 6.9
4 oe NA E. a. 11.8 55.31 31.7] 1:2
5 29.31 2441 O71 12.2} 98 1 12.2 $ Lai i Ak
6 64 16 16 4 a a
Race A
Stage of Caterpillars I | 11 | tto rv v vI | vit | vit] rx | x
3 Da ae ok La 19 43 38
4 lo e 2d ae 35.51 58.1 6.4
5 14.8) 42:01 471 95] 386I is
6 100
Race G
Stage of Caterpillars ý | nu | m | Iv | v Vid yn) Vee ie x
|
3 By REE 11.8 31.2
4 as iy Ye OFLAF A 338i 28.2 1.4
5 201.73 148) 18.125 = 9
6 100 |

TABLE VI
EXAMPLES OF SHIFTING AS OBSERVED IN INDIVIDUALS WITH DIFFERENT
UMBER OF MOULTS

Class of the Individual after Moult
Race Individual

2 | 3 4 5 6 Sex
Hon ZA3.6 Ix VII II ref
Ho oe, ZA3.7 VIII I g'
aa ZA3.4 1 Ill I Q
coda ZA3.1 VIII VII I Q
A ZA4.1 VI Ill g
"OA ZA4.3 Vv Ill g
a e ZA4.4 VI V IV I Q
et ZA4.8 VII V III II ọ
T ZA4.13 VI y y I Q
Go ZA6.11 VIII VIII VI II Ore
ae eon ZA6.7 VIII VI VI IV g
ware ey: ZA6.5 VIII VIII VI II 1%
ee ZA6.6 yu A VILE IV IH iope
To nl LE | VI VE eee ur A OG A

out $ going into details, which would, necessitate a multi-
tude of en) | following points

f on Eo lO al

34 THE AMERICAN NATURALIST [ Vou. LIT

are of importance: (1) F, between a non-shifting light
race and an always dark race is intermediate, or some-
what lighter in the beginning. But by progressive stages
the hybrid caterpillars shift over into the dark classes.

First row: hybrid between the two par races after se
moult. Beca row: Hybrid (left) and = Rte ap race > al after pe
moult. Drawings by Dr. Poppelbaum, 1912

Fig. 2 represents caterpillars from a cross of this type.
(The exact curves that belong with these pictures are re-
produced in our ‘‘ Einführung in die Vererbungswissen-
schaft,’’ 2d edition, 1913, p. 170, Fig. 66, as an example

No. 613] EXPERIMENTS CONCERNING EVOLUTION 35

of change in dominance during development.) (2) The
same thing happens in certain F, crosses, reversing com-
pletely during larval life the original ratio of lights and
darks. (3) The different races involved are characterized
by a difference in the speed of differentiation, as shown in
the actual curves. This velocity is also caused by genetic
factors. Where these recombine with the pigmentation
factors, the entire situation of the F, curve is shifted
(without changing the 3:1 ratio), showing that the visible
effect of the pigmentation factors is bound to a certain
velocity of differentiation. (4) The shifting of the type
of pigmentation from light to dark during larval life of
certain races or hybrids is a process which progresses
constantly with time. This is seen when isolated indi-
viduals are studied which belong to races that differ in
regard to the number of moults and exhibit the shifting
simultaneously. There are races where all the male cater-
pillars have four moults and the females either four or
five; other races where the males have four, the females
five; others where both sexes have five moults; and in the
last case even a sixth moult occasionally occurs. In these
cases we see that every new moult produces a further
shift to the dark side of the curve, showing that the class
of pigmentation to which a full-grown caterpillar belongs
is in this case a function of the time of differentiation.
The same fact can be demonstrated in a shifting race by
prolonging the time between two moults by starvation
(which succeeds only to a certain extent). In experi-
ments of this sort it has been possible to get the shifted
type of pigmentation, characteristic of the fourth stage,
in some individuals in the third stage. Table VI also
contains a few random data on the first point. (6) In
shifting hybrid cultures there appear comparatively
often mosaic-caterpillars, showing different classes of
marking on right and left sides. The distance between
these two different classes is approximately kept up when
shifting occurs during development. The following ex-
ample demonstrates this fact:

36 THE AMERICAN NATURALIST [ Vou. LIT

TABLE VII

Class After Second to Fifth Moult

2 | 3 | 4 | 5
| i
| Right | Left | Right | Left

Right | Left | Right | Left

Witty bli winn ll ee ad eee

These caterpillars always give normal moths and nor-
mal offspring.

A careful consideration of these poni: shows clearly
what these multiple allelomorphs for pigmentation really
are: They are different quantities of the substance which
we call a gene which act according to the mass-law of
chemical reactions, i. e., produce a reaction or accelerate
it to a velocity in proportion to their quantity. In our
special case it means that the factor stands for a metabolic

QUANTITY OF PIGMENT

Fre. 3.

activity proceeding with a definite velocity dependent
upon the quantity of the factorial substance present.
This activity finds its visible expression in the deposition
in the skin of increasing quantities of certain products of
protein decomposition which as chromogens are oxidized
into melanin pigments. The effect of the different quan-
tities of active substance (enzyme?) which we call the
multiple allelomorphs upon the progressive pigmentation

No. 613] EXPERIMENTS CONCERNING EVOLUTION 37

of the caterpillars is then represented by the graph on
p. 36 (Fig. 3).

Given a definite quantity of the factorial substance and
identical conditions, the velocity of the reaction is con-
stant. Thus the final result depends upon the amount of
the factorial substance present and the independently in-
herited rapidity of differentiation, which determines the
situation of the growth-stages on the abscissa (the dotted
lines). Thus the above quoted facts as well as the multi-
tude of details not mentioned can be easily derived from
this graph. The last named mosaics are of course the ex-
pression of small differences in the velocity of differen-
tiation in symmetric halves of the body, which are well
known to embryologists.

These conclusions in regard to the real character of
multiple allelomorphs are the same as those derived from
other characters in the same objects. In our work on
intersexuality we were able to prove, to as great an extent
as a genetic proof can possibly be carried, that the dif-
ferent geographic races of the same moth differ in re-
gard to the absolute and relative quantities of the sub-
stances, which we call the sex-factors. In the genetic
language of the present day we should call them, there-
fore, multiple sex-allelomorphs, a conception which indeed
we have always used (without this recent term) since our
first report about this work in 1911. In the case of inter-
sexuality we can furnish facts very similar to those about
the caterpillars, if we consider certain features of the
wing colors. In normal males a certain amount of pig-
ment covers the entire wing, whereas the female wing is
unpigmented. This pigment is formed by the oxidation
of a chromogen deposited within the scales. There it
flows from the wing veins with the blood. By a detailed
analysis we are able to show that an intersexual male is a
genetic male which developed as such up to a certain
point when the development suddenly began to continue
under the aspects of femaleness. One of the results of
male metabolism is the phoma es of these chromogens

38 THE AMERICAN NATURALIST [ Vou. LIL

in late larval stages. This production is therefore
stopped when female metabolism sets in; when then the
time arrives in development, when the chromogen spreads
over the wing scales, its available amount is proportional
to the relative lateness of the reversal of sex. Therefore,
with increasing intersexuality, the pigment flowing from
the veins covers a smaller and smaller area of the wing,
finally being confined to the neighborhood of the veins.
As? the analysis of the other intersexual organs allows
an accurate determination of the time factor involved, we
have here a very close physiological parallel to the facts
about the caterpillars.

In most other cases of multiple allelomorphism only
the results can be seen, and it will be difficult to work out
the time factor, which proves that the multiple allelo-
morphs are different quantities of an active substance.
(Some botanical subjects ought, however, to be favor-
able.) But in comparing the other facts about multiple
allelomorphs with our cases, we feel confident that, where-
ever a similar analysis can be applied, the results will be
the same. For example, all the cases of quantitatively
different pigmentation, which are of multiple allelo-
morphic nature, like Castle’s hooded rats or our different
cases of melanism in moths, show*that the effect of the
different factors is that different quantities of pigment
spread from different ‘‘points of outlet,’ which of course
are hereditary traits of the species or group; the similar
effect, therefore, leads to suspect a similar cause.

If our conclusions regarding the nature of multiple
allelomorphs are accepted, it must lead to a different intel-
lectual attitude toward the problem of variability of
genes, which is so important for evolution. The opposi-
tion to the view has been, we believe, primarily on aprio-
ristic grounds. In the long controversies of recent years
regarding the interpretation of Castle’s work the logical
side of the case seems to have always been in the fore-
ground. The same is the case when E. Baur calls our

2See pictures in Jour. Exp. Zool., 22, 1917, pp. 614-15.

No. 613] EXPERIMENTS CONCERNING EVOLUTION 39

views in regard to the variability of the sex-factors
a priori inadmissible. We believe that this intellectual at-
titude toward the problem is the result of Johannsen’s
doctrine of agnosticism in regard to the nature of the
gene, which resulted in a kind of mystic reverence, ab-
horring the idea of earthly attributes for a gene. (Our
distinguished opponents will excuse this somewhat ex-
treme statement.) If, however, it can be proven that
genes are substances with the attribute of definite mass,
it would be illogical to deny their variability. Nobody
will claim that a gene is a substance that passes unaltered
from generation to generation. The elementary facts of
development and regeneration show that this substance
grows, at least, and increases in quantity. If, now, the —
substantial basis of heredity in the sex-cells is established |
by the assembling of all the factor-substances in their
characteristic quality and their correct quantity, the sit-
uation is the same for the gene as for any/other organic
process: the varying conditions of the surroundings of
the gene cause a certain amount of fluctuation in its quan-
tity. This conclusion entirely changes /the logical aspect
of the question, whether or not a change of the gene by
selection of variants is possible.

The strongest point.of the anti- selectionists was that it
is absurd to assume that a selection of somatic fluctuation
has anything to do with the characters of the germ-plasm.
With the quantitative view, however, which we believe to
have proven in two elaborate cases, this situation changes.
The somatic character in question, say amount of pig-
mentation, can only change toward a plus or minus side.
This change is caused directly by a difference in the
velocity of the reaction of some metabolic process which —
results in the deposition of pigment. Such a change of
velocity of reaction, however, can be produced either by
the action of the medium, and then it is a modification, or
by fluctuation in the quantity of the gene, causing increase
or decrease in the velocity. The resulting variation is of
course, phenotypically, the same. Selection, therefore,

40 THE AMERICAN NATURALIST [ Vou. LIL

may be ineffective, if a modification only is selected; it
will be partly successful if a combination of plus-quantity
with plus-modification is selected; and fully successful
if the exclusive result of plus-quantity of the gene is se-
lected. The deus ex machina modifying factor, which,
moreover, does not fit the decisive genetic facts in the
most discussed case of Castle’s rats nor our cases, thus
becomes superfluous.

It is, moreover, perfectly logical to assume that selec-
tion of either plus or minus quantities of the genes
changes the mode of the fluctuation of this quantity corre-
spondingly in the succeeding generation. If the different
quantities of the substances, which constitute the systems
of multiple allelomorphs, are inherited, then every other
quantity is also inherited. If the presence of the quan-
tity p in the germ cells of the parents causes the reap-
pearance of the quantity p in the germ cells of the chil-
dren, the same fact applies to the quantities q, r, s—to
every quantity which is present or has been selected. Se-
lection can, therefore, change the quantity of the gene,
and also, therefore, the somatic characters caused by
quantitative differences in the gene, until the physiolog-
ical limit is reached. This limit may be the limit for the
character in question—for example, no pigment, self-
color—or it may be the limit set by the necessary coor-
dination of developmental processes. For example, in
the development of a moth a certain gene causes, at a cer-
tain moment—during pupation—the evagination of the
imaginal disks of the antenna. The correct quantity of
the gene causes this process to take place at the correct
time. A quantitative variation of the gene would cause
the evagination to take place at the wrong time. We
have, indeed, had strains of caterpillars where in many
individuals this process took place in the last stage of the
caterpillar, giving caterpillars with pupal antenne, The
quantity of the gene in question was in these cases not co-
ordinated with the other genes and the action was pro-
duced too early. It is evident that quantitative changes

abe a Z :
la EE TA ON pS
O ee a ee E AEE

di ices ell

No. 613] EXPERIMENTS CONCERNING EVOLUTION 41

of this kind will lead to physiological impossibilities,
monsters, etc. Here, then, is again the limit for selection
of factorial quantities. It need hardly be added that such
selection is necessarily orthogenetic.

Our own experiments in this line are, as far as they go
at present, in perfect accord with Castle’s work. We have,
moreover, applied another type of experimental test,
namely, selection in F,. If a given pair of multiple al-
lelomorphs differs in regard to the quantity of the fac-
torial substance and this quantity is subject to fluctuation
around a mean, the variability of the character in F, is
caused by the usual agencies producing fluctuations as
well as by the different combinations of the parental
quantitative values. Selection in F, ought, therefore, to
influence the curve in F, in a certain number of cases,
namely, when the plus or minus individuals are genetically
plus or minus. Within the normal segregation of light
and dark individuals in the 3:1 ratio a shifting of the
mean for lightness and darkness must take place. In a
series of such experiments we had a number of positive
results. The following Table VIII may serve as an ex-
ample:

TABLE VIIT

F, WITH SELECTION IN F, From Cross K X S

In Third Stage,

vi vu ¡var Ix | x

I elmin

5

s AS | 131127 9.1 30 154 = | g
Minus selection.. ........... 25 15.7|26.9115.51 17.2 3.216.651.. 1. 1.

We believe that these facts and interpretations have a
definite bearing on the problem of evolution. The first
step in the differentiation of species which occurs in na-
ture seems to be the formation of geographic races. The
entire bulk of modern evidence in ecology tends to show
the existence of clearly defined local forms for very re--
stricted areas. For example, the ichthyologists differ-
entiate forms of Salmonids and Coregonids for prac-
tically every river and lake; in the same way in the lower

42 THE AMERICAN NATURALIST [Von LI

organisms, like Daphnids and Rotatoria, different forms
appear in different regions. The ornithologists describe
different races for every river basin of the affluents of the
Amazonas; the mammalogists do exactly the same thing

for every area which was thoroughly covered. Where
- breeding experiments have been carried on it has been
shown that the geographic races may be perfectly fertile
with each other and may produce fertile offspring. In
some cases, however, the transitional stages toward steril-
ity are found. Thus the production of intersexual moths
in crossing geographic races can be regarded as a step
toward increasing incompatibility, which in one of the
crosses attempted by us was an absolute one. In other
cases only a small percentage of the offspring of the hy-
brids could be reared, as in the crosses of the North and
South European Callimorpha dominula. We, therefore,
with many evolutionists, feel convinced that the geo-
graphic races are the most important visible steps in
species-formation in nature.

If we now look into the characters distinguishing geo-
graphic races, we very often find certain qualitative dif-
ferences most conspicuous, for example, exchange of red
and yellow color in the moths. A close study of definite
examples, however, reveals that these differences are
often more conspicuous than important. This is shown
by the only group of information in the animal kingdom
which we have both by ecological and genetic work—the
geographic variation of land snails. The facts about the
extreme variability of Helix, Achatinella, Partula, ete.,
are well known, as well as the irregularities in the con-
finement of definite types to definite localities. We have
been so fortunate as to gain some insight into these facts
through a very interesting collection which Dr. Haniel
made in Timor and studied under our direction (not yet
published). It was evident here, as in the other cases,
that a series of unit factors for number, color, form of
bands and ground color, which recombined freely, was
involved. And practically all the combinations could be

No. 613] EXPERIMENTS CONCERNING EVOLUTION 43

reduced to the genetic factors which Lang worked out for
Helix. But, exactly as in the classic cases, there was no
possibility of stating a definite relation of these factors
to the grouping according to localities. In some locali-
ties certain factors or combinations did not occur, but the
attempt to classify the material along this line proved a
failure. However, every group from each locality ex-
hibited beside these factorial recombinations certain
quantitative characteristics of size, proportions, etc., of
the shell which were characteristic for definite localities.
These, however, are the characters which probably fall in
line with those caused by the quantity of the genes.

The difficulties which the facts of geographic variation
create for the conception of species-formation by selec-
tion have often been discussed. Bateson in particular
(“Problems of Genetics’’) serutinizes them from the
modern genetic point of view. They are indeed insu-
perable if all characters which show variations and recom-
binations are considered from this point of view. The
extreme irregularity, for example, of the local combina-
tions of types of shells in Helix, Partula and Achatinella
makes it impossible to regard them as local adaptations.
This is certainly true, but may be without any bearing on
the species question at all. The factors and recombina-
tions occurring in Helix, Achatinella and Prodromus are
more or less the same, just as are the recombinations of
coat colors in different rodents. They constitute a set of
mutations and their recombinations which are proper to
the type of germ-plasm of the group. They occur, re-
combine or fail to appear as chance wills, and seem to have
no special selective value. We do not think that these
are the characters which play a part in the evolution of
species; they are, in most cases, independent of adapta-
tion.

There are, however, reasons for supposing that such
differences of characters as are based on the quantitative
differences of the gene are those which are influenced by
selection and are important for the formation of the first

Hiei

44 THE AMERICAN NATURALIST ~- [Vou.LII

steps toward diversification of species. We base this
opinion on the following facts:

One of the few cases where selection in nature has ap-
parently been seen at work under our eyes is the much-
quoted case of melanic moths. ‘We started in 1908 to
work out the case of the nun, Lymantria monacha. The
dark varieties of this moth have been known as rare oe-
currences for over a century. But only during the last
decades have they spread and almost replaced the white
forms. The analysis of the genetics of this case shows
that the dark form is a dominant mutation to the white
and that the many different stages of darkness, which
form a complete series from white to black, are produced
by sex-linked multiple allelomorphs. (Unfortunately,
the interesting details can not be given at present.) How
is it, now, that these combinations have come to replace
the original form? Many hypotheses, some of them very
strange, have been put forward; but it seems to us that
the case is comparatively simple. The dark forms are
stronger, more lively, better fliers, as far as we can tell
from our experience with the animals in captivity. They
are also larger (see Fig. 102, p. 267, in our “Einfúhrung
in die Vererbungswissenschaft,’? 2d ed., 1913). The
melanism is in this case, therefore, only the most con-
spicuous superficial feature of a quantitative and pro-
gressive change in a gene which causes a definite meta-
bolic condition, resulting in hardiness as well as in the
deposition of more pigment in the wings. The quantita-
tive change has here a superficial expression and is there-
fore easily recognizable. But this visible pigmentation
is not the really important character. How is it, then,
that these melanic forms, and other forms in similar man-
ner, have established themselves so suddenly? We may
venture to point to the facts that the selection, as has
often been stated, has occurred especially near the larger
cities, and that the period during which this selection has
taken place is the period of industrial development, i. e.,

of restriction of forested areas near the cities. It is,

IO a e a AS Aee sh
eS ER ARA R AEN PELE E E a A A E

So ee

No. 613] EXPERIMENTS CONCERNING EVOLUTION | 45

furthermore, the period of scientific and intense forestry
and of economic entomology. Here we have the probable
agencies that made life difficult for the moth and gave a
great selective value to that advance in hardiness which
lies behind the melanie appearance.

We should point out here the difficulties which arise in
the criticism of definite views of evolution on the basis of
facts not analyzed genetically. The selective value of a
climatic character may often be doubted on the ground
that the'same type occurs in a very different area ad-
mixed with the local form. But genetic analysis may
often show that what appears to be the same type is in
reality a different thing. The north European Aretiid,
Callimorpha dominula, has wings marked with red; the
Italian form has wings marked with yellow. In certain
localities (one of them near Berlin) a yellow sport of the
red form regularly appears, apparently the same form as
the Italian one. We, as well as others, have crossed these
forms. The yellow sport is a simple recessive to red and
segregation occurs in the 3:1 ratio. The Italian yellow
form, however—at least the ones from the Abruzzi, which
we used—crossed with the red northern form, produces
intermediate orange in F, and in F, every shade from red
to yellow. The two yellows, which look alike and prob-
ably are chemically alike, are nevertheless products of a
different metabolic process. In the sport the same met- —
abolie process which usually leads to red pigment is
changed by mutation only to the extent of the color change
in the end-product. Inthe southern form a different type
of metabolism results in the formation of yellow pigment, |
and the cross is therefore an entirely different cross, with

different results? As a matter of fact, the latter cross
` shows very much diminished fertility, as Standfuss has
already pointed out. This shows how unsafe the ground
is on which criticism of evolutionary questions without
genetic test is based. - That our example is not an excep-
tion is proved by the fact that Standfuss long ago formu-
3 We may point out that herefrom a rational interpretation of dominance

and blending can be derived. — | :

46 THE AMERICAN NATURALIST [ Vou. LIT

lated the rule, that when two forms coexist in the same
locality and are able to interbreed, they do not produce
intermediates; but when the forms are geographically
separated as local races, crosses between them result in
a series of intermediates. Bateson says: ‘‘In this apho-
rism there is a good deal of truth.” We think that the
rule expresses the difference between a non-adaptational
chance mutation and the adaptational change in the fac-
torial quantities which may lead to a similar-looking, but
physiologically different character. This character, al-
though, like the non-adaptational one, is itself of no
selective value, is the result of a general physiological
change which does have.a selective value.

This will become still more evident if we return once
more to the study of the gypsy-moth. In studying the
relations of the different geographic races as character-
ized by the multiple-allelomorphic characters in question,
we find that these characters are paralleled closely by dif-
ferences in the life-cycles. Without going into details,
we may state as a fairly general rule that the races with
high degrees of pigmentation in the later stages are the
ones which show a fast development, comparatively short.
larval life and a long period of hibernation. The light
races have a comparatively long larval period and a cor-
respondingly short period of hibernation. The former
races, furthermore, inhabit the areas where a long and
cold winter occurs, while the latter are endemic in places
which have a hot summer, early spring and mild winter.
One might think that these different characteristics were
simply the direct effect of temperature conditions. But
that this is not the case is shown by the constancy of the
differences when the races are bred in a different climate
and also by experiments on the physiology of hibernation,
which have convinced us that the time relations of the
life-cycle are—of course, within the limits of fluctuation—
a heritable trait of rhythmic character. These facts show
where the adaptational character of the differences of the

geographic races lies: the adaptation which fits the differ-

No. 613] EXPERIMENTS CONCERNING EVOLUTION 47

ent milieus is the life-cycle (in a broad sense). The visi-
ble distinctive characters of the races—aside from addi-
tional mutations of a non-selective nature—are nothing
but the products of reaction of different types of metabo-
lism, allied with the different time relations of the cycle.
The method of the formation of geographic races in this
case must, therefore, be the following. The first con-
quest of a new territory is of course only possible when
the animal is preadapted, along general lines, to the new
medium. But that it can maintain itself depends upon its
power of special adaptation. The gypsy-moth, for exam-
ple, has repeatedly been brought into England, but it has
never established itself there. In the case of this form
the special adaptation means the coincidence, in the first
place, of the life-cycle with the seasonal cycle in nature.
And it is here that all the discriminating effect of selec-
tion comes in. The quantitative changes of the genes
which cause the time relations of the cycle are then the
material for selection, and selection acts according to Dar-
winian principles until the equilibrium is established.
Thus the genetic study of the quantitative changes of the
gene reveals anew the truth of Darwin’s conception.
Furthermore, we see here how sterility of hybrids or com-
plete incompatibility of new forms may arise. We have
proved that the quantitative differences of the sex-factors,
which are themselves nothing but adaptations to the time-
relations of the cycle, are among the characteristic differ-
ences of these races.* There are, moreover, responsible
for the incompatibility in regard to sex which results in
intersexuality after crossing. Changes of exactly the
same type may easily make any cross-breeding impossi-
ble, since no organism can develop unless all the processes
of differentiation are coordinated in respect to their ve-
locity. Here we see, seen why goog ipe races are
so often uniform and ar d by certain traits of

rere also Pfliiger’s and R. Hertwig’s work with frogs and Cuénot’s
th starfish, ae similar facts in regard to geographic varia-
dics of sexuality. :

48 THE AMERICAN NATURALIST [ Vor. LIT

a quantitative character even when additional mutations
and their recombinations make them at first sight appear
diversified. This uniformity indicates the adaptational
type produced by selection of the quantitative variations
of some vital gene; the differences are only a difference
in apparel.

In conclusion, we may point out three groups of facts
which, of the greatest importance for evolution, have
always been a hard nut for the mutationists to crack.
The first is the series of temperature-experiments in Lepi-
doptera—and similar experiments in Amphibia, Crus-
tacea, etc.—that lead to the production of aberrant forms
which resemble closely certain geographic varieties. But,
with the exception of certain often-quoted cases, these
aberrations are not hereditary. Inthe light of our experi-
ments these facts are not surprising. The effect of the
temperature experiments is to change the normal time-
curve of certain metabolic processes. The effect is, there-
fore, due to this change of one of the variables of the reac-
tions in question. The quantitative change of the sub-
stance of a gene, however, which we found to be at the
basis of the geographical variations, also produces a dif-
ference in respect to the time-curve and therefore the
same effect, this time a heritable effect. If we now select
the plus individuals in this type of experiment—and this
applies to all analogous experiments—we may simply se-
lect a modification. But we also may select the combina-
tion of a plus-modification with a plus quantity of the
gene in question. If the experiment is repeated, the next
generation will then show a still stronger reaction, or, if
the experimental influence is not repeated, there will be
an after effect of the experiment on the parents. It 1s
remarkable that such results, which were to have proved
the inheritance of acquired characters, always turned out,
when characters relating generally to the life-cycle were
in question, characters which also appear in the geo-
graphic races of the form. Extreme mutationists used to
deny or disregard these facts. Here we have a simple

No. 613] EXPERIMENTS CONCERNING EVOLUTION 49

explanation for them which both does justice to the facts
themselves and falls in line with modern genetic views.

Furthermore, we now see the exact meaning of Dar-
win’s view, which he had to express in a somewhat am-
biguous way on account of the lack of experimental data
which would have permitted clearer expression. His
essay of 1842, the forerunner of the ‘‘Origin of Species,”
begins with the words: ‘‘An individual organism placed
under new conditions sometimes varies in a small degree
and in very trifling respects, such as stature, fatness,
sometimes color, health, habits in animals and probably
disposition. . . . Most of these slight variations tend to
become hereditary.” This statement shows clearly what
Darwin had in mind. If he assumes that some variations,
which are produced by change of conditions, are some-
times non-heritable, but tend to be inherited, we can now
explain what this means. The variations which, as geo-
graphic races, form the first steps in the formation of new
species are indeed exactly the same whether or not they
are inherited. Their direct physiological cause is also
identical, being a change in the rate of a definite process
during differentiation. Only the ultimate cause is differ-
ent; in the one case the original quantity of the gene de-
termines the rate of differentiation—which then is heredi-
tary—from the beginning; in the other case an outside
factor is active, retards or accelerates the same reaction to
the same degree. With this additional bit of interpreta-
tion, Darwin is right, after all.

The other group of facts includes certain details of
mimicry (mimetism). We believe that the general prin-
ciple of mimetism has been fully explained genetically by
Punnett. But there are certain details which his selec-
tionist opponents point out which constitute strong evi-
dence against Punnett’s view. We think that the most
valid argument against the Mendelian view of mimetism
has been derived from the facts about the parallel geo-
graphic variation of model and mimic. If our genetic
conception of geographic variation is correct, this point

50 THE AMERICAN NATURALIST [ Vou. LII

is not difficult to understand. If the resemblance of
model and mimic is based on the presence of similar
chance- or non-chance combinations of genetic factors,
and if geographic variation consists in the specific adapta-
tion of the quantity of certain genes to a required veloc-
ity of some vital reaction, it is very natural that similar
genes in model and mimic should be in exactly the same
situation and should undergo parallel changes.

The third important set of facts to be considered is the
problem of domestication. Darwin’s view is well known,
as well as the solution of a great part of the problem
through Mendelism. The latter shows that selection of
the recombinations after cross-breeding (besides picking
of mutations) is the chief source of success in domestica-
tion. (See our demonstration of this fact regarding the
improvement of pigs in “Einfihrung in die Vererbungs-
wissenschaft,’’ 2d ed., 1913, pp. 276-80.) That this fact
was well known to Darwin is shown, for example, in his
report about Lord Orford’s greyhounds (‘‘ Variation of
Animals,”” etc., Ch. 1). But he believed, in addition, in a
positive effect of selection of small variations. Wher-
ever he tabulates such characters, most or all of them are
quantitative characters of a kind which we can assume
to be dependent upon the presence of definite quantities
of a gene. Here we may have the solution of the diffi-
culties which the problem of domestication affords in
spite of mutation and recombination. No doubt the high
capacity for fattening was crossed into our hogs with
Asiatic forms. But selection of plus-quantities of the
responsible gene enabled us to obtain the character as
it stands to-day.

OSBORN ZOOLOGICAL LABORATORY,

YALE UNIVERSITY
New Haven, CONN.

MATERNAL INHERITANCE IN THE SOY BEAN

H. TERAO

THE ImPERIAL AGRICULTURAL EXPERIMENT STATION, TOKYO, JAPAN

THe soy bean, Glycine hispida Maxim., shows as differ-
ent types two cotyledon colors, yellow and green. The
beans with yellow cotyledons have two types of seed-coat
colors, namely, green and yellow, while the beans with
green cotyledons have always green seed-coats.t The in-
heritance of these types of cotyledons and of seed-coats
has been ‘proved by the author’s experiments to be ma-
ternal. A brief notice of the experiments will be given
in the following.

The green and yellow colors of cotyledons and seed-
coats are obviously attributed to chlorophyll, which,
on the ripening of the beans, is either changed from green
into yellow or remains green. Further, according to the
author’s observations, the chlorophyll in the vegetative —
parts of the plant shows the same behavior as the chloro-
phyll of the cotyledons; in other words, the leaves and
stems of the varieties with yellow cotyledons turn to a
yellow color when they are gradually dying coincident
with the ripening of the beans, while those of the varieties
with green cotyledons remain green sometime after the
dying of the whole plant. These facts suggest that the
two types of cotyledon colors may represent two kinds of
chlorophyll, one which changes into yellow under certain
physiological conditions and one which is not so affected.
The chlorophyll of the seed-coats, however, seems to be-
have somewhat differently from the chlorophyll in all

1Black and brown pigments also appear in the seed-coats of certain
varieties. These pigments are entirely independent of the green and yellow
colors here referred to in their inheritance, but they make the latter colors
invisible or at least indistinct. By proper crosses, however, one can test
whether a seed-coat covered by the black or brown pigment belongs to the
green or the yellow category. —

51

held A a e ot z y. ¿IE = na o —_
à j k > Pa ES ES A fa eee aly ANERE Pinhal A o A ee

[=]
=
(a
3
O
>
E __— _ _ _P_ _-_ z n mn
se? | v98'99 vez 18383 S6F'9T 10962 | PUI ON
z 2 gre OF ETT 918 SFT |U weg ‘ON
2 JUBISUOD JUBISUO0) JUB]SUuO0) yueysuoo quggsuoo o.
= -MOJ[OA -MO[[9A -MO[[9Á -MOT[OÁA -u9018 -uəə13 | *** “WR Ay
E Carsci%eso) (rez: %9m) | (yese: vrr) | | <
E 99 691 SZT : SIF | SIST. | 9€7 9F8 886 988'IT | 913 GSE SFT | 6318 PU "ON
= none: uaei3 | MO[9A | MO[SA udsds mops | uəəı3 | mos moped. > Mops | 9913 uoois | ua. u99Ia [> “pul
= gee Pe Mo os Ne ee | | | |
NZ g BSA 6 $ v GAZ OF £ € vz vz ““puy ‘oN
q U99.13 MOT[OA u99.13 MOTEL uəəı3 | MOTTO 199.13 Mood | u9d913 | uəəı3 | uo9l8 | woes |`" `` pura
S | | EE
= aes mops Ja MO[p9A | uoaa | uəəIF 199.13 u9913 | 9913 ab ‘dee a o eT
E uasa MOT[9Á Mojs MO[p9A | usa13 | MOTO mops Mops | woos | uses | uəəı3 | uəə17 |'''oppuog
= | suong
| suo suo 1809 suo 3809 suo a

E 1009-P098 | 014900 o. -015300 -po08 -01400 ia HUSO “pees -014100 -pos | 21800.
N 3 IA ‘ON Burssolo | A ‘ON 3UJS010 AI ‘ON Bussolo E III ‘ON 8uyssoip II ‘ON eS I ‘ON Suyssoip

SLVOO-GETG HO ANV SNOUWIALO) dO SUOTO) MOTIHA ANV NYAAH AO HONVIIMMIENT MHL ONIACOLS NI Adv SUSSOU) NVI A08

I Wav

N
10 >

No. 613] INHERITANCE IN SOY BEAN 53

other parts of the plant, since, as was already noted, yel-
low cotyledons are accompanied by green seed-coats in
certain varieties.

The crossing experiments which have been made by the
author since 1910 with these different types of beans have
produced the results shown in Table I, the main facts
being summarized as follows.

I. The F, cotyledons of the crosses reciprocal to each
other are of the same character as the female parents.
In respect to the cotyledon colors, the F, and following
generations show the characters of the F', generation ex-
clusively, instead of a Mendelian segregation between the
yellow and green colors. Hence we are probably dealing
with characters which can be inherited only through the
female parents.

TI. The inheritance of the seed-coat colors is a more
complicated phenomenon. In the cross ‘‘green cotyle-
dons, green seed-coat’’ (2) X “yellow cotyledons, yellow
seed-coat’’ (g), the green seed-coat is inherited through
the female parent exclusively, just as in the case of the
cotyledon colors; but in the reciprocal cross the green
and yellow seed-coats show Mendelian segregation, the
former being dominant.

The maternal inheritance observed above was not due
to self-fertilization succeeding failures in artificial cross-
ing, because several other characters showed inheritance
through the male parents.

An interpretation of the inheritance phenomena under
consideration is suggested as follows. In the first place,
let us refer again to the two different kinds of chlorophyll
assumed to be concerned in producing the green and yel-
low cotyledons; namely, the chlorophyll which can be
changed into yellow and the chlorophyll which remains
green. (These will be denoted respectively as “(Y)”
and ‘‘(G)’’ in the later descriptions.) These character-
isties of chlorophyll may be due to heritable traits of the
chromatophores or of the cytoplasm, and not to hered-
itary elements in the nucleus. As, on the fertilization of

54 THE AMERICAN NATURALIST [Vor. LII

the egg-cell, the chromatophores and the eytoplasm of the
female gamete will probably remain as such without
being supplemented by those from the male gamete, their
characteristics would naturally be inherited only through
the female parent. In the second place we may assume
that a pair of Mendelian factors is concerned in the inher-
itance of the colors of the seed-coats. The factor “H”
inhibits the chlorophyll ‘‘(Y)’’ in the seed-coat of the
beans with yellow cotyledons from changing to yellow,
producing beans with yellow cotyledons and green seed-
coat; the absence of the factor ‘‘H,’’ expressed by ‘‘h,”’
allows the seed-coat of the bean with yellow cotyledons to
remain yellow. The seed-coat of the bean with green
cotyledons remains green no matter whether the factor
“H?” is present or absent, because the beans of this kind
have the chlorophyll ‘‘(G)’’ which is incapable of chang-
ing the color.

The justice of the contention regarding the bean with
green cotyledons, moreover, is supported by the following
observations. The F, families of the crosses ‘‘green
cotyledons, green seed-coat’’ (2) X “yellow cotyledons,
yellow seed-coat’’ (4) were actually composed of two
kinds of individuals which were distinguishable from each
other by a slight difference of the intensity of green color
in the seed-coats, and the numerical relation between these
two kinds of individuals was approximately the Mendelian
mono-hybridal segregation ratio, the darker seed-coat
being dominant to the lighter one. Again, in the F, gen-
eration of these crosses, there were obtained three types
of families, two which were uniformly of the darker and
of the lighter seed-coats respectively and one which was
a mixture of both. By comparing the green seed-coats of
the female parents in these crosses with those of the prog-
eny, the former was found to belong to the darker class
mentioned above. These variations in the green color of
the seed-coats may be regarded as being due to the in-
fluence of the Mendelian factors ‘‘H’’ and ‘‘h’’ respec-
tively on the chlorophyll ‘‘(G)’’; from which it follows

No. 613] INHERITANCE IN SOY BEAN 55

that the method of inheritance in the beans with yellow
cotyledons obtains also in the beans with green cotyle-
ons.
Keeping these statements in mind the cases in Table I
may be illustrated as follows:

Crossing No. II

Parents (G)HH (9)X(Y)hh (9) (G)HH (9)X(Y)HH (2)
Cotyledons green yellow green yellow
Seed-coat green yellow green green

o d de.

1 (G)Hh (G)HH
Cotyledons green green
Seed-coat green green
Fs (G)HH (G)Hh (G)hh (G)HH
25% 50% 25% . 100%

Cotyledons green green
Seed-coat green _ green

ossing No. I

Parents Y(bh) (9)X(G)HH (9) Y(HH) (9) x(G)HH (g)
Cotyledons yellow green yellow green
Seed-coat yellow green green green

Fi Y(Hh) Y(HH)
Cotyledons yellow yellow
Seed-coat green green

F: = (Y)HH (Y)Hh (Y)hh (Y)HH

25% 75% 28% 100%
Cotyledons yellow yellow A yellow
Seed-coat green yellow green

If the foregoing interpretation really represents the
facts in this investigation, we may consider also crosses
in which forms such as (G)Hh, (G)hh, and (Y) Hh were
used as the parents, since in these crossings phenomena
different from those in Table 1 would be expected. These
expectations have been fulfilled in further experiments in
which individuals from the previous experiments repre-
senting different intensities of seed-coat color were used
as the parent plants. The results of these crosses, accom-
panied by interpretations, are shown in Table II.

56 THE AMERICAN NATURALIST [ Von. LIT

TABLE II

CROSSES MADE AMONG THE PROGENY OF THE HYBRIDS SHOWN IN TABLE I

Parents Fi Fe
No. of No. of
Female | Male | Character | Indi- | Char- | Indi-
viduals | 2Cte | viduals

Crossing No.
a O dra yellow | green yellow | 22 yellow | 2,381

d-coat yellow | green yellow

ance (Y) hh | (G)hh| (¥)hh | 100%

bo
bo

Crossing No.
WEEE os ees Cotyledons | yellow | green yellow 18 yellow | 1,963

Seed-coat green | green mo 10

j(Y)Hh | 50%

Interpret. | (Y) Hh) (G)hh 4 (y)hh | 50%

ig sie Cua Cotyledons | yellow ¡ green yellow 9 yellow | 1,108
Seed-coat green | green o 7

yellow | 2
fi HH| 25%

Interpret. (Y) Hh| (G) Hh < (Y) Hh| 50%

(Y) hh | 25%

The maternal inheritance described in this paper seems
to be essentially the same phenomenon as the inheritance
of the character ‘‘albo-maculata’’ which was studied by
Correns? in Mirabilis Jalapa and also by Baur? in Antir-
rhinum majus. In each case one is dealing with chroma-
tophore characters.

HARVARD UNIVERSITY, BUSSEY INSTITUTION,
2 Correns, C., Zeitschr. f. ind. Abst. u. Vererbungslehre, Bd. I, 1909,
pp. 291-329; Ibid., Ba. II, 1909, pp. 331-340.

3 Baur, E., Zeitschr. f. ind. Abst. u. Vererbungslehre, Bd. IV, 1910, pp.
81-102.

SHORTER ARTICLES AND DISCUSSION

ao oes IN MAIZE: THE C ALEURONE FACTOR AND
WAXY ENDOSPERM!

In 1912 Collins? presented data which showed a linkage be-
tween waxy endosperm and aleurone color in certain hybrids of
Chinese and American corn. A summary of the F, data in Table
II, p. 579,? gives the coefficient of association as .821. This is
equivalent, approximately, to a 3.5-1 gametic ratio, and a cross-
ing-over percentage of 22. The percentage of waxy grains is
about 21 and colorless about 25. This is good evidence that
Collins is dealing with material heterozygous for waxy endo-
sperm and heterozygous for only one factor in aleurone. In the
back eross data in Collins’s Table IV none of the ears shows the
1:1 relation between colored and colorless expected from plants
heterozygous for one color factor. The material in that table
apparently involves more than one aleurone factor in the het-
erozygous condition and before such data may be considered in
any linkage study they must be corrected for this or the true
values for the percentage of crossing over can not be ascertained.
The coefficient of association need not be used if we are dealing
with back cross data. If it seems desirable to use the coefficient
of association with this sort of data new tables should be calcu-
lated from the gametic series n:1:1:n corrected for the respec-
tive aleurone factor conditions.

The advantage of back cross data is obvious. Data of this
nature obtained by the writer from crosses of plants heterozygous
for one aleurone factor and for waxiness with double recessive
plants are presented in the table on p. 58.

Families 6, 99, and 100 are derived from dlid corneous
seeds heterozygous for aleurone and waxiness. Families 9 and
101 are colorless waxy plants. The first nine ears give an average
crossing over of 26.7 per cent. or a gametic ratio of 2.75:1. Ears
8 and 9 show repulsion instead of coupling, but this does not

1 Paper No. 66, Department of Plant Breeding, Cornell University, east,
N. de

““Gametic Coupling as a Cause of spread AMER. NAT., +, pp.
569-590. 1912.
57

58 THE AMERICAN NATURALIST [ Von. LIT

necessitate a separate summary for the crossover and non-cross-
over classes of the coupling and repulsion families. The devi-
ations from the average are more than twice the probable error
in ears 1, 3, 4, and 9, between one and two times in ears 5 and 7,
and less than the probable error in ears 2, 6, and 8.

Back Cross DATA: HETEROZYGOUS CORNEOUS COLORED X Waxy COLORLESS —

a z om PESE AES gee

gi momoe (ERER BEDE) sisas E85 88258 gls Ends

z Estes eses “555 ESER EEEE] | RF IES
1| 100(6)X101(1)|147| 58| 65|133| 280:123| 30.5 | 3.8 | 1.49 | 2.55 | 49.1| 47.4
2| 100(1)X101(3)104| 24| 32| 67| 171: 56| 24.7 | 2.0 |1.98| 1.01 | 43.6| 40.1
3| 101(9)X100(9) 102| 60| 43|137| 239:103 30.1 | 3.4 | 1.61 | 2.11 | 52.6| 57.6
4 10005) X 101 (4) 170 49 39 06: 3| 4.4|1.50| 2.93 | 44.4] 47.0
5 101(3) X 100(8) 124 53| 601153| 277:113/29.0| 2.31 1.51) 1.52 | 54.6| 52.8
6| 101(5)X100(6) 71| 24| 9| 31| 102: 33 24.4| 2.3/2.57| .89| 29.6] 40.7
7 igo POCO 10 42| 631124| 264:105| 28.5 | 1.8|1.55| 1.16|50.7| 45.0
8| 101 (2) Xx 99(1) 46/111/103| 32| 214: 78 26.71 0|1.75| .00|46.2 49.0
9| 99 (19% 101 (0)! 69229195 60! 424:129| 23.3 | 3.4|1.27 2.68 | 46.1] 52.3
Total | 2,277:828| 26.7 0. 47.7| 48.4
6 (6)X 9(6) 11811141128 131| 249:242 49.3 | 22.6 1.35 116.74 | 52.7| 49.9

Owing, perhaps, to the difficulty in separating waxy from
corneous grains the percentage of waxy grains for the total of
the 9 ears is only 47.7 + .6. The deviation from the expected 50
per cent. is nearly four times the probable error, indicating a
poor fit. The percentage of colorless grains is 48.4 and the de-
viation is two and two thirds times the probable error. Here the
separation is accomplished with a somewhat greater degree of
accuracy. On the whole, the data seem to show conclusively
that we are dealing with a linkage between waxy endosperm
and one of the aleurone factors.

Ear No. 10, which is derived from a non-linkage family and
included in the table for comparison with the first nine ears, is
also the result of a back cross. The per cent. of crossing over is
49.3, which is practically equivalent to independent inheritance.
The deviation from 26.7 per cent. of crossing over is 16.74 times
the probable error and the odds against this being due to ran-

n sampling are enormous.

EVIDENCE THAT THE C Factor FOR ALEURONE COLOR Is
CONCERNED
Mien the linkage data are interesting and valuable, it is
elapse of greater interest and value in the study of maize in-

No. 613] SHORTER ARTICLES, DISCUSSIONS, REVIEWS 59

heritance to know which of the aleurone factors is concerned in
this linkage. This may be determined, where the inhibitory
factor I is not concerned, by crossing plants grown from either
colorless waxy or colorless corneous grains taken from families
showing linkage with plants that are homozygous recessive in
turn for one of the aleurone factors and homozygous dominant
for the remaining factors. Plants of this nature were available
in Professor Emerson’s aleurone testers.

At least five factors are now known to be concerned with de-
velopment of aleurone color. They are known as the A, C, R,
Pr and I factors. A, C and R are necessary for the development
of red. The dominant Pr factor changes red to purple. That
the Pr factor is not concerned in this linkage is evident from
the fact that the linkage relation is observed in segregations of
colored and non-colored aleurone and waxy and corneous endo-
sperm regardless of whether they are purple or red, and from
the fact that only red aleurone seeds were used in obtaining the
back cross data which show the linkage. That the inhibitor is
not involved is inferred from the fact that the segregation on
the original parent ear was 195 corneous colored, 95 corneous
colorless, 95 waxy colored and 15 waxy colorless which is ap-
proximately a 3:1 segregation of colored to colorless aleurone.
The relation of the A, C and R factors to waxy endosperm re-
mains to be accounted for. This may be determined as stated
above by the use of aleurone testers, which are named for the
pair of factors which is homozygous recessive. The following
diagram explains the method of testing for these factors.

| The Constituti 1 Grains from the 3 : 1 Ear may be
Constitution of Aleurone Testers i ; E
aaCCRR | AAccRR | AACOrr
A tester, aaCCRR ......... No color | Color | Color
ARA o Color | No color | Color
R tester, AAC CHRO A Color | Color | No color

If the C factor were linked with the factor for waxy endosperm
no color should appear in the F, from a cross between the C
R and A testers colored ears should be obtained. Crosses of this
nature were made in 1916 and 1917 with the following results:

60 THE AMERICAN NATURALIST [ Vou. LIT

COLORLESS ALEURONE WITH C TESTER

Year | Parentage ie niece | Color of Aleurone
TT | 1 (2) X6857 (9) 360 | No color
1916..... 2 (14) X6857 (12) 240. EE
ie 2 (9) x6857 (4) 500 |
MGR a0 101 (1) X 7506: (2) 250 |

COLORLESS ALEURONE WITH R TESTER
1916....: | 1 (5) X 6867 (19) - 480 Colored
ibero | 2 (1)X6868 (6) 200

COLORLESS ALEURONE WITH A TESTER
1917.....| 101 (7) 7505 (2) | 150 | Colored

The above data are believed to show conclusively that the C
factor for aleurone is linked with the factor for waxy endosperm,
because the & and A testers caused the development of aleurone
eolor when crossed with colorless individuals from the same fam-
ily while the C tester did not.

Non-LINKAGE OF ALEURONE COLOR AND Waxy ENDOSPERM

It is interesting to know that in ear 10, which shows no linkage,
the C factor for aleurone was not heterozygous. A colorless
waxy individual from this family crossed with the C tester pro-
duced an ear consisting of approximately 300 seeds, all of which
showed colored aleurone.

Another waxy colorless individual of the same family crossed
with the R tester produced an ear with 25 seeds, all of which
were colored. Thus the C and R factors are eliminated with
some degree of assurance. The A factor apparently is heterozy-
gous in this family, but unfortunately the writer has obtained
no crosses with the A tester to verify the conclusion reached by
the process of elimination. This may be considered as indirect
proof of the C factor as the aleurone factor which is linked with
the waxy endosperm factor.

SUMMARY
- (a) Collins has presented conclusive evidence of the linkage

between waxy endosperm and aleurone color. The writer has
_ presented additional evidence from back crosses, which shows the

No. 613] SHORTER ARTICLES, DISCUSSIONS, REVIEWS 61

intensity of the linkage in the material at his disposal to be
equivalent to 26.7 per cent. of crossing over.

(b) It has been shown directly, by means of crosses between
colorless individuals in a linkage family and aleurone testers and
indirectly by means of aleurone tests with a non-linkage family
where the A factor and not the C factor is heterozygous, that the
C factor for aleurone is linked with the factor for waxy endo-
sperm.

T. BREGGER
CORNELL UNIVERSITY

INHERITANCE IN ORTHOPTERA

IN a recent paper (Nabours, 717) Nabours has continued his
admirable studies of inheritance in Paratettix. The paper is
backed up with an abundance of data, from which a number of
facts are deduced. In his discussion, he attacks certain modern
hypotheses, and since it appears to me that his strictures are
not entirely justified, I venture here to review the evidence, and
make certain comments on it.

The following facts, several of which were well brought out
in a previous paper (Nabours, 14), are presented :

1. A large number of distinct, true-breeding forms of Para-
tettiz occur ‘‘in nature.” Of these he has collected at least
fourteen or fifteen. He no longer looks upon ee as a distinet
Species, and he has dropped the “specific names”” he suggested
for them in the previous paper.

2. The distinguishing characteristics of these forms fall into
two groups in their mode of inheritance: (a) Fourteen color pat-
‘terns act as allelomorphs to each other. (b) A fifteenth pattern
is ‘‘allelomorphie only to its absence.”’

3. One of the characters of the ‘‘multiple allelomorph”” group
does not always act as an allelomorph to the other members of the
group. This is the character I of his first paper, which was
noted then for the same behavior. Rather, it behaves (to put it
briefly, but in words very different from Nabours’s) as if it
were closely but not completely linked to the others.

Because I wrote a review of Nabours’s first paper on this sub-
ject (Dexter, 14), I feel a certain responsibility for what I think
are mistaken viewpoints concerning the multiple-allelomorph-
nature of this group.

62 THE AMERICAN NATURALIST [Vou. LIT

Bellamy, working in the same laboratory, contributes also an
excellent study on the same subject, but based on a different
genus of Tetrigine (Bellamy, 717). Both he and Nabours have
apparently accepted as proven that there is here a large group
of determiners allelomorphic to each other, and I am quoted in
support of this idea. Inasmuch, however, as I made certain
reservations which are important in the light of the new data pre-
sented, I beg to quote a portion of my former paper.

As Sturtevant has pointed out, for any case to which the idea of
multiple allelomorphism is applicable, an equally valid explanation
may be found in “complete linkage” of the factors concerned. To
decide in any case between the two explanations would be impossible.

If however linkage were not complete, a “cross-over” class
might occur, and this would suffice to rule out the explanation based on
multiple allelomorphs. Such a cross-over class perhaps is furnished

y the BEI individual.

I then suggested that since the BEI individual had been lost
before it could be tested, the cross be repeated, and said:

If then BEI forms should oceur again, and in these, when mated to
other forms, the factors B and I should be found to stay together to
the same extent as before they separated, it would show that close
meee, rather than multiple Waite dear Spim this particular
instanc

Nabours has repeated this experiment, using the character S
instead of E, and has again obtained such a cross-over, BIS.
With this individual, he has carried out breeding tests.

Apparently forgetting what had been pointed out in my paper,
he says:

The significant feature is the complete combination or linkage, ap- -
parently permanent, of the factor for S, and the factor for the modi-
fied I. . . . This combination, IS, becomes a new form, a new multiple
allelomorph (italics not original), pairing with and allelomarphie to
any other multiple allelomorph with which it has been tried . t is
not possible for me to suggest the means by which the dosbinalión or
linkage was effected.

(One must protest against the use of words which permits a
single determiner to be called a ‘‘multiple allelomorph. ’ |

The answer that he was unable to give is obvious. Perhaps
there are thirteen characters here whose determiners are allelo-
morphic to each other. That is possible, perhaps probable,

No. 613] SHORTER ARTICLES, DISCUSSIONS, REVIEWS 63

though unproved. But I is not a member of that group, but is
only linked to it, being, as we may say, in the same chromosome.
The work that Nabours has done makes that certain, and disposes
also, by the way, of the likelihood that non-disjunction explains
the similar case in the first paper (Bridges, ”

Nabours has made a sort of mystery of the character called G
in his first paper, but now called 6, which, he says, is “‘ only allelo-
morphic to its absence.” Ignoring the philosophy of this state-
ment, he has shown that @ mendelizes independently of the other
characters. He suggests that such determiners may be of fre-
quent occurrence. He has shown that by substituting Greek
letters for English letters the formule will work out as well as
they did before, and has naively applied the method to the case
of comb inheritance in poultry. His difficulty is simply caused,
and Bellamy has pointed out its solution:

It need only be assumed that the determiner is borne by some other
chromosome.

In Drosophila some four or five years ago, the determiner for
bent wings was the only one known for the fourth chromosome
group. If at that time we had known only one other set of char-
acters in Drosophila, viz., that of the white-eosin group, the situ-
ation would have been parallel to the one described by Nabours.
We might speak of a half dozen or so of ‘‘characters allelo-
morphic to each other,’’ and of one, bent, ‘‘allelomorphic only
to its absence.’’ Later on, when we found other characters
whose determiners were located in the fourth chromosome, we
should modify our theory. Nabours’s industry in his research
makes me feel safe in prophesying that he will yet discover some-
thing linked to 6.

He says parenthetically that two other characters ‘‘apparently
of the nature of @’’ have been discovered. It is important to
find out their linkage relations and we shall wait eagerly to hear
of them. In the meantime we must conclude that he has discov-
ered the beginning of at least two chromosome groups.

er

PAN
1917. Studies of Inheritance and Evolution in Orthoptera, Iv.
Journal of Genetics, Vol. E ou

64 THE AMERICAN NATURALIST [ Vou. LIT
— C. B.
916. Non-Disjunetion as a Proof of the Chromosome Theory of
redity. Genetics, Vol.

Dexter, J. 8.
1917. Nabours’s Breeding Experiments with Grasshoppers. AMERI-
CAN ‘NATURALIST, Vol. 48, p. 317.
Nabours, R. K.
Journal

1914. gS of Inheritance pii Evolution in Orthoptera.
enetics, Vol. 3,
1917, Studies o of Inheritance u Evolution i in Orthoptera, II and III.
of Genetics, pp. 1-5
JOHN S. DEXTER

UNIVERSITY OF SASKATCHEWAN

THE
AMERICAN NATURALIST

Voz. LIT. February-March, 1918 No. 614

INTERNAL FACTORS INFLUENCING EGG PRO-
DUCTION IN THE RHODE ISLAND RED
BREED OF DOMESTIC FOWL

A SURVEY OF THE PROBLEM or Eee PRODUCTION AND A
PRELIMINARY ANALYSIS OF AN EGG RECORD INTO ITs
CONSTITUENT ELEMENTS

DR. H. D. GOODALE

MASSACHUSETTS AGRICULTURAL EXPERIMENT STATION, AMHERST, Mass.

INTRODUCTION

A survey of the problem of egg production, such as is
made in the present paper, seems desirable at the present
time because of the great interest taken in breeding for
increased egg production. While the various factors dis-
cussed are familiar, to a degree at least, to most poultry
keepers, nevertheless they are ignored in breeding prac-
tise and reliance placed upon the numerical record alone
as a sufficiently detailed and accurate description of a
hen’s performance, although, as will be pointed out in a
later section, identical numerical records result from
quite diverse combinations of factors.

The point of view which we have been led to adopt may
be stated in one form as follows: The egg record of a hen,
expressed as a given number of eggs per unit of time and
taken by itself, is not a sufficient measure or description
of egg production, even under a favorable environment,
for the record is the result of the interaction of a number
of innate factors. Some of these factors, such as rate of
growth, are quite distinct from egg production, while

66 THE AMERICAN NATURALIST [VoL. LII

others, such as rhythm, are almost inseparable from egg
production itself. The numerical record of a hen shows
only the number of eggs laid, but does not show the com-
ponent elements which enter into the making of such a
record. All these various elements must be studied in-
dividually and the influence exerted by each on egg pro-
duction worked out. Moreover, the mode of inheritance
of the separate factors must also be determined.

Further, it should be noted that the interrelation of
the various factors is so complex that it is difficult to
describe each by itself. In nearly all cases the bearing
of some other factors must be considered to a certain ex-
tent, at least, along with that factor which is specifically
under discussion.

It is important to observe that while the results ob-
tained for the Rhode Island Reds described in this paper
differ in several respects from those obtained by Pearl
(712) for Barred Plymouth Rocks, these differences are
inherent in the birds themselves and are on a par with
the visible differences, such as color, that exist between
the two breeds. Pearl has anticipated that differences in
fecundity in various strains and breeds are likely to be
found. He states as follows:

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 races or breeds of poultry than those used in the present ex-
periments may show a somewhat different scheme of inheritance of
fecundity. . . . I wish only to emphasize that nothing is further from
my desire or intention than to assert before such investigations have
been made that the results of the present study apply unmodified to all
races of domestic poultry.

It is clear, then, that a complete knowledge of fecundity
and its inheritance in domestic birds can only be ob-
tained by a careful study of egg production in all breeds
and perhaps even in several strains of the same breed.
As shown later on, one of the several factors that deter-
mine winter egg production is characteristic of Pearl’s

No. 614] EGG PRODUCTION 67

Barred Plymouth Rocks, while another is characteristic
of my Rhode Island Reds.

The data in this paper are obtained from a flock of
220 March and April hatched pullets placed in the laying
houses in the fall of 1913, together with the data on winter
egg production from the flock (numbering 482 pullets)
placed in the laying houses in the fall of 1915, although
the composition of this flock was not the same as that of
1913-14, because it had been altered by the addition of
several other strains in order to overcome the unsatis-
factory vitality of the original flock. The addition of
new blood apparently increased the variability in some
respects as shown by the statistical constants (cf. Figs.
1 and 2, also 10 and 10a). The winter production of the
flock of 1914-15 was decidedly poor and apparently not
normal, probably due largely to environmental condi-
tions, and hence data from this flock have not been used.

It is impossible within the limits of this paper to pre-
sent detailed data on all points discussed. To the reader
who is unfamiliar with egg records, it may be said that
an inspection of the records reveals the essential nature
of the problems.

The original flock came mainly from one of the leading
showroom strains of the country, to which were added a
few individuals from another showroom strain. Neither
strain, so far as known, had been especially bred for egg
production, nor had any of the strains added in 1915.

Ways or Measurin6G Eee PRODUCTION

It has been customary in times past to determine a
hen’s egg production by her record expressed in the num-
ber of eggs per year, the year usually running from No-
vember 1 through the succeeding October 31. At other
times the first-year record of the hen has been taken as
the time unit, beginning with her first egg and running
365 days therefrom. More recently, the Maine Exper-
iment Station has used the period beginning with the first
egg of a pullet and extending to March 1 as the unit of

68 THE AMERICAN NATURALIST [Vou. LII

measurement, since March 1 serves as a convenient cal-
endar date, near the end of the winter cycle. Still more
recently the same workers have suggested that even a
shorter period would be desirable, because it is held that
a hen only reaches her highest possibilities under favor-
able conditions. Recently the Utah Station (Ball, Tur-
pin and Alder, 714) has suggested that for Leghorns the
records be kept for three years, since hens that lay poorly
the first year often lay much better during the second
or third. Rice, however, (’13) has published data on
this point, which show that such birds are the exception
rather than the rule.

A year, however, may be considered to be a natural
unit. During this period the whole eycle of seasonal
changes is gone through with. Moreover, this period
bears a definite relation to the bird’s life cycle, for its
beginning may be taken to correspond to the beginning
of egg production in the fall, while its close roughly corre-
sponds to the cessation of egg production the next fall,
usually coinciding with the onset of the fall moult,
though, of course, in some individuals, the biological year
exceeds 365 days. Thus, the year would seem to mark a
pretty definite period in the life of the bird as to her in-
nate capacity for egg production. In this paper we have
used both winter and annual periods as measures of pro-
duction, as the necessities of the moment required.

There are some objections to each of the two common
methods of determining the point at which the year
begins. If the year begins with the first egg of each indi-
vidual, the differences in age at which the first egg is
produced are neglected. If a given point in the year is
chosen and the production of all individuals within a year
from this date recorded, differences in time of hatching
are neglected. Possibly a more satisfactory method
would be to take 365 days from the beginning of egg pro-
duction in each flock of equal age, or else from the aver-
age date at which production begins.

The terms ‘‘high producer”” and ‘‘low producer”” are

No. 614] EGG PRODUCTION 69

frequently encountered, but each is used very loosely.
The use, either of the term ‘‘higher producer’? or ‘‘low pro-
ducer’ without qualifications of any sort can scarcely be
iently precise. Unless qualified by the word
annual, the term ‘‘ high producer”’ in this paper will be un-
derstood to refer to the winter record only. Pearl (712)
has defined a high producer as a bird that lays over 30
eggs during the winter, a mediocre producer as one that
lays during the winter but that lays fewer than thirty
eggs, while a zero producer does not lay at all during the
winter. As will appear later, the use of the numerical
value of the record as its sole characteristic is insuffi-
ciently precise. The term ‘‘true mediocre producer”’ will
be used to denote a mediocre producer in the sense
(Pearl’s) explained below, while the term ‘‘mediocre
(under 30 eggs) producer”” will be used elsewhere.

The Influence of External Factors.—A brief consid-
eration of the relation of external factors to egg pro-
duction is necessary before considering internal factors.

External factors may be divided into two classes: first
those that operate rather directly upon egg production,
and secondly those that operate indirectly, through their
- influence on the organism as a whole.

Under the head of direct factors should be mentioned
housing, climate, food, general care, ete. It should go
without saying that the birds must be properly fed and
kept under conditions generally recognized as suitable
for maximum egg production. It is not yet clear, how-
ever, that the optimum conditions are fully known, or
that they can be obtained at will, for with the present ap-
pliances for keeping poultry, only the crudest sort of
approximation can be made toward securing a uniform |
environment. For example, one is never certain with
open-front houses that a draft may not strike one portion
of the flock, while on the roosts, but not another. There
are many little things of this sort which can not at pres-
ent be controlled, nor is it definitely known in what way
these ““little things” influence egg production. Some

70 THE AMERICAN NATURALIST [Vor. LIT

appear to be without any influence whatsoever; others
appear to. be of varying degrees of importance.

Thus, it is not easily possible to overemphasize the im-
portance of the environment in relation to egg produc-
tion. At best, certain elements of the environment are
partially controlled and similar conditions supplied to
the members of the flock under experimentation, but it is
impossible with the best practical facilities at present
available to furnish identical conditions to all individuals
of the same flock. At the very best one can only go
through the motions of providing such conditions. More-
over, one may be forced to modify the procedure selected
in order to keep the birds in good condition. Further-
more, individuals or strains may not react in the same
fashion to identical conditions.

The difference in the reaction of individuals of the
same strain to similar conditions, particularly when these
conditions fall near the eritical point for the strain (or
species), is a matter of considerable importance, espe-
cially when a character such as egg production is under
study, and more especially when it is impossible to con-
trol certain important elements of the environment. As
long as the environment is not too far from the optimum,
birds of low vitality, for example, may do quite as well as
birds of much higher vitality, but when the environment
approaches either end of its range, then its effects begin
to manifest themselves.

A full discussion of the possible influence of the en-
vironment, either directly or indirectly, upon egg pro-
duction as a whole or upon any of the several factors in-
fluencing production is outside the scope of this paper.
While the reader should bear in mind the possibility that
the environment has introduced disturbing factors, every
effort has been made to keep all controlable elements,
such as feeding and housing uniform.

Turning now to internal factors, we find that these also
may be considered under two heads. We have little to
do with the factors falling under one of these heads, for

No. 614] EGG PRODUCTION 71

their effect is exerted only indirectly. They undoubtedly
play an important part in egg production, but like many
external factors they are without influence unless they
fail in some way. Such factors are the capacity to digest
and assimilate food, to excrete waste matters properly,
etc. It is not my purpose at this time to discuss such
factors. Those internal factors with which we are mostly
concerned are those whose relations to egg production are
much more obvious. They are rate of growth of the
chick, cessation of growth, the attainment of both bodily
and sexual maturity, moults, the size of the bird, the
stamina of the bird, the presence or absence of cycles,
litters or clutches of production, the rhythm of produc-
tion, the rate of production for definite time intervals,
age at first egg, and broodiness. Some of them are
clearly separable from egg production. Others are so
closely interwoven that it is impossible to say that they
are not phases of egg production. Whether or not this
is so, is of no immediate importance from the standpoint
of inheritance, since the result will probably be the same
whether they are treated as genetic factors that are sep-
arable from egg production or treated as groups into
which egg production itself may be divided. These fac-
tors may be regarded as phases of egg production if one
desires, but on the whole it has seemed profitable to re-
gard them as factors influencing egg production.

Rate of Growth, Bodily Maturity, Cessation of Growth,
Sexual Maturity’

- These interrelated factors are closely interwoven in
their effect on egg production. Under normal conditions
it is clear that sexual maturity is indicated by the begin-
ning of egg laying, and may be measured by a bird’s age
at her first egg, i. e., the length of time elapsing between
the date hatched and date of first egg. Sexual maturity,
however, demands certain antecedent conditions before it
can become manifest. Among other conditions is a cer-

1 Unless otherwise stated, reference here is to the female only.

72 THE AMERICAN NATURALIST [VoL. LII

tain body size, which depends upon the rate at which the
individual grows, as well as the limiting size for that in-
dividual. That is, size at a given age is the result of rate
multiplied by time, up to certain limiting values deter-
mined by the genetic composition. Cessation of growth,
however, does not necessarily coincide with the onset of
sexual maturity nor even with general bodily maturity.
Although it is certain that the hen is heavier in her second
autumn than at the beginning of egg production, our data
show that there is little or no growth during the first
winter. We must, then, distinguish between sexual ma-
turity, which is capable of manifesting itself as soon as
the body reaches a certain size, from that maturity which
is not attained until long after the adult size is reached.
At present the relation between sexual maturity and bod-
ily maturity has not been worked out. Some extreme
phases, however, of the interrelation appear a priori
probable. Chicks that grow very rapidly naturally tend
to reach sexual maturity at a very early period in their
life. They may or may not start in laying immediately
after reaching full size. Other birds grow very slowly
and can not lay before a certain size is reached. There-
fore, they must of necessity reach sexual maturity rela-
tively late in life. It may be impossible for birds of this
sort to reach sexual maturity before spring if hatched
during the usual breeding season (April, May). The
general effect of slow growth, then, will be to lower the
record made by such individuals, although they may be
otherwise identical with those that grow more rapidly.

Combined with the factors mentioned are the factors
that limit the size finally reached. As pointed out above,
size results from rate of growth times length of period
through which growth continues. Each factor is deter-
mined in part by the environment and in part by the
genetic constitution of the bird.

The following combinations of dl (Table 1) and
their effect on egg production may be assumed. Each
factor is treated as though it were wholly independent

No. 614] EGG PRODUCTION Bo

of the others. Early sexual maturity is assumed to be a
constitutional tendency to begin laying as soon as a suffi-
cient body size or body maturity is reached, while late
sexual maturity is assumed to be a tendency to delay pro-
duction until after body maturity is attained. This head-
ing, however, does not refer to the objective attainment of
sexual maturity which is shown by the column on ‘‘time
of first egg.” The length of the growth period also is
assumed to be determined by the attainment of bodily
maturity.
TABLE I

VARIOUS COMBINATIONS OF HYPOTHETICAL GROWTH FACTORS WITH THEIR
EFFECT ON WINTER Eee PRODUCTION

| Probable Time of First | hy sone Winter

Rate of Growth | Sexual Maturity Growth Period Egg £ Production
| | |
Rapid | Early Short ae High
sr Late Short Low
Early Long Relatively late Medium
fe Late Long Late Ow
Slow | Early Short Relatively late Medium
r Late Short Relatively late Low
Early ong Late Low
Late | Long Very late Zero

It appears from this table that early sexual maturity
can become fully effective only when combined with rapid
growth during a short growth period.

The effect of the activities of some of these factors as
bearing on winter egg production may be given more
specifically as follows: If we measure egg production by
the number of eggs laid before the 1st of March, assum-
ing for the moment that this point represents, approxi-
mately at least, a definite point in the history of the egg
production of each individual, it follows that the birds
hatched during April and May, or to take a definite point
for the purposes of illustration—April 15—which mature
at five months, as is sometimes the case, will begin to lay
September 15 and will lay a large number of eggs before
March 1, provided, of course, that they do not moult. On -
the other hand, true mediocre productivity (slow rate)
associated with early maturity will tend to force a bird

74 THE AMERICAN NATURALIST [Von. LII

out of the class of mediocre producers, when measured by
a specific number of eggs, into that of the high class. If,
then, one is dealing with a flock in which these degrees of
maturity exist, it is evident that extreme care must be
taken to avoid confusion due to differences in maturity
or rate of growth.

Differences in maturity may be observed among the
males as well as the females, although there is no precise
objective point at which a male may be said to have be-
come mature, which is comparable to the first egg of a
pullet. On the whole, the larger birds tend to mature
later than the smaller, though the rule is by no means
rigid, since some small birds grow slowly while some
large birds grow quickly. Since age at first egg is so
large a factor in determining the kind of record a bird
makes, one has a physiological character in the male of
considerable value as an index of his capacity for produc-
ing females that will mature at a given age.

The age of a bird when she produces her first egg does
not coincide necessarily with bodily maturity, theoreti-
cally at least, although it seems that a certain size must
be reached before the bird can begin to lay. On the other
hand, the relation between body size and age at first egg
as frequently encountered is of a sort such that the larger
birds tend to lay at a later absolute age than the smaller
ones hatched the same day. There are many exceptions,
however, to this rule. It would, perhaps, be better ex-
pressed to say that more heavy birds lay late in life than
early, while more of the lighter birds lay early than late.
For one of the flocks, the coefficient of correlation between
age at first egg and weight has been calculated and found
to have a value of + .5473 + .0216.

The influence of the date at which the first egg is produced
as well as the relation of age at first egg to the number of
eggs laid during the winter months is shown in the series
- of records shown in Figs. 3 and 4 (Page 78). These
records have been selected in such a way that the rate of
- production is nearly constant, although the date of hatch-

No. 614] . EGG PRODUCTION 75

ing of the individual birds covers a period of five weeks.
The records are to be read as follows: The number in the
upper left-hand corner is the hen’s number. The ver-
tical mark in each square indicates that an egg was pro-
duced on that day. The totals are given for each month
while the figure at the extreme right of the row headed

PERCENT

111-380 1811% 198200

AGE AT FIRST EGC
Fig. 18 Graph showing the percentage of the flock beginning to lay at the
ai 5 days) ao ad by the ay limits. Flock hatched in March, ume and
1915. M = 263.19, S. D. = 37.71; C. V.. = 43.00, C. V. =

“February”” is the total number of eggs for the winter
period. The records for March and April have also been
included in order to show the type of record made by
birds that begin to lay very late in the season. These
records show clearly that no sharp dividing line exists in
the number of eggs laid. On the contrary, it is clear that
birds hatched at the same time begin to lay at widely dif-
ferent dates and that in consequence differences in egg
yield for the winter period result. That this result is of
general applicability to our flocks is shown by the fair

sIn calculating the C. V- for the data given in Figs 1 and 2, the
mean was taken as the difference between the mean age and the lower end
of the range of Fig. 1.

76 THE AMERICAN NATURALIST . [VoL. LII

amount of homogeneity in the flocks in respect to rate of
production as described below.

Graphs showing the age at first egg for the flocks of
1913-14 and 1915-16 are shown in Figs. 1 and 2. The
former (Fig. 2) is unimodal and has a narrow base, the
shape of the curve indicating a high degree of homo-

PERCENT

Sos BOM AG AON & eS

“hoe AT FIRST ked
Fic. 28 Graph showing the percentage of the flock beginning to lay at the

Fy (in a rere by the class limits. Flock hatched in March and April,
M= Ag, S. D. = 24,34, C. V.s = 28.41, C. V., = 9.32

geneity in the flock. As might be expected from the na-
ture of the data (which is affected by the environment in

only one direction, i. e., toward a retardation of the age
at first egg) the lower part of the right-hand side slopes

No. 614] EGG PRODUCTION qi

off more gradually than the left. The mean has a value
of 261.18 days.?

The curve for the flock of 1915-16 (Fig. 1) is somewhat
unlike the preceding. It is distinctly bimodal, but it is
not altogether clear that this bimodality indicates two
genotypes, for it may be due to chance alone. The base
is broader than for the 1913-14 curve, indicating less
homogeneity of the flock in this respect, although the
same gradual slope on the right-hand side is apparent.
There are reasons, however, for believing that the left-
hand side of the graph for the flock of 1913-14 was short-
ened by the methods of handling the pullets that fall.
The mean has a value of 263.18 days. The difference be-
tween this graph and the first is undoubtedly due to the
changes in the composition of the flock as described in an
earlier paragraph.

Graphic representations of the day on which the vari-
ous members of the flock produced their first egg are
shown in Figs. 5 and 6. The data for the two flocks, i. e.,
1913-14 and 1915-16, are divided into groups according
to the month in which the pullets were hatched. Each
dot in the figure represents the first egg of a pullet
and is placed in a square corresponding to the date on
which the egg was laid. If more than one pullet began
to lay on a given date, there is a dot for each pullet.

There are some interesting differences and resem-
blances between the groups mentioned in the distribution
of the first egg through the various months. In all in-
stances the pullets laying for the first time come in slowly
during the first few weeks. Then follows a period of six
to eight weeks during which the new pullets come in at
a faster and fairly uniform rate. This period is followed
by a third period when new pullets come in slowly, the
last of the period representing the stragglers. The fairly
uniform scatter is due in part to the inclusion of several

2 In this paper we have given only those statistical constants that appear
to be particularly pertinent and as a rule have omitted the probable error,
especially where ‘‘n’’ is large, unless there has been special reason for
inserting it.

i

= No. 4846 er Harcnep Marca 21, 1915. Aam ar First Eco, 191 Days

Date Estela
1915-16 1 213|4|5|6/|7|8/|9|10/11/12/13|14|15/16/17/18/19/|20| 21) 22 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 Totals
Sept......| kid E-
Oa. PF Pe eee ee) eee ei ls abi IPED GOA EE AI
Nov 41 Meet AA IEA lib APA ELIT ae AN EA LE
Dec N| / cae PGs SS aa ot eee ees oe ee ae
Ca 1 rect il Rl fl hi eel eli) lee PEE AAAA PAI
eG RR Phere E ELE LLE EFE A 20
| AY | ul A
eh ee Fle reel iit ie) AAA Te lees
ape Ll Rede Plt Ae et bec EEN N Ny naa
No. 4921 oe Hatcuep Marca 21, 1915. Aam at First Ece, 219 Days
Revie * 112|38|41|5/6/|7|/8|9]/10/11/12/13/|14/15/16/17/|18|19|20/21/|22| 23 24 | 25| 26| 27| 28 | 29 | 30 | 31 | Totals
Sept Po...» : E : | a 0
Oct... EET El oh 3
Nov / l / ¡A / / / / IE | inf icf |E 7
De EA eno ese, ERr EISA AAAF
D Pilg a a APE eal Pie ER EAA ARE las
E > 1
Md IFI BPA 1a PRS AN EEE LARA IBAS 18
3 Deo Ge
Mee iis | Ae Fife li ial tr Pre Peery eee] EE
ore la tee lis lll lil leleletels ty llas]

Fras. 3 AND 4. Daily records of Rhode Island Red pullets hatched in 1915, arranged in order of decreasing total winter eggs to show
the effect of approximately equal rates of production, but of different dates of first egg on the total winter production, as shown by the num-
eral at the right of the February record. March and April records are included to show that no fixed date can be selected as a dividing
line, l=an egg; N=on nest but did not lay; B, L.=removed to broody coop; A =returned to pen

8L

LSITVAALVN NVOIJAWV AHL

TIT "0A |

Firas. 3 AND 4.—Continued,

No. 4514 Hatcuep Marcu 7, 1915. AGE AT First Eca, 266 Days
Date 3 : ;
1915-16 T ón 3 617 |81}9 110/11}12/18 | 14| 15| 16/17) 18) 19) 20) 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28| 29 | 30| 31| Totals
Nova.. 0. N N|/ IB 2
Dh. ay eS NE LEE LEE eae Fee Wee See dled
ERY Ll ill Ili ar ab page. tit ¡Y I | 21
Pe el / i l {li iif Lli Iil AWAY LLEI 20
| i.
Mee MU ATI ANN ¿MONTA nl i} i} i] las
Apr.. lila AGE | E Bot A Y INTE
No. 5297 Harcuep APRIL 11, 1915. Acre ar First Eca, 268 Days
bet 1 343 6/1718 )9 110/11) 12)13 | 14/15) 16) 17) 181/19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30! 31 | Totals
Nov.) c.: E o
Don. oa 0
Jan: A oi A T pa aa TAS AA a ea
Pop.: fly oo eevee Fa i-feel Wey ee) AA NA || [ls
41
BL
Mar. ry alar ear eel [eo BH ao
NOS a| |N a kes eo ene?) a Sh A UP

[PT9 “ON

NOILIAdGOUd DOA

62

HATCHED Marcu 14, 1915. Aam at First Eco, 312 Days
3/4 8/|9¡10/11/12/13/14/15/|16/|17/18/19|20/21|22|23/|24/|25/|26|27| 28} 29 | 30| 31 Totals —
B o
0
/ rd EEIN E Hr FELY rae eee
J4 l a ANS H 23
ORE? |:
AN O CRA ANNAN MT EA 188
© il) Peer tee al A al lla ll los
Hatcuep APRIL 4, 1915. Aam ar Frrst Eca, 294 Days
I 3) 4 8 | 9 |10}11)}12)13)] 14/15/16) 17| 18! 19 |20 |21 |22 23) 24 25 | 26| 27| 28} 29 | 30} 31 | Totals
| mu
| 0
/ LALA lal 7
eb. LLE JO ANS NINA NE eee we! ON 22
| | mm”.
Moz. us|} | Ree AE AAA ee || al
PE AININ / / AAA Ad Ed IN 20

l
Fias, 3 AND 4,—Continued.

08

LSTIIVIALVN NVOIJIHMV AHL

ITI “10A]

HATCHED. APRIL 4, 1915. Acre at First Eca, 304 Days

pl

12/13) 14) 15) 16/17) 18} 19

20

Pert yt LAS

/

Peery te eae eT

/

PEN PE]

/

14, 1915. Aem at FirsT Eca, 336 Days

12/13/14 /15/16/17/|18 19

20

|

"IO y

¿CENAS SS,

FIGS. 3 AND 4.—Continued,

[F19 “ON

NOILOMaOYd 99A

18

No. 5192 HATCHED APRIL 4, 1915. Aam ar First Eca, 334 Days
=k te 1 9 | 10/11) 12] 13| 14| 15| 16/17] 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 26 | 27 | 28 | 29 | 30 | 31 | Totals
Hore E Bo
A 0
ae BE
a | P
Mar...... see <A A I ee
Ayee rr A TT Arr
No. 5201 HATCHED APRIL 4, 1915. Agr ar First Eco, 352 Days
a 1 9 110! 11/12] 13/14] 15! 16/17) 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 27 | 28 | 29 | 30 | 31 | Totals
Nov.. n E o
m. 0
Jan.. 0
w... y y 0
0
wW Pe TTT iy art) ths
Apr.....-| / cla YE A A AAA JE

Fics. 3 AND 4.—Concluded.

38

ISTIVINOIVN NVOIVAMNV AHL

117 “10 A]

HatcHep IN March, 1915

Date |1|2|3|4|5|6|7|8 9 |10 |11 |12 |13 | 14 | 15 | 16 | 17 [18 | 19 | 20 | 21 | 22 | 23 | 24 25 26 | 27 | 28 | 29 | 30 | 31 | oe

[FI9 “ON

Nov. . ... Lt” pide .. . ... . .. .. . ... .. . ae .. . .. .. . . . . ... . 27.2
o ... |. . . ... prs . oe ... lo . eee E E $ .. oe |. . . .. . .. . .. .. . 24.4

Jan. essees E E =? add . .. . . . . . . . eee les 12.2

HATCHED IN Aprin, 1915

Date [| 1/2)3) 4) 516) 7) 8 | 9 | 10/11) 12/13/14) 15) 16 |17 | 18 | 19 20 | 21 | 22 | 23 | 24 | 25 | 26 |27 |28 |29 | 30 | 31 | ŞE

NOILINGOUd DOA

Nov. * . . . . . . .. .. oe . . oe ee see 14.5
Dea 3 . . ... |. .. ... +. .. . . ts} Le . E id fa ... . . . eee eee lee eee ae E 34.6
Jan.. i .. .. eee | eee pied . . oe eee .. . : . je .. . . . .. . .. . . ... 25.7

eee eee

Man... eee oe. be ne Y ha $ y . . .. . 10.1

€8

AM y . a

Harcuep in May, 1915

4/15/6|7/8 9 |10/11|12/13 14 15 16 | 17 | 18 | 19 | 20 21 22 23 | 24 | 25 | 26 | 27 28 2930/31 | o,

v8

* . . . . . . ee ee . .. eee 12.5

eee .. ...j. . oe ee oe ... e... ¡.. oe 114.109 . . s.. | ee . . ...|» eee
: ? : 37.5

oe ooo je . . . .. . eee lee . se ae ee . ... . . 32.4
.. ... . . . . . . . . . . 15.4
iwa is En : 2.2

ISITVUAIVN NVOIJANV AHL

Fig. 5.—Concluded. Diagram to show the date on which the first egg of each member of the flock was laid. Flock of 1915-16,

TIT 104]

No. 614] EGG PRODUCTION 85

hatches on one chart, “na also to the ungrouped data, for
if the data be grouped in 10-day periods, a curve is ob-

40

"PERCENTAGE
e
o

s

Fic. 5a. Graph mia the percentage of the flock of etme that began
to lay in the month indica From Fig. 5. ———— March, — — April,
and ----- May hatched pcia respectively.

100

90
80
70
60
$
E
5”
x
ka
a
40
JO
20
10
ea A AA
Fig. 5b. Graph showing the percentage of the flock of 1915-16 sod waa desa
to bade DO or an ber naga indicated. From Fig 5 reh,
il y hatched pullets, respectively:

tained similar to the one that results from the combina-
tion of the age at first egg curves of several consecutive
hatches. |

Harcnep IN Marcu, 1913

- Date $ 2 3ļ|41516/7|8]|9/10|11|12|13 |14 |15 |16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 24 | 25 | 26 | 27 | 28 | 29 | 30

98

Nov. oe a . . .. .
: Deo... O hel oe a e e me ne 0. ee fees repel © 7 . ar lre fs - | 58.5
de “h ' . i : : . . 15.4
Go o , : : 6.2
a

HATCHED IN APRIL, 1913

Per

ISIITVAALVN NVOIMANV AHL

Date [112131 4(5| 6] 7] 8 | 9 [10/11 | 12] 13/14) 18 | 16 | 17 | 18) 19 | 20 |21 | 22 |23| 24 | 25 20 27 28 20 |30 81 | Gent
cn ee : |
Oct... essre A
Nae.. a . . . .. . . a is . roo jo 16.5
Dec...... S ee . . . .. . . . . eS .. .. .. oe ee . eco [too fo .. ee |oo ed .. . 38.6
Jan. 4 is eee |. se ee .. eee [eee lee . . .. ee . . . . ee . . .. . 35.4 a
— E
«Feb... 4 . . . .. . . . . . . | 8.7 -
ha E
Mar... . : | | 0.8 ma
A... | | |

Fic. 6.

Harcuep 1n May, 1913

Date a 4(5/6|7/8|9|10/11/12/13/14/15|16/17|18|19|20/21|22|23|24|25|26 | 27 | 28 | 29 | 30 | 31 Tae
Sept......
Oita..
NOF ira
Doaa . . . . . . 25.0
Jan ee E ‘ ' : : y ; * | 50.0
Yee i : y : Es 25.0

AD oe

Fia. 6.—0oncluded.

Diagram to show the date on which the first egg of each member of the flock was laid. Flock of 1913-14.

Hatcuep Marcu 22, 1915

Date | 1 4|5/6| 7 | 8] 9 |10)11|12| 13/14 15/16/17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 can
Sept... 4
Oct. . .. . . . . . 25.0
NOV ao 1 . . 11.1
Pes ae a > : è he aa e : vs 36.1
A ob : 19.4
Pob 1 : 5.6
Mar........ 0.0
Aok ae. í F

Fie. T. isa le show the date on which the first egg of each member of the flock was laid. This figure is based on a new sample of the
text.

original strain that made their winter record in 1915-16.

See t

NOILOAGOUd DOA [FI9 ‘ON

18

88 THE AMERICAN NATURALIST (Vor: LIT

In any one hatching group a period of several months
elapses between the date the first pullet begins to lay and
the date the last member of the flock starts. This period
is longest for the March-hatched birds, apparently be-
cause the warm spring weather brings all the stragglers
to laying and because the March-hatched birds are the
first to lay in the fall. For the May-hatched birds the
period between first and last pullet is shorter because they
begin to lay later in the fall than the March-hatched birds.

For the March-hatched pullets of 1915-16, the initial
period is nearly twice as long as for the April or May
pullets. The date of the first egg of the first pullet is ap-
proximately a month later for the May than for the April
pullets. The data, however, for 1913-14 are not quite
comparable with those for 1915-16. In the first place it
was impossible in 1913, because of lack of room, to begin
putting the pullets into the laying quarters until late in
October, while some were not finally in place until about
the middle of November. The birds therefore did not get
settled down at once. The March- and April-hatched
pullets both began to lay at approximately the same time
and although most (77 per cent.) of the March birds had
commenced laying by January, a considerable percentage
(viz., 44.9 per cent.) of the April pullets did not begin to
lay until after. Jaunary 1, which is approximately the
same percentage (viz., 49.7 per cent.) obtained for the
April pullets of the 1915-16 flock. It should be noted,
too, that 73.9 per cent. of the March pullets of this year
began to lay before January 1, so that the effect of the
delay in housing the 1913-14 flock shows itself principally
in a retardation of the first eggs of the March pullets,
forcing a larger percentage of first eggs into December
than would be normal for that flock. There is a further
difference in the two years. The percentage of the April
hatched pullets laying after February 1 was about 24
times as great for the 1915-16 flock as for the 1913-14
flock, the ratio being 24 per cent. for the former to 9.5 per
cent. for the latter.

No. 614] ' EGG PRODUCTION 89

There is another way in which the close relation be-
tween age at first egg (also date of first egg) and the
winter egg record can be shown, for it follows that the
higher the age at first egg the lower the winter record

TABLE II
AVERAGE WINTER EGG PRODUCTION FOR EACH WINTER MONTH OF 1915-16,
BY HATCHES

D Hac pas Se | Ne | D |

e gos, [Bema E gon | ee | tr | ae Fe,
February. 7...| 12 |66.83|.5.9 | 14.3 | 11.9 5.2 | 2.6 | 10.8 | 16.2
February 14... 8 | 56.63 | 5.3 9.9 8.8 6.6 5.5 9.4 | 11.3
February 21...| 13 !65.92| 1.8 | 10.5 | 10.9 | 10.1 8.8 | 10.8 | 13.1
February 28...| 18 | 41.44] 0.0 2.0 3.7 6.3 8.8 9.4 111.3
March Vin... 24 |44.88| 0.0 1.5 5.2 s2 | 206: j 10,6 9.9

March 14..... 33 4.15 | 0.0 0.0 0.8 5.1 7.7 8.3 | 12.2.
Mareh 21 23... 48 |40.67 | 0.0 0.3 2.7 7.2 | 11.6 9 8.9
March 28..... 17 .00 | 0.0 0.0 0.0 44 1118 | 1254 113
io 29 9.31 | 0.0 0.0 0.1 1.0 8.0 9.4 :| 10.7

April 11 sais. 63 | 31.30) 0.0 0.0 0.0 1.9 6.8 -| 10.91 87
April 183.4705) 47 | 29:26 | :0.0 0.0 0.0 0.6 5.9 | 10.4 | 12.4
A O os 34 |27.35| 0.0 0.0 0.0 0.4 3.3 | 10.6 | 13.1
May 2 30 2 0.0 0.0 0.0 0.0 1.3 6.9 | 12.0
Moy Pesii 30 | 20.37 | 0.0 0.0 0.0 0 1.8 701 1126
Mar i0., v5 19.) 18.95 |. 0.0 0.0 0.0 0.0 0.1 4.3 | 14.6
May ZA 3 17.78 | 0.0 0.0 0.0 0.0 0.2 4. 12:9
May 30 20.111.354: 0.0 0.0 0.0 0.0 0.3 3.1 8.0

should be. In calculating this coefficient of correlation

it is necessary that the birds should all be hatched at the

same time, so that for our flocks, which were hatched at

intervals of one week, it would be necessary to form as

many correlation tables as there were hatches. The prob-

able results did not seem to warrant the labor involved, at
: TABLE III |

AVERAGE WINTER EGG PRODUCTION FOR EACH MONTH OF 1915-16, GROUPED
By MoNTH HATCHED

|Average
‘ Sep- No- De- Feb-
Month no, Pm ABE | er October ver | cr Jomar hay
tion
February..... 51 56.2 27 8.3 8.3 dh 6 0i | 129
ae T REE 122 39.8 0.0 0.4 2.3 6.2 10.4 | 10.0 | 10.3
ADA. acs 173 | 29.8 ¡ 0.0 0.0 0.02 LE i 0.5 | 12.0
18.1 0.0 | 0.0 0.0 0.0 0.8 5.4 | 11.9

90

THE AMERICAN NATURALIST

[Vou. LIT

least not at present, so that the coefficient was calculated
for one of the largest hatches only. The value found, r=
— 829 + .029, is in full agreement with the hypothesis
that the winter egg production of a flock all hatched at
the same time, depends largely upon the age at which the

first egg is produced.

TABLE IV
AVERAGE WINTER Ege PRODUCTION FOR EACH WINTER MONTH OF 1913-14,
BY HATCHES

lay verage
N er! | |
sar mA Pullets bei August anne | October Makai Eo ¡January Pb sa
inHatch duction | |
e Mie sc lo de a Ge, Pp | 3.7 | 12.7 | 186 | 18.3
March 16..... ie aks A 16 | 11.3 | 168 | 151
arch 23..... eae ee eed Ae een Fas 0.3 | 10.3 | 19.4 | 17.1
March 30..... eke ee ee ee. 0.0 | 7.7 | 13.9 | 14.9
Annt 6o... 29 | 48.6 2.4 | 11.4 | 18.0 | 16.9
oped ID EN TA O ne eee 12 |8.4 | 16.1 | 156
April 20...... Mr eee ea EA POTS 0.2 | 3.3 | 15.9 | 15.8
ADA ec. ES A O E 0. 4.1 | 125 | 15.9
Se Soa 9 | wae ee i... 0.0 4.0 | 19.6 | 18.7
May ll....... B PS POROUS... 00 tsaa | 86) 18.3
TABLE V

AVERAGE WINTER EGG PRODUCTION FOR EACH MONTH OF 1915-16, or Stock
FROM ORIGINAL SOURCE

Number oe
DME IS | Pullets | "pro. | August | tember | October! vember | cember | January] ruary
in Hatch duction
March 22.....| 35. | 41.11 |, 0.0 0.0 2.3 4.2 7.9 12.8 | 13.9
April BF oe. 16 | 23.31} 0.0 0.0 0.0 0.0 2.6 8.4 | 12.3
TABLE VI

A COMPARISON OF THE PROPORTION OF THE FLOCK LAYING EITHER MORE OR
LESS THAN , TOGETHER THE MEANS

OF THE RESPECTIVE GROUPS, GROUPED ACCORDING TO THE MONTH
PRODU: HAVE BEEN OMITTED A AND ALSO THE

_ Few BIRDS THAT LAID EXACTLY 30 Eces

Over 30 Under 30

o Meher Bre ay age ag dere Seaman apd
February... 63.0 | 43 84.3 15.7 18.8 8
March..... 54.9 77 68.8. 31.3 16.4 35
Aon. 46.3 88 58.7 41.3 16.0 62
May o 40.5 29 26.6 73.4 16.1 80

No. 614] EGG PRODUCTION 91

The influence of the time of hatching on winter egg pro-
duction is shown in Fig. 8 and Table IT and III based
on the records for 1915-16. The lower graph in Fig. 8 is
for birds hatched in March; the middle graph for birds
hatched in April; while the upper is for the May-hatched
birds. Similar data for 1913-14 are given in Table IV.,

PERCENT
de

BN NGAGAN@S

O FS 610 11-15 16-20 21-25 26-30 SI-35 36-40 41-45 46-50 51-55 56-60

PERCENT
osuna SN

1-73 76-80

0 FS 610 1-15

PERCENT
A A

PS G10 11-15 1220 31-25 2630 3135 36-40 41-43 46-0 SISS Sé-60 6l-6S 66-70 1-TS 7-80
ECOS

o

Fig. 8. The effect of time of hatching on winter egg production, the graph

. M. for March =
ullets anal i April. Upper curve, pullets hatched in May

epee = 35.4, May = 22.5; S.D. for March = 23.50, April =17.91, May

= 15, a C 38 for March = 5.5, April =55, May = 69.2.
I med to the pone that for Rhode Island Reds, the zeros peep a
j March 1 ma no

wholl artificial class for reasons given in ae a hence,

red division point in t de ne at which the birds begin to lay. aee
the zero class has been kept separate and ws sea in calculating these con-
stants. If M includes the zero class, its value for March is 39.8, for April 29. 9.81

tan
and for May 18.1.

92 THE AMERICAN NATURALIST | [Vot LH

while Table V. should also be examined in this connection.
It is clear from these graphs that the earlier hatched
birds -are superior to the later hatched for winter egg
production. In the former group there are fewer zero
and more very high producers than in the last. There is
also a marked difference in the number of birds in each of
the several hatching months that laid over thirty eggs
during the winter period. (See Table VI.)

The variability in the age at which the first egg is pro-
duced influences the winter record strongly, so much so
that we have been led to believe that it is the most im-
portant determining factor for egg production during the
winter months for our flocks.* It leads to the abandon-
ment of the view that those records that fall below 30 eggs
are made by true mediocre producers, substituting there-
for the view that many, perhaps most, of them are late-
maturing high producers. Now, the variability in age at
first egg, shown in Figs. 1 and 2, is considerable. If this
variability could be eliminated—that is, if it were possible
to have each individual of a flock of birds hatched April
1, begin to lay on a definite date, say December 1—the
birds would make records which would differ from each
other in the proportions given by the graphs of rates of
production. The fair degree of homogeneity of the flocks
in respect to rate of production is shown in part by the
coefficient of correlation between the number of days from
the first egg laid up to March 1, and the number of eggs
produced during that period. The coefficient was found
to have a high value, i. e., for 1913-14, r—-+ .8618 +
.0125 and for 1915-16, r— + .7878 + .0128. That is, the-
number of eggs laid is a fairly definite function of the
length of the laying period. These coefficients are a
rough index of the amount of homogeneity in the flock
respecting the rate of production, since a high coefficient
implies a fair amount of homogeneity in the flock (cf.,
however, the statistical constants for rate) for if an egg a
day is taken to represent the maximum production while

3 The results obtained from our breeding tests substantiate this point.

No. 614] EGG PRODUCTION 93

the minimum is represented by a single egg laid during
the winter and that one at the beginning of the winter
period, no correlation would exist if the scatter is per-
fect between these extremes. Experience shows, how-
ever, that conditions approached by the maximum rate
of production are much more common than those rep-
resented by the minimum so that the coefficient, even
though it has a high value, shows only that the rate of
production is comparatively uniform. It does not prove
that the flock is composed exclusively of high producers,
for, since it is an average figure, the flock may still con-
tain some true mediocre producers. A certain degree of
correlation, however, is to be expected in any flock, so that
the mere existence of a small positive correlation is of
little value, though a low value for the coefficient would
imply that there was considerable variability in rate of
production. It is quite clear that if a considerable per-
centage of the flock made records like those shown in Fig.
12, the variability in rate would be much greater than
observed, and the coefficient of correlation between length
of laying period and number of eggs would be smaller.

Size.—Size does not of itself seem to have any specific
relation to a bird’s ability to lay because birds of all sizes
may lay equally well once they have started. It is true,
of course, that very large birds rarely make high records,
but as there are very few large birds, the chance for a
combination between very high egg production, itself un-
common, and large size is rather remote. The converse,
however, is not apparently true, for small birds fre-
quently make good records. Since, however, birds that
are too small are not desired by poultrymen while as a
rule large birds are considered desirable, very small birds
are not often trap-nested, so that a strict comparison is
at present impossible.

In another way size seems to exert some influence on
the record a hen makes. On the average, as shown by
the coefficient of correlation between age at first egg and
weight, birds of large size reach this size later in life

94 THE AMERICAN NATURALIST [VoL. LII

than small birds do. That is, large size usually, but not
always, results from long-continued growth rather than
from very rapid growth and as long-continued growth
- naturally tends to postpone the date at which the first
egg is laid, the large hen, other things being equal, can
not lay as many eggs. It is possible, however, that the
case may actually be the converse, viz., the hen may grow
large because she does not lay, though there is no definite
evidence for this point of view. While it is easy to find
many instances of small birds that mature late in life,
instances of large birds maturing at the same age as
birds of approximately half their body weight have not
been observed. The reason for this is probably to be
found in the consideration that while some large birds |
may grow more rapidly than some small birds, it is always
possible for some small birds to grow as fast as it is ever
possible for a large bird to grow and hence to mature
that much earlier.
(To be continued)

THE CASE OF THE BLUE ANDALUSIAN!

WILLIAM A. LIPPINCOTT

Kansas AGRICULTURAL EXPERIMENT STATION, MANHATTAN, Kansas

Tr blue Andalusian has become the classic in animals
as an example of a heterozygote phenotypically inter-
mediate between the parental types. It has also served
as an illustration of the failure of dominance for those
opponents of Mendelism who consider dominance one of
its fundamentals. Furthermore, it has been in constant
demand as a classroom example of blended inheritance.

The main facts concerning the breeding behavior of
blue Andalusians are, accordingly, more or less familiar.
In spite of the long-continued efforts of their breeders
they do not come true to color as a breed, but. continually
throw a certain proportion of off-colored progeny, or
‘wasters, of two kinds. One is self (entirely) black.
The other approaches white, but displays considerable
pigment, and is referred to variously as white, splashed,
and splashed-white. Since an examination of a large
number of birds of this type shows the pigmented feathers
to be blue in all sections of the female and in those sec-
tions of the male which carry blue feathers in the blue
Andalusian male, they will be referred to throughout this
paper as blue-splashed.

““Splashed” refers to the fact that the pigment does not
regularly appear in any particular group of feathers or
in any definite region. Feathers located apparently at
random on any part of the body may be pigmented over
their entire surface or may show only slight traces of pig-
ment. Not infrequently both of these conditions are
present in the same individual.

Since the blacks and blue-splashed breed true when

1 Contribution from the Department of Experimental Breeding, Wis-
consin Agricultural Experiment Station, No. 12, and from the Kansas Agri-
cultural Experiment Station. :

95

96 THE AMERICAN NATURALIST [VoL. LII

mated inter se, they are considered as being homozygous.
If crossed they invariably produce blues.

These facts have led to the current view that the case
involves a single allelomorphic pair of characters. The
blacks and blue-splashed represent the homozygous con-
ditions, while the self blue is the heterozygote between
the two. When blues are interbred, blacks, blues, and
blue-splashed are produced in a ratio approximating
1:2:1 for these classes, respectively, which seems to cor-
roborate this view.

Although the blacks and blue-splashed breed true for
color, they are not recognized by fanciers as breeds or
varieties, and it is doubtful whether they would continue
to exist if, much to the disgust of the breeders of blue
Andalusians, they did not continue to appear as ‘‘wast-
ers’’ among the progeny of blues. The blues, on the
other hand, are quite widely bred. They are officially
recognized by the American Poultry Association as a dis-
tinct breed and have their place in the American Standard
of Perfection. It is interesting in this connection to note
that the numbers of blues they throw on the Mendelian
expectation barely gets them into the Standard, since the
rules of the Association are that no breed can be officially
recognized as such unless a minimum of 50 per cent. of
the offspring come reasonably true to type (American
Poultry Association, 1910, p. 328, Constitution, Article
XT)

The blues are quite uniformly bluish-gray throughout
the body, with certain exceptions in the males to be noted
later. Emphasis has usually been laid on their distinet-
ness from the black and the blue-splashed birds, but it
seems important to note their resemblance to these two
classes. In the first place, they are like the blacks in being

self-colored, that is, all feathers in all parts of the body
are pigmented. In the second they resemble the blue-
splashed in that the color of the individual pigmented
feathers is blue rather than black, save in certain sections
of the males of both classes, where the feathers showing

No. 614] THE BLUE ANDALUSIAN 97

pigment are a glossy black, apparently a secondary sexual
characteristic. The blue appearance is due to the distri-
bution and arrangement of the pigment granules in the
feather structure, as will be described later. The fact
that the splashed birds are splashed with blue (with the
exception noted above) rather than black is important
and appears not to have been noted, or at least not em-
phasized, by previous writers.

As an example, Punnett (1911, p. 70), in discussing the
breeding behavior of the blue Andalusian, says: ‘‘It
always throws ‘wasters’ of two kinds, viz., blacks, and
whites splashed with black’’ (italics mine). In the ma-
terial which has come under my observation, consisting of
upwards of one hundred birds in the unrelated flocks of
the poultry departments of Kansas State Agricultural
College and the University of Wisconsin, no individual
has been noted in which the pigmented areas were not dis-
tinctly bluish-gray, except that those pigmented feathers
or parts of feathers appearing in the hackle, back, and
saddle of the male were glossy black. These sections, it
should be clearly understood, are also glossy black in the
blue Andalusian male. There are occasionally flecks or
small spots of black, appearing in the blue-gray feathers,
and even in the white feathers of the blue-splashed birds.
This is also true of the blues and, indeed, is not a rare
occurrence in both dominant and recessive white races of
other breeds. It does not in the least affect the fact that,
in the material so far observed, the white birds have been
splashed with bluish-gray rather than black in those sec-
tions where the blue Andalusian is also blue. This con-
clusion is borne out by the results of a microscopic ex-
amination.

In an effort to determine the fundamental differences
between the three Andalusian phenotypes, a careful study
of feathers from numerous individuals of each phenotype
was made. A detailed account of the results of the study
will be published in a later paper. For present purposes
a short account of the most obvious differences will serve.

98 THE AMERICAN NATURALIST [Vou. LII

The pigment in all three phenotypes is black. The dif-
ferences in appearance are due to the distribution and
arrangement of the pigment or to its absence.

The pigment in a black Andalusian feather is in the
form of rod-shaped granules, which almost completely fill
each cell. They extend to the very tips of both curved
and hooked barbules, and into the tiny hooklets given off
from the barbs of the latter class. The cell boundaries
are usually visible, due, apparently, to a slight contraction
of the pigment, leaving very narrow pigment-free spaces
between the cells. The former position of the nucleus of
each cell is almost always plainly visible, due to an accu-
mulation of pigment at its border, and to a narrow area
surrounding it that bears relatively little pigment. In
appearance, size and distribution the pigment granules in
feathers from the black Langshan seem to be identical
with those of black Andalusians.

The feathers from blue Andalusians differ from those
of the blacks in two important particulars, namely, the
restriction of the pigment in the feather structure and the
shape of the granules. In blues of average shade, pig-
ment fails to appear in the extremities of the barbules of
both types. The hooklets are also entirely pigment-free.
Though not always the case, the curved barbules usually
carry rather more pigment than the hooked barbules,
since the pigment extends further toward the distal end.
As a usual thing that part of the hooked barbule which
bears the hooks is free from pigment and does not differ
in appearance, by transmitted light, from the same por-
tion of a similar barbule from a white feather.

In the pigmented portions the pigment is usually mark-
edly contracted or clumped within each cell, leaving a
pigmentless space about the border much wider than is
the case with blacks. These spaces are not always clean
cut, but may be broken by invading rows of granules, or
isolated granules may be found scattered within them.
As a usual thing the nuclear boundaries in the cells of
blue-gray feathers can only be made out with difficulty,

if at all.

No. 614] THE BLUE ANDALUSIAN 99

In cross-section, the pigment granules are seen to be
scattered through the cortex of the barb and along the
boundaries of the medullary cells. They are not re-
stricted to the apex of the barb, as is reported by Lloyd-
Jones (1915, p. 472, Figs. 37-39) in the so-called blue
pigeon.

The predominating shape of the pigment granules in
feathers from blue Andalusians is round. There may
be a few elliptical granules ¿nd occasionally one which
can not be classified otherwise than as a rod. These are
quite rare, however, and one may carefully serutinize sev-
eral blue-gray feathers without finding any but round or
very slightly elliptical granules. These round granules
quite frequently appear in straight rows, giving the effect
of a string of beads.

While the granule shape may have an appreciable
effect in giving the bluish-gray cast found in blues and
blue-splashed, it seems more likely that, as suggested
above, the bluish appearance is due to the restriction or
arrangement of the pigment. While the condition is not
precisely the same as in pigeons, as described by Cole
(1914, pp. 324-325) and Lloyd-Jones (1915, pp. 472-473),
the optical effect appears to be from essentially the same
causes, namely, the clumping of the pigment within the
cells, and the reflection from this pigment through more
or less transparent layers of keratin. It appears, how-
ever, that in the blue Andalusian the contrast between
the pigment-free ends of the barbules and the pigmented
barbs and barbule-bases is of more importance in produc-
ing the bluish effect than is suggested for the pigeon by
these writers.

A characteristic of the typical blue Andalusian not be-
fore mentioned is that the contour feathers on the female
and the breast feathers on the male present a laced ap-
pearance. This results from a black edging on that por-
tion of each feather which is exposed when in its natural
position. In this part of the feather the barbules on both
sides of the barb are alike, being without hooks. The

100 THE AMERICAN NATURALIST [Vou. LII

cells in these barbules are more heavily pigmented than
is true of the rest of the feather and the granules are rod
shaped. In the regions where the black is giving way to
blue, both round and rod-shaped granules are found.

All pigmented feathers secured from several blue-
splashed females show identically the same pigment ar-
rangement and granule shape as predominates in the
blues. This holds true whether the portion examined
comes from a feather that gs pigmented throughout, or
from one that is almost wholly white, with but a trace of
pigment showing. In feathers which are pigmented
throughout, the same relation regarding the lacing occurs
as in homologous feathers in blue females.

The statements of the foregoing paragraph apply
equally well to the feathers of those sections of the blue-
splashed male which are blue in the blue male.

As previously mentioned, in both blue-splashed and
blues, as well as in other self-colored races, black flecking
or spotting not infrequently appears. Such spots, whether
taken from a blue feather from a blue individual, or from
a blue or an almost white feather from a blue-splashed
bird, invariably show rod-shaped granules, while the sur-
rounding area, if blue-gray, shows round granules. These
spots are apparently entirely independent of the factors
and conditions discussed in this paper and their appear-
ance is comparatively limited. If hereditary, they prob-
ably depend on other factors. In handling blues and
blue-splashed, however, one can not help being impressed
with the possibility that these spots are caused by some
interference with the full expression of the factors re-
sponsible for the arrangement and rounding of the pig-
ment granules. Whether this interference is hereditary
or environmental is as yet undetermined.

One further fact concerning the blue Andalusian males,
already alluded to, is of interest. The long feathers of
the neck (hackle) and saddle are glossy black. This is
apparently a secondary sexual characteristic, though it is
as yet undetermined whether it is due to the presence of

No. 614] THE BLUE ANDALUSIAN 101

testicular secretion or the absence of ovarian secretion.
The black feathers from both sections show rod-shaped
granules predominating. There are numerous elliptical
granules and a few round granules present. The pig-
ment is not restricted as to distribution in the feather
structure and is found even in the tiny hooklets of the
hooked barbules, being in all these respects similar to the
analogous feathers on a black male. These same condi-
tions prevail in homologous pigmented feathers in a blue-
splashed male.

The foregoing describes the conditions that usually
prevail. There is some variation in all conditions de-
scribed. In pure-bred blue Andalusians, for instance,
there frequently appear areas that are not the usual clear
blue-gray, but are dull and smoky. In such regions both
round and rod-shaped granules are found in about equal
numbers.

Bateson and Punnett (1906, p. 20) make note of the fact
that the adult color of Andalusians may be determined
from the down color of the young chicks. Examinations
of the down show the same differences in granule shape
that are observed in the adults. The blue and blue-
splashed chicks for the most part show nothing but round
granules in the down, while the blacks show rods.

It is of interest to note in this connection that a section
from that portion of a barred Plymouth Rock feather
where the black bar is giving way to the white, and the
eolor is dull gray or dun with no bluish cast, there is a
dilution of pigment as to amount, but no restriction as to
arrangement or distribution. The pigment is fully ex-
tended through the barbule cells and consists of rod-
shaped granules. There simply appears to be less pig-
ment. While this is the usual condition, here, too, there
is variation. At least one barred Rock individual was
found whose feathers showed numerous round granules,
though the rods predominated.

While it is generally accepted that blue Andalusians,
when mated inter se, produce blacks, blues and blue-
splashed in the ratio of 1 black to 2 blues to 1 blue-

102 THE AMERICAN NATURALIST (Vou. LIT

splashed, exact data on this mating, as well as on the back
crosses to black and blue-splashed, are really very mea-
ger. Bateson and Saunders (1902, p. 131) first sug-
gested that the blue Andalusian was probably a hetero-
zygote. Bateson and Punnett (1905, p. 118) quoted Mrs.
Blacket Gill, a fancier of blue Andalusians, to the effect
that blues mated to blues gave 22 blacks, 36 blues and 17
white-splashed (i. e., blue-splashed). They secured stock
from Mrs. Gill and made matings which gave the follow-
ing results:

By the blue Y the white 2 gave 34 blue, 20 white-splashed, and the
black 2 gave 27 blue, 19 black. In each case the result is qualitatively
what would be expected if the blue is a heterozygote of black X splashed
white [italics mine]; but whether the departure from equality indicates
that some gametes bear the unsegregated blue, or may merely be taken
as individual irregularities, can not yet be stated.

The same blue cock was bred with a black hen from Experiment 40
(in which the dark birds were unexpected), F„ from White Wyandotte
X Wh. Legh., giving as offspring 10 black, 15 slaty black to bluish.
Hence, therefore, it is evident that the black $ was a homozygous black.
The 10 blacks are the result of the union of the black gametes from the
Andalusian d with those of the 2, and the 15 slaty resulted from the
meeting of the black of the hen with the white-splashed from the
Andalusian.

Bateson and Punnett (1906, p. 20) give the following
summary of the data upon which the case of the blue An-
dalusian largely rests at the present time.

In Report I it was suggested that the blue colour of the Andalusian
is probably heterozygous, and in Report II (p. 118) figures were given
in support of this view. During the past two years additional evidence
has been acquired, and every form of mating has now been tested, with
the following results: y

Result
No. of Experiment Nature of Mating
Black Blue | Wh. Spl.
Rep. H piis 2. | Bitte a X bed ees ek 22 36 17
Ll o sie bit e DMG o eee. ae 42 29
(Total numbers for blue and bue... aS e ye 41 78 39)
: tation (inserted ae ps, ea ee EE 39.5 | 79 39.5]
Rep. T, pls. 3 | WE SL 9 <bhie S... — 34 20
Hon ET e XX bhie oca la 19 27 —
Em. A ets vi a. O's KWH. ap. Oi... nae ha a
MoO: ts Blac M wh. DE dg. dai ey — 20 —
SS ene [Black $X black ociosos 25 TO | K

No. 614] THE BLUE ANDALUSIAN 103

The colour of most of the chickens was determined in the down. In
the blacks the down is black with the exception of the ventral surface,
the tips of the wings, and sometimes parts of the head, which are white.
The down in the blues is slaty-blue, similarly marked with white, whilst
in the white splashed it is of an exceedingly pale blue tint as a rule,
though sometimes practically colourless.

The above figures bear out the view we previously expressed as to the
heterozygous nature of the blues, . . .

The only other definite figures that have come under
the writer’s notice are from W. J. Coates, a blue Anda-
lusian breeder of East Calais, Vermont, quoted by Platt
(1916, p. 665) and referred to by Pearl (1917, p. 149).
These are for matings of blue to blue and are as follows:

White (Blue-
Mating | Splashed) Blue Black Dark Red

F 20d Se ee ae 4 10 3 1
T e E EE A a E 4 5 2 0
EE E E E E 3 3 0 3
DT Al eee eee es 0 12 1 0
PCa eve cae ee 3 3 1 0

| 14 33 7 4

The fact that birds showing dark red appear is unusual
and would seem to indicate that the Coates stock differs
in its genetic constitution from the majority of the mem-
bers of the breed, unless the occasional appearance of red
is a fact usually suppressed by breeders.

Bateson and his co-workers make no attempt beyond

‘that quoted above to account for the hereditary behavior
of Andalusians and appear content to rest the case on the
assumption that ‘‘blue is a heterozygote of black x
splashed white.’’

The fact that ‘‘blue’’ is not a true intermediate be-
tween black and blue-splashed does not seem to have re- .
ceived due consideration. While the blue-gray bird is in
a sense intermediate between self black and an individual
that approaches white more or less closely, this inter-
mediacy is more apparent than real. As previously
pointed out, it is not intermediate in regard to either of
the conditions involved when they are considered sepa-

104 THE AMERICAN NATURALIST [Vou. LII

rately. It resembles the black phenotype in being self-
colored and the blue-splashed phenotype in having the
pigment restriction within the barbules, which gives the
blue-gray effect.

The 1:2:1 ratio may therefore be analyzed as follows:

Pigment not re- Pigmen re- Pigment re-
stricted in bar- stricted ES bar- stricted in bar-
bule cells; es- bule cells; egs- bule cells; not ex-

Phenotypes ..... tended through tended through tended through

(Mie plumage (*“self”” plumage (self); plumage; pheno-
condition); phe- phenotype blue, type blue-
notype black. splashed.

good Boe pheno- } 1 9 1
— —
Ratio for restric- 1 3
tion in barbules
Ratio for exten- a Ra
sion in p um- | 3 i 1
ERE

In reality, then, the 1:2:1 ratio is the result of the com-
bination of two 3:1 ratios.

The foregoing facts appear to lend themselves equally
well to two interpretations. The first is that there are
two pairs of allelomorphic factors at work. The second,
that there is one pair of true allelomorphs (i. e., factors
having identical loci on homologous ehromosonive),
neither of which is recessive to the other in its manifesta-
tion in the phenotype.

The suggestion of two pairs of allelomorphic factors to
explain the case of the Andalusian is not a new one.
Goldschmidt (1913, p. 274) makes such a suggestion.
After pointing out that the offspring of a pair of blues
are black, blue, and ‘‘schmutzigweiss’’ in the ratio of
1:2:1, and that all three phenotypes carry pigment, he
. proposed two factors to account for the condition. The
one is an “Entfaltungsfaktor,'” which brings about a full
development of the pigment. He represents this factor
by “Q” (Quantität) which is possessed by the black race.
The other factor, which is possessed by the ““Weisse?””
race, he calls a “Mosaikfaktor,?? which finely divides the
pigment. This factor he designates M (Mosaik). He

No. 614] THE BLUE ANDALUSIAN 105

finds it necessary to postulate further that Q is closely
linked with m, and M is closely linked with q. Assuming
pigment (P) to be present in all cases, he represents the
““black”? gamete as (mPQ), the ‘‘white’’ gamete as
(MPg), and the F, blues as (mPQ)(MPg). The blue
results from bringing M and Q into the same zygote.
The monohybrid ratio results when the blues are inbred,
however, because of the close coupling of the factors
within the parentheses.

The Hagedoorns (1914, p. 179) also make use of two
coupled factors in accounting for the hereditary behavior
of blue Andalusians. They state:

A blue Andalusian fowl, when mated by us to “recessive” white
hens did not produce as many blue as white chicks, as should result on
the hypothesis, that the white Andalusian is a recessive white (blue and
black Andalusians being heterozygotes and homozygotes for one single
genetic factor), but exclusively blacks and blues in equal proportions.

To account for this result they propose a gene A which
is present in black Andalusians, but absent in the ““white””
Andalusian. The blacks, conversely, lack a gene B which
is present in the ‘‘ whites.’

This factor B, present in a pigmented fowl, actively “dilutes” the
colour. It has no effect in the white Andalusians, because these, as they
lack A, are not pigmented [italics mine]. We should therefore expect
dilute black (blue) young from the cross black white, which, inter se,
would give AB, Ab, aB and ab offspring. Now, there is no evidence
that in Andalusians there are ever produced aabb animals, or AABB.
There seems to be a mutual repulsion between A and B, so that no AB
or ab gametes are ever produced. In some varieties of fowls this repul-
sion does not seem to exist, as pure strains of blue chickens occur.

Unless their material differs from any that has come
under my observation the Hagedoorns err in assumin
that what is frequently termed ‘‘the white Andalusian””
carries no pigment, and Goldschmidt’s suggestion accords
more closely with the facts. Further, if the ‘‘recessive’’
white to which they refer was an Andalusian, the produc-
tion of equal numbers of blues and blacks from a blue X
white (blue-splashed) cross is difficult to understand.
The expectation would be equal numbers of blue-splashed

106 THE AMERICAN NATURALIST [Vou. LII

and blues. If, as I suspect, the ‘‘recessive’’ white was a
true recessive from another race, their results can only
be interpreted by assuming that the ‘‘white’’ gametes as
well as ‘‘black’’ gametes produced by the blue fowl car-
ried a factor necessary for pigment production, which
was lacking in the recessive whites.

If the latter is the case it accords with results I have
obtained the past season. Among several matings made,
preliminary to a further study of Andalusian blue, a white
Wyandotte Y (R 840 from the University of Wisconsin
flock) was mated with blue-splashed Andalusian 22 M
409 and M 539 (also kindly furnished by the poultry de-
partment of the University of Wisconsin). From M 409
seven chicks were hatched, all of which were unmistakably
bluish-gray. Six chicks which failed to hatch, but which
did develop far enough for the color of the down to be de-
termined, were also all blues. From M 539, brought in
late in the season with the hope of increasing the numbers
of chicks from this type of mating, three chicks were se-
cured, which were again all bluish-gray. On the assump-
tion that Wyandotte white is recessive (I am surprised to
find no statement to this effect in the literature) these
results would seem to indicate that a factor necessary for
pigment formation as well as one causing the character-
istic arrangement or restriction of the pigment found in
blues, and both lacking in the Wyandotte, were furnished
by the blue-splashed Andalusian. And further that a
factor for the extension of this pigment to all feathers on
the body was furnished by the Wyandotte. The blue off-
spring from this mating are assuredly not intermediates
between a pure white parent and one that appears to be
nearly white.

It is significant to note in this connection that the blue-
gray offspring of the white Wyandotte X blue-splashed
Andalusian cross show pigment granules that are pre-
dominatingly round, In some individuals they all appear
to be round, while in others some rods may be made
out. The down of black chicks, offspring of a blue Anda-

No. 614] THE BLUE ANDALUSIAN 107

lusian Y and a white Plymouth Rock 9, showed only rod-
shaped granules. Feathers from a blue-gray individual,
whose dam was a blue-splashed Andalusian and whose
sire was a crossbred, the offspring of a Houdan Y X
single-combed white Leghorn ? cross, showed only round
granules.

If, as Goldschmidt assumes, his factors mQ and Mq are
so closely linked that they never separate, and behave
only as a single pair of factors, it is simpler to assume
that there is but one pair of factors. As already pointed
out, however, the discontinuity in the gradations from
blue-splashed to black is such as to lead one strongly to
suspect that two pairs of factors are at work. This dis-
continuity is greatly emphasized in the case of the blue
offspring from the white Wyandotte X blue-splashed An-
dalusian cross. Itis perhaps not impossible that a single
pair of factors should bring about the result found in
Andalusians, but it is so unusual as to make the assump-
tion of two pairs of factors reasonable.

If this assumption is correct it must be further as-
sumed, as Goldschmidt implies but does not state, that
the black and splashed races each contribute a dominant
and a recessive factor, and that in the blues we have the
expression of both dominants, namely, the extension of
pigment to all feathers, furnished by the black (or, in the
Wyandotte cross noted above, by the white) parent, and
the restriction of the pigment in the feather structure in
such a way that the effect is bluish-gray, furnished by the
blue-splashed parent. It is of interest in this connection
to note that the blue condition produced by the restriction
of the pigment in the barbule cells is recessive in pigeons
(Cole, 1914, p. 325), while in Andalusians, on the above
assumption, it is dominant.

While exact data concerning the breeding behavior of
blue Andalusians are gly meager, the experience of
breeders generally seems to be in accord with such data
as there are, and with the interpretation offered by Bate-
son and his associates. In order to account for the fail-

108 THE AMERICAN NATURALIST [Vou. LII

ure to secure a dihybrid ratio from the mating of blues,
one is driven to assume linkage, and apparently a quite
close linkage, of the dominant of one allelomorphic pair
to the recessive of the other. As I hope to make clear,
however, the linkage may not be complete, since it would
easily be possible for crossing-over to occur occasionally
with very slight likelihood of detection.

Goodale (1917, p. 213) has very recently shown that
crossing-over occurs in the sex chromosome of the male
fowl, though he has not as yet presented his evidence in
detail. The universality of the laws of heredity through-
out the plant and animal kingdoms is such that it would
be a matter of surprise if crossing-over in fowls did not
also occur in chromosomes other than the sex chromosome.

There is at present no certain criterion by which to pre-
dict whether, having assumed crossing-over in the auto-
somes of fowls, it occurs in one sex only or in both; and
if in but one sex, which it may be, unless one dins to
suppose that it occurs only in the sex homozygous for the
sex chromosome, as in Drosophila. If it occurs in both
sexes, it is apparently so rare an event in Andalusians
that the probability of securing two cross-over individuals
in a mating made for purposes of analysis is so small as
to be almost negligible.

After a somewhat extended microscopic study of blue-
gray feathers from blues, blue-splashed and certain
crosses, it seems more in accordance with their apparent
action to refer to the factor responsible for changing black
into bluish-gray as a restrictor, designated as R rather
than M (Mosaikfactor) as was done by Goldschmidt.
Similarly in place of Q (Quantität) I would suggest E, as
responsible for the extension of pigment to all the feathers
of the body.

Using this terminology and assuming for the moment
complete linkage, a cross between individuals of the black
and blue-splashed races, respectively, would appear as
follows:

Er Er = black X eR eR = biiabdaabadl:
F, Er eR = blue;

No. 614] THE BLUE ANDALUSIAN 109

F, 1 Er Er=black + 2 Er eR = blue + 1 eR eR = blue-
splashed.

The gametes produced by the F; ( blues) are Er and eR.
If crossing-over should occur there would be occasional
ER and er gametes produced. It is highly interesting to
note that if these two classes of cross-over gametes were
produced in equal numbers, as would be expected, and the
individuals producing them were mated with ordinary
blues, exactly the same phenotypic ratio would result as
from the unions of the non-cross-over gametes, viz.:

F, crossover gametes _ Ordinary gametes of F, blue
, Or Er, e

F, ER Er= blue, ER eR = blue, er Er=black, er eR
= blue-splashed.

This is the usual ratio of 1 black to 2 blues to 1 blue-
splashed and would, from the very nature of the case,
escape observation as involving crossing-over unless care-
ful analysis were made of the hereditary constitution of
these particular F, individuals.

Such analyses would not be impossible, though they
might be long and tedious. The matings which would un-
cover any of the cross-over types, if offspring were pro-
duced in sufficient numbers to make it fairly certain that
one were not dealing with chance variations in the ratios,
are given herewith. |

Cross-over blue of ER Er constitution mated with an
ordinary blue would give the following expectation :

ER Er = blue cross-over X Er eR = ordinary blue;

F, ER Er= blue,
ER eR = blue,

Er Er = black,
Er eR = blue,

or 3 blues to 1 black, while the ordinary blues would give
the normal 1 black to 2 blues to 1 blue-splashed.

Similarly this same individual mated with ordinary
blue-splashed would produce all blues instead of the ordi-
nary expectation of 1 blue to 1 blue-splashed, viz.:

110 THE AMERICAN NATURALIST [Vou. LIT

ER Er = blue cross-over X eR eR = ordinary splashed;

F, ER eR = blue,
EreR = blue.

If blue cross-over of the type ER eR were mated with
ordinary black the expectation would be all blues instead
of the usual blues and blacks in equal numbers, viz.:

ER eR = blue cross-over X Er Er = black;
Fi ER Er = blue,
eR Er = blue.
This second type of blue cross-over individual, ER eR,
mated with ordinary blue, would give an expectation of
3 blues to 1 blue-splashed instead of the ordinary 1:2:1
ratio, Viz.:

ER eR = blue cross-over X Er eR = ordinary blue;

F; ER Er = blue,
ER eR = blue,
eR Er = blue,

eR eR= blue-splashed.

If black cross-over Er er were mated with ordinary blue
the expectation would be 2 blacks to 1 blue to 1 blue-
splashed instead of the ordinary ratio of equal: numbers
of blues and blacks, viz.:

Er er = black cross-over X Er eR = ordinary blue;
j Er Er = black,
Er eR —blue,
er Er = black,
er eR = blue-splashed.

This same individual Er er (black cross-over) mated
with an ordinary splashed bird would give an expectation
of half blues and half blue-splashed instead of all blues,
as in the case of ordinary black and blue-splashed, viz. :

Er er = black cross-over X eR eR = ordinary blue-
splashed;
F, Er eR = blue,
er eR = blue-splashed.

No. 614] THE BLUE ANDALUSIAN 111

Finally, blue-splashed cross-over eR er mated with or-
dinary blue would give an expectation of 1 black to 1 blue
to 2 blue-splashed instead of the ordinary expectation of
equal numbers of blues and blue-splashed, viz. :

eR er =blue-splashed cross-over X Er eR = ordinary

blue;
F, eR Er = blue,
eR eR = blue-splashed,
er Er = black,

er eR = blue-splashed.

The possible matings not indicated in the foregoing are
those which would produce the same phenotypic ratios as
if ordinary individuals (i. e., non-cross-overs) of the same
appearance as the cross-overs were used. Such matings
are naturally of no value for analysis.

If it should later be shown that crossing-over does occur
as suggested above and there are two pairs of factors con-
cerned, there is the possibility of occasionally securing
ER gametes. This in turn would seem to make possible
the blue Andalusian breeder’s long-time dream of pro-
ducing blues that ‘‘breed true.’’ With the appearance of
the double recessive gamete er another race of Andalu-
sian would apparently become possible, which, if the
factors assumed in this paper are correct, should be white
splashed with black instead of with blue.

The second possible interpretation of the facts so far
established is that my postulated factors R and E occupy
identical loci on homologous chromosomes, neither being
recessive to the other in its phenotypic expression. For
the present at least any evidence that this is the correct
interpretation will be largely negative and come from con-
tinued failure to find cross-over individuals with regard
to R and E. If these cross-overs should not be found it
might at first appear that the interpretation of the case
of the blue Andalusian is in all probability exactly what
has been suggested from the first, namely, that blue is
. a heterozygote intermediate between the parental types.

112 THE AMERICAN NATURALIST [Vou. LII

Such an interpretation makes the characters black and
blue-splashed the allelomorphs.

The practise of referring to characters that seem to be-
have in an alternative relationship in heredity as allelo-
morphs, instead of factors occupying identical loci on
homologous chromosomes, is, it is to be hoped, passing.
That it has lead to a misinterpretation in the present case
is shown by the fact that all the offspring of certain pure
white birds mated with blue-splashed ones are blue. The
E factor must have come from an individual that was ho-
mozygous for it and devoid of pigment. It appears rea-
sonable to expect that among the F,’s from the white
Wyandotte X blue-splashed Andalusian cross will appear
pure whites that carry the R factor. If this proves to be
the case the allelomorphs are two factors, R and E, which
act on black pigment. R arranges and restricts the pig-
ment in the feather structure so that it gives a bluish-gray
appearance. ŒE extends any black pigment present to all
the feathers of the body. One and probably either or
both may be present without any phenotypic expression
whatsoever. In fact, for every sixteen F, individuals
from this cross four pure whites are to be expected in
which the genotypic ratio with regard to R and E is1:2:1,
exactly as in the F,'s from a cross of a black and a blue-
splashed Andalusian. One of these whites will be homo-
zygous for R like the blue-splashed Andalusian. One
will be homozygous for E like the black Andalusian. And
two will be heterozygous for E and R, as are the blue
Andalusians. But because there is no black pigment
present these differences in the genotype do not affect
the phenotype. For the sake of clearness the expectation
of this cross is shown herewith, carried through the F,
generation. P is taken to represent a factor necessary
for the formation of pigment which is present in the blue-
splashed Andalusian, but absent in the white Wyandotte,
] es E and R are represented as allelomorphic to each

other.

No. 614] THE BLUE ANDALUSIAN 113

White Wyandotte Y X blue-splashed Andalusian 9.
ppEE PPRR
i PpRE = all blue;
F, 6 blues: 3 blue-splashed: 3 black: 4 white.

Strat 1PPRR 1 PPEE 1ppRR
4 PpRE 2 PpRR 2 PpEE 2 ppRE
1 ppEE

This same ratio (6:3:3:4), which is to be expected on
either interpretation, has been reported by Baur (1914,
p. 95) for crosses between a white-flowered race and cer-
tain plants bearing ivory-colored flowers, of the snap-
dragon (Antirrhinum majus).

Recessive mutations are of comparatively frequent oc-

currence. Dominant mutations, though much less fre-
quent, have been described so often that they can not be
reasonably doubted. There appears to be no reason, a
priori, why a mutation might not occur where the mutated
factors’ potency of expression in the phenotype is approxi-
mately equal to that of the normal factor. That this has
occurred, not once, but several times, might be the inter-
pretation placed on the striking allelomorphic series re-
ported by Nabours (1914, p. 141) for the color patterns
of the grouse locust (Paratettix).
-= Upon which of the two alternative interpretations is
correct appears to depend the possible success or the futil-
ity of the search for true breeding blues. The first makes
it possible. The second appears to close the door of hope
in the Andalusian breeder’s face unless hope is seen in
the progressive selection of the darker blue-splashed in-
dividuals. It does not appear possible, on the basis of
present known facts, to reach a conclusion. Extensive
matings are being made for the coming breeding season
which it is hoped will throw further light on the matter.

SUMMARY

1. This paper shows that blue Andalusians are like
black Andalusians in that they are self-colored. They

kd

114 THE AMERICAN NATURALIST [Vou. LII

are like the blue-splashed in that homologous pigmented
feathers in both sexes have the same condition with ref-
erence to the restriction of pigment in the feather struc-
ture.

2. The fundamental phenotypic differences between
black, blue and blue-splashed Andalusians are briefly
described.

3. It is pointed out that the 1:2:1 ratio is in malty a
combination of two 3:1 ratios.

4. The condition in the blues is shown to be due to the
combined action of two factors R and E. R acts on black
pigment, restricting its distribution in such a way that it
gives the characteristic blue-gray appearance. E extends
black pigment to every feather on the fowl’s body.

5. It is impossible to decide on the basis of present
facts whether R and E are located on identical loci of
homologous chromosomes or are the dominants of two
pairs of factors, each linked to the recessive allelomorph
of the other.

6. It is shown that if the latter is the condition, crossing-
over might occasionally occur between R and E with small
likelihood of detection.? If crossing-over does occur, RE
gametes are possible, which appears in turn to make pos-
sible true-breeding blues.

ACKNOWLEDGMENTS

It is a pleasure to acknowledge my indebtedness to Dr.
Leon J. Cole and Professor Jas. G. Halpin, of the Uni-
versity of Wisconsin. The blue Andalusian problem was
undertaken at Dr. Cole’s suggestion and the work is
being continued under his direction. I have consulted
him freely during the preparation of this paper. Pro-

2In ordinary practice poultry breeders make what are kno own as ‘‘pen
matings,’’ that is, one male is mated to a number of females and the off-
spring from these females are not kept separate. The exact parentage of
any individual is therefore known only with regard to its sire, since its dam
might be any one of the females in the group. As the detection of erossing-
over depends upon the results of individual matings, it would be practically
impossible to discover it under these conditions.

®

No. 614] THE BLUE ANDALUSIAN 115

fessor Halpin very kindly placed stock and equipment at
my disposal for an entire breeding season, without which
it would not have been possible to carry on the breeding
work reported herein.

BIBLIOGRAPHY

American Poultry Association
1 The American Standard of Perfection. 331 pp. Pub. by Amer.
Poultry Assoc.
Bateson, W., and Saunders, E. R.
19 Re note to the Evolution Committee of the Royal Society, I,

pp.
Bateson, W., er Punnett, R. C.
1905. Reports to the Evolution Committee of the Royal Society, II,
99-131

1906. Reports to the Evolution Committee of the Royal Society, III,
pp. 11-23.

Baur, E.
1914. Einführung in die experimentelle Vererbungslehre. 2. neube-
sea Auflage. viii+401 pp. ‘Berlin: Gebriider Born-
aeger

Cole, L. J.
1914, Studies on Inheritance in Pigeons: I. Hereditary Relations of '
the Principal Colors. R.I. Agr. Expt. Sta., Bull. 158, pp. 311-
380, pls !
emana R.
1913. Einführung in die RAET OE Zweite Auflage,
xii + 546 pp. Leipzig: W. Engelma
soas H. D
1917. Crgeatug-over ì in the Sex Chromosome of the Male Fowl. Science,
N. S., Vol. 46, No. 1183, p. 213.
Hagedoorn, A. L., and A. C.
1914 Studies on Variation and Selection. Zeitsch. a Pot Abstam-.
ererbungslehre, Vol. 11, No. 3, pp. 145—
ne Jones, O.

1915. Studies on Inheritance in Pigeons: II. A Microscopical and
‘Chemical Study of the Feather a Jour. Exp. Zool.,
Vol. 18, No. 3, pp. 453-495, pls. 1-7

Nabours, R. K.
1914. par of Inheritance and it hee in Orthoptera I. Jour.
. Vol. 3, No. 3, pp.
Pearl, R.
1917. The Probable Error of a Mendelian Class Frequency. AMERI-
AN NATURALIST, Vol. 51, No. 603, pp. 144-156,
Platt, F. L,
1916. ‘‘Western Notes and Comment.’’ Reliable Poultry Journal,
ol. 6
Punnett, R. C.

THE ROLE OF FACTOR MUTATIONS IN
EVOLUTION

ERNEST B. BABCOCK

PROFESSOR OF GENETICS, UNIVERSITY OF CALIFORNIA

Tux essential features of the mutation theory of evolu-
tion, as proposed by de Vries in 1901, are discontinuity
and heritability of those variations which make evolution
possible. New forms arise from preexisting forms by
saltation; they occur in all directions; they are heritable;
some of them are advantageous to the species and these
are preserved by natural selection. These features are
still recognized as the definitive elements of the mutation
theory, but biologists are gradually changing their point
of view concerning the real nature of mutations them-
selves. De Vries had worked with entire plants as units.
He was searching for evidence of species in the making.
He believed he had found this evidence when he discov-
ered his new evening primroses at Hilversum and found
that they transmitted their divergent characters to their
progeny. The evidence appeared none the less clear to
him, even though the parent species when tested did not
always breed true, but continued to produce not only the
forms first discovered, but also new ones which did not
exist in the original population.

It is not my purpose to discuss the cenothera data which
have accumulated in such enormous bulk in recent years.
Goodspeed and Clausen, in their papers on species hybrids
and the reaction-system concept of Mendelian heredity,
have provided a strong argument for attributing many
of the so-called mutations among the evening primroses
to antecedent hybridization between distinct species.
- These authors have shown that ‘‘the occurrence of the
‘mutants’ (in @inothera) and their subsequent behavior
in hybridization admit of Fog arrangement and inter-

No. 614] FACTOR MUTATIONS IN EVOLUTION 117

pretation without any necessity for assumption of exten-
sive germinal changes.’’ On the other hand, Muller’s re-
cent investigation of balanced lethal factors in Drosophila
led him to conclude ‘‘that some (if not most) of the so-
called mutations in O. lamarckiana are but the emergence
into a state of homozygosis, through crossing over, of
recessive factors constantly present in the heterozygous
stock.’’ If this is correct and these recessive characters
arose as factor mutations, it is obvious that, in basing his
theory of speciation by mutation on the evidence from
(Enothera, de Vries builded better than he knew!

During the decade following de Vries’s announcement
of his theory biological interest shifted from the general
problem of evolution to the more specific problem of
heredity. The rediscovery of Mendel’s law at once
focused attention upon the inheritance of particular char-
acters. Then began the era of experimental evolution in
which, under the leadership of Morgan, most remarkable
progress has already been made. The traditional prob-
lem of heredity, its mechanism, has been solved. We
know, not only that the ultimate hereditary units are
germinal, but also that they are located in that particu-
lar portion of the germ cell called the chromatin, and there
is an ever-growing body of evidence proving that each
hereditary unit occupies a particular locus in a particular
chromosome. These hereditary units have been desig-
nated by various terms, but are most commonly referred
to as genes, genetic factors, unit factors or simply factors.

The germ plasm has come to be recognized as an ex-
ceedingly complex stereochemic system, and, as Reichert
has pointed out, on account of the impressionability and
plasticity of such a system the germ plasm must be ex-
ceedingly sensitive to changes in internal and external
conditions. That factors, however, are relatively stable
entities is being clearly evidenced all the time. But occa-
sionally they undergo definite alteration, doubtless as the
natural result of some new or peculiar set of internal con-
ditions. These alterations in genetic factors, or factor

118 THE AMERICAN NATURALIST [Vou. LII

mutations we shall call them, have been most thoroughly
investigated in the common fruit fly, Drosophila ampelo-
phila, in which species Morgan and others have discov-
ered over 150 factor mutations, each of which is inherited
in strict conformity with Mendel’s laws, when tested in
contrast with its normal mate as it exists in the wild
type. This has become the classical evidence for the
theory of factor mutations. Furthermore, it is now com-
mon practise to refer all Mendelizing characters to their
hypothetical representatives in the germ cell, the genes
or factors; and we speak of a pair of contrasted charac-
ters which are irherited according to Mendel’s rule for
the monohybrid as due to a single factor difference.

It is well known that Mendelizing characters exist gen-
erally throughout all groups of sexually reproduced or-
ganisms. Therefore it appears that factor mutations are
of general occurrence. The data on factor differences will
undoubtedly continue to increase in volume, and as they do
our knowledge concerning the relative frequency of fac-
tor mutations will become more precise. In general Men-
delian phenomena have been observed mostly in conspicu-
ous characters, but certain minute character differences,
such as forked bristles in Drosophila and size and shape
of starch grains in peas, are inherited in Mendelian
fashion. Factor mutations, therefore, are sufficient to
explain the origin of all differences between varieties, and
doubtless they provide the necessary point of departure
in the origin of new races. If the new characters thus
produced are beneficial or advantageous, then natural
selection will cause them to be preserved. Sumner has
recently discovered a number of interesting mutations in
deer-mice (some as yet unpublished) and has shown that
isolation may assist in differentiating local races. The
writer is not unmindful of the earlier discussions of Wag-
ner and others, and later of Jordan and others, on isola-
tion as a cause of evolution. Many biologists are still
inclined to think of geographical differences as the deter-
minative condition in the production of new species. For

No. 614] FACTOR MUTATIONS IN EVOLUTION 119

example, Harrison has discovered that certain species of
moths, which are natives of different continents, but which
resemble each other so closely morphologically as to be
sometimes indistinguishable, exhibit extreme physiolog-
ical differences. These physiological divergencies were
indicated by the failure of hybridization between these
species to produce offspring which were viable, or, if
viable, which were fertile. Harrison concludes that geo-
graphical differences play a very important part in the
production and accentuation of such physiological diver-
gencies.

The rôle of the environment in the production of factor
mutations is still an unsolved problem. As Caullery
points out, the rôle of external factors in the production
of mutations is no longer very clearly or directly appar-
ent. It even appears that factor mutations occur in ‘‘all
directions’’ quite independently of those elements of the
environmental complex which are outside the organism.
This does not mean that factor mutations are not caused.
Like any other natural event, they must be dependent upon
or conditioned by certain antecedent events, and, a priori,
there is no reason why such antecedent events should
not occur outside the organism. In other words, it is
reasonable to suppose that specific elements of the ex-
ternal environment might induce permanent alterations
in genetic factors. But, as yet, such a specific relation
between the external environment and factor mutations
can not be said to have been determined beyond reason-
able doubt. On the other hand, migration, isolation and
geographical differences along with other elements of the
environment play an important róle in the selection of mu-
tations, and must, therefore, be recognized as of funda-
mental importance in organic evolution. It is conceiv-
able, indeed, that, given the occurrence of factor mutations,
the continuous impingement of some definite element in
the environmental complex during long periods of time
might condition a definite orthogenetic trend in phylogeny,
as in the evolution of the elephant and the horse. But

120 THE AMERICAN NATURALIST (Vor LIT

may we safely assume the oceurrence of the necessary
factor mutations? The fact that such mutations are
known in many species and that in Drosophila the same
mutations have arisen anew in the same loci of homol-
ogous chromosomes of different pure strains would cer-
tainly indicate that we may. A factor mutation probably
involves some sort of change within the group of similar
molecules occupying a particular locus in a particular
chromosome. Obviously the number and direction of the
changes possible in such an entity are limited and the sum
of the limits of change in all the loci in the chromosome
group of a given species would define the limits of factor.
mutations for that species. The limits and direction of
these mutations must have some bearing and may have
intimate bearing upon orthogenetic trend.

Factor mutations produce new morphological and
physiological characters such as distinguish the forms,
races or varieties of existing species. As they occur
generally in animals and plants at the present time, we
may safely assume that they have occurred in preexisting
organisms more or less frequently, and, therefore, that
they have played a definite róle in evolution. But just
how extensive is this róle? Can we account for the whole
process of organic evolution including the origin not
only of species, but also of genera, families, orders and
phyla upon the basis of factor mutations? To be worthy
of serious consideration a theory of evolution must ae-
count for the development of the organic world as we
know it at present. Can the hypothesis of evolution
through factor mutations fulfil this requirement?

Tt is well known that in many genera some of the species
differ in their chromosome number. Do factors play a
róle in determining chromosome number? It is possible
that they do. It is conceivable that a factor mutation
might arise which would so alter the physico-chemical re-
lations between different parts of the chromosome as to
cause the chromosome to break at some point. Yet
chromosomes are genetic units of a higher order than

No. 614] FACTOR MUTATIONS IN EVOLUTION 121

factors, each chromosome containing many factors and in
general behaving as a continuous entity. Glaser has re-
cently suggested that while the chemical forces determin-
ing the specific structure of individual molecules may be
precisely analogous to those which account for the nature
of the hexose molecule, for example, yet aggregation into
linear series in the case of the chromosomes very likely
involves elements not strictly molecular. It seems to be
necessary, therefore, to postulate some process by which
these major entities become altered in number or recom-
bined in entirely new systems. We are dealing here
with phenomena of a different sort for factor mutations,
` and the latter appear, therefore, to be of slight signifi-
cance in the origin of species having unlike chromosome
numbers. Alterations in chromosome number may be
brought about either by the unique or irregular behavior
of one or more members of a chromosome group or by
hybridization between species. Natural hybrids between
distinct species of plants are of not infrequent occur-
rence and, according to Lotsy, species hybrids are known
in the following groups of animals: Echinodermata,
Vermes, Arthropoda, especially the insects, Mollusca, Am-
phibia, Aves in which even generic hybrids are known,
and among mammals where there are several well-known
cases of fertile hybrids. As for unique chromosome be-
havior, several different types have been discovered and
are known to occur occasionally. Those which contribute
directly to chromosome group evolution are: (a) non-dis-
junction of homologous chromosomes in the heterotypic
or true reduction division, preceding gamete formation;
(b) failure or retardation of the reduction division, re-
sulting in chromosome groups of three, four or more
times the haploid number of the parent species; (c) frag-
mentation or loss of one or more chromosomes, resulting
in gross changes in the germinal reaction system and
hence potentially in new species. The occurrence of the
last type has not been proved, but, from his cytological
investigations of several species of Drosophila, Metz in-

122 THE AMERICAN NATURALIST [Vou. LII

fers that within this group of species there has been a
definite evolution of chromosome groups.

The known methods of species formation, therefore,
may be described as follows: (1) Factor mutations, caus-
ing more or less extensive heritable somatic changes, some
of which are adapted to the environment and persist.
These, under the influence of natural selection, provide
the means for gradual differentiation of groups having
the same chromosome numbers. Presumably these groups
would be recognized at successive stages in the process
as geographical or ecological forms or races, distinct
varieties and, ultimately, related species. (2) Chromo-
some group alterations, which produce new and sometimes
inconstant forms, but which may also produce true species.
(3) Species crosses, which are known to give rise to new
types, some of them constant, but mostly inconstant
forms, all of which are cryptomeres, potentially capable
of throwing new combinations of parental characters in-
definitely. The possibility should also be noted here that
new constant types, having different numbers of chromo-
somes from the parent species, might originate as species
hybrids.

Factor mutations occasionally produce new dominant
characters. This fact, now fully established by various
investigations, is of considerable theoretical significance.
It has been a common practise during recent years to ex-
plain the origin of recessive characters as due to verlust
mutations, i. e., mutations due to the “loss”? of factors.
This conception has been associated with the much used
though inadequate “presence and absence”” hypothesis,
according to which the only relations which can exist
with respect to a certain factor depend on its presence or
absence from the hereditary material. Difficulties are
met when attempts are made to explain the origin of domi-
nant mutations in terms of this hypothesis, for in such
cases it is necessary to assume that a factor has been
added to the hereditary material. As a result of employ-
ing the presence and absence hypothesis in genetic nomen-

No. 614] FACTOR MUTATIONS IN EVOLUTION 123

clature we have the conception of evolution, recently sug-
gested by Bateson and expanded by C. B. Davenport,
which holds that ‘‘the foundation of the organic world
was laid when a tremendously complex, vital molecule,
capable of splitting up into a vast number of kinds of
other vital molecules, was evolved,’’ and that the process
of evolution may be described as the unpacking of this
‘‘original package’’ by the process of loss of factors or
portions of factors. Various lines of evidence indicate
that changes in species are the result of some process of
factor changes. But those who adopt a physico-chemical
conception of factors and factor changes will find it un-
necessary to imagine the ‘‘primordial ameeba’’ in which
was laid the foundation of the organic world as possessing
in some mysteriously generalized condition all the genetic
factors that comprise the hereditary complex of the genus
Homo, since it is not by the ‘‘loss’’ or ‘‘fractionation”’ of
factors that hereditary changes have been wrought. A
gene does not ‘‘drop out”” or ‘‘split up”” into two or more
—rather a gene or factor is altered so that its reactions
condition a different somatic end product. It is not by
loss of factors, but by changes in the composition of fac-
tors, supplemented by intercrossing, that new races, new
varieties, and new species having the same chromosome
number, originate.

Definite organization of the ‘‘hereditary substance par
excellence,’’ the chromatin, probably occurred in certain
prototypes of existing organisms, in which the chromatic
substance was not differentiated from the remainder of the
cell plasm. The recent papers of Minchin on the evolution
of the cell and of Troland on the enzyme theory of life
show that the most probable early course of evolution was
from that unorganized state typical of the Chlamydozoa,
which are supposed to consist of free chromatin material,
up through advancing degrees of differentiation between
the specialized hereditary substance and the remainder of
the protoplast. Between this fairly satisfactory concep-
tion of the earliest steps in evolution and the ever-

124 THE AMERICAN NATURALIST [VoL. LII

strengthening evidence that factor mutations furnish the
means for evolutionary change within existing species, it
must be admitted that there is a wide gap which needs to
be filled. May we not look to future studies on the
phylogeny of the chromosomes to supply this need in some
measure? I venture to suggest that more work like that
of Metz on the chromosome groups of related species will
prove to be an important source of further light on this
problem.

We may now consider the róle of factor mutations in
more detail. It is certain that even those species having
the same chromosome number differ as a rule in many
unit factors. Hence in order to explain the origin of one
from the other it is necessary to assume one of three pos-
sible methods of procedure.

1. There may have been one or more factor mutations
having manifold somatic effects. That profound somatic
differences such as would distinguish species from one an-
other are sometimes produced by single factor mutations
is proved by the results of the crosses between the oak-
like walnut and its parent, the California black walnut,
which I have described in earlier papers. The mutant
form, unlike most so-called monophyllous varieties, differs
from its parent in every gross external feature, yet it
behaves as a simple recessive in F, and F,. However,
factor mutations which induce such extensive somatic
changes seem to be exceedingly rare.

2. There may have been simultaneous mutations in sev-
eral factors, thus producing full-fledged an individual of a
new species. Objection to this hypothesis is found in
the observation that factor mutations always or nearly
always occur singly, i. e., a single factor mutation in a
given individual at a given time. This observation is

ble since the probability of the oc-
currence of two: factor mutations in the same individual
at the same time, according to the principle of least
‘Squares, would be the inverse ratio of the square of the
number of typical (unchanged) individuals in the popu-

No. 614] FACTOR MUTATIONS IN EVOLUTION 125

lation. Thus, for example, if one factor mutation occurs
in one individual among say 1,000, then the probability
of two factor mutations occurring in the same individual
at the same time would be once in 1,000,000 times. More-
over, individuals showing even one factor mutation are
comparatively rare. Hence, it appears extremely doubt-
ful that any species have arisen through simultaneous
factor mutations in single individuals.

3. Single factor mutations may have occurred in dif-
ferent individuals within a group either simultaneously.
or successively. This hypothesis implies that individuals
possessing certain mutant characters are capable of main-
taining themselves in the wild state, an assumption which
is justified by the fact that factor mutations are known
which have not impaired vitality and fertility, nor re-
duced the general adaptability of the organism. Further
evidence to support this hypothesis is found in the wide-
spread existence of composite species. Although these
heterogeneous groups have been classified as species, they
are really aggregates of numerous distinct varieties or
subspecies. Indeed, their Mendelian behavior indicates
that many advantageous physiological characters, such
as earliness or lateness of maturity, resistance to disease,
high fecundity or possession of a certain pigment, origi-
nate through factor mutations. Such mutations fre-
quently, though not necessarily, involve morphological
ehanges also. In all except strictly autogenous (self-fer-
tilized) species new combinations of mutant characters
would occur through intercrossing, thus increasing the
chances of beneficial or advantageous results to the
species. Populations of such species consist of indi-
viduals of heterogeneous germinal constitution, and fre-
quently disadvantageous or even lethal factor mutations
persist in heterozygous condition, but soon make their
presence manifest when inbreeding or self-fertilization is
practised. Autogenous species, on the other hand, are
composed of individuals of homogeneous germinal con-
stitution (pure lines) which have arisen through the re-

126 THE AMERICAN NATURALIST [Von. LII

currence of factor mutations. When in a certain germ
cell of such an individual a mutation occurs which will
produce a detrimental or lethal effect, the offspring will
either die or fail to reproduce as the case may be. Hence
pure lines possess no lethal or highly detrimental factors;
yet these biotypes may vary between wide limits in their
morphological and physiological characters.

There is no necessity whatever for the simultaneous
appearance of mutations in order to establish new forms.
«Only three conditions are necessary in this method of

-evolution: the existence of species during long periods of
time; repeated occurrence of some factor mutations re-
sulting in new characters advantageous to the species; and
the transmission of these mutations from generation to
generation. All these conditions are known to exist. In
fact, the repeated occurrence of the same mutation in the
same locus of a particular chromosome has been observed,
as well as the occurrence of different factor mutations
producing similar somatic variations,

Factor mutations, therefore, provide the means for
eradual evolution within species; only a few, to be sure,
out of many factor mutations being preserved, but these
few being sufficient, with the frequent aid of migration or
isolation by geographical barriers, to build up new groups
which can be recognized only as distinct species. But
these new species, it will be understood, would have the
same chromosome numbers as the ones from which they
arose. We are, therefore, considering here only one of
several methods by which new species originate. Strictly
speaking, the only true mutations are factor mutations,
as they are the only known germinal variations.

CONCLUSION

Factor mutations occur in accordance with the general
scheme of the mutation theory as formulated by de Vries.
They arise suddenly, they occur in all directions, they are
heritable, and some of them are advantageous to the
species and are preserved by natural selection. When so

No. 614] FACTOR MUTATIONS IN EVOLUTION 127

preserved they give rise to new forms or races, and when
fostered by man they make possible new horticultural
varieties of plants or new breeds of animals. It is prob-
able, as Morgan has shown, that factor mutations alone
have furnished the necessary germinal changes to make
possible the evolution of the elephant’s trunk and similar
cases of orthogenetic development which have been dis-
covered by paleontologists. But factor mutations alone
are not sufficient, so far as we know, to account for the
origin of species of different chromosome numbers, much
less for the appearance of phyla and genera. It is to be
hoped that light will be shed on these more obscure phases
of the general problem of organic evolution through a
combination of taxonomic, genetic, cytological and physio-
logical researches. It would seem that the solution must
involve the expression of relationships between organic
groups in terms of the morphology and physiology of
the chromosomes.

LITERATURE CITED
Babcock, E. B.
1916. Studies in Juglans, III; Further Evidence that the Oak-like
Wal ng Sit ae by Mutation. Univ. of Calif. Pub. in Agr.
Bei, 1
Bateson, Wm.
1914. Address of the es of the British Association for the Ad-
vancement of Scie Sci., N. S., 40, pp. 287-302; 319-333.

2,

Caullery, M.
1916. The Pal hoes of the Problem of Evolution. Sci., N. S., 43,

pP.
Clausen, R. E., an si eos T EL
1916. Hereditary Reaction System Relations—an Extension of Men-
delian Concepts. Proc, Nat. Acad. Sci., 2, p. 240.
Davenport, C. B.
A The Form of Evolutionary Theory that Modern Genetical Re-
search Seems to Favor. Am. NArt., 50, pp. 449-465.
Glaser, O, C.
1916. The Basis of Individuality in Organisms. Sci., N. S., 44, pp-
cei

Goodspeed, T. snd Clausen, R. E.
1917a, og a: of the F, Species Hybrids between Nicotiana syl-
vestris and Varieties of Nicotiana Tabacum, with Special Ref-
erence to the Conception of Reaction System Contrasts in
Heredity. Univ. of Calif. Pub. Bot., 5, pp. 301-346.

128 THE AMERICAN NATURALIST [Vor. LII

1917b. REIR, Factor Differences versus Reaction pen Con-
asts in Heredity. Am. NaT., 51, pp. 31-46 and 92-101.
Harrison, J. W.
1916. Studies i in the Hybrid Bistoninæ. Jour. Genetics, 6, pp. 95-161.
Jordan, D. 8.
1905. The Origin of Species through Isolation. Sci., N. S., 22, pp.
45-562.

Lotsy, J. P.
1916. Evolution by Means of Hybridization. The Hague.
Metz, C. W.
1914, Chromosome Studies in the Diptera, I. A Preliminary Survey
of Five Different Types of ad Groups in the Genus
- Drosophila. Jour. Exp. Zool., 17, Ao
19164. ES Studies in the Dipte oo The Paired Associa-
of the Chromosomes in the Motera, and its Significance.
Pe 21, pp. 213-264.
19160. arpa Studies in Diptera, III. Additional Chromosome
oups in the Drosophilide. pan Nar., 50, pp. 587-599

chin, E. na
1916. The Ss ai of aes Cell. Am. Nar., 50, pp. 5-38; 106-118.
Morgan, T. H., Sturtevant, A. H., Muller, H. J., and Bridges, C. B.
1915. The eva of Sane Heredity. New York.
Morgan, T.
1916. A Critique of the Theory of Evolution. Princeton.
Muller, H. J.
1917. ar Gnothera-like Case in Drosophila. Proc. Nat. Acad. Sci.,
3, pp. 619-626.
Reichert, E. T.
1914. The Germplasm as a Stereochemic System. Sci., N. S., 40, pp.
649-661,

Troland, L.

1917. Biological Enigmas and the Theory of Enzyme Reaction. AM.
AT., 51, pp. 321-350,

Vries, H. de.
1901. Die Mutationstheorie. Leipsic.
Wagner, M

1868. “*Die Darwinische Theorie und das Migrationsgesetz der Organ-
ismen,’’ cited =e Kellogg, V. R., in Darwinism To-day, N. Y.

PHYSIOLOGICAL PROBLEMS IN THE LIFE-
HISTORIES OF ANIMALS WITH PAR-
TICULAR REFERENCE TO THEIR

SEASONAL APPEARANCE!

PROFESSOR VICTOR E. SHELFORD
UNIVERSITY OF ILLINOIS
I. [INTRODUCTION

Tue fact that plants flower, fruits ripen, insects appear
and disappear in succession throughout a growing season
needs no statement even to the savage huntsman or the
city flat dweller. The variations of the usual succession
of appearances with peculiar seasons, unusual weather,
etc., are general guides to many operations of primitive
agriculture and matters of comment by all out of door
people. Seasonal succession has long been scientifically
investigated (see Alee, *11; Forbes, 16; Harvey, ’08;
Hough, ’64; Johnstone, 08; Shelford, '13). Only re-
cently has careful investigation of it been stimulated by
the general interest in modern ecology and economic prob-
lems. The importance of a knowledge of delayed germi-
nation of seeds to the agriculturist (Crocker, ’06) has
further stimulated work along a line throwing light on the
general subject. The analysis of the physiological causes
of the normal succession of biological events in any season
calls on many of the laws of biology to formulate merely
the outline or even a portion of a life history, as, e. g., the
answers to questions such as why potato beetles appear
from hibernation at a certain time, and not earlier, and
deposit eggs on plants of the genus Solanum. Further,
as soon as we concern ourselves with the analysis of the
causes of the irregularities of appearance in any season,
the evident complication of problems is such that one may

1 Contribution from the Illinois Natural History Survey and from the
Zoological Laboratory of the University of Tlinois, No. 100.

129

130 THE AMERICAN NATURALIST [ Vou. LIT

undertake to discuss them without apology. The prac-
tical significance of variations to agriculture is shown by
the destruction of the wheat crop in the Southwest by the
wheat aphis. This was due to differences in response to
weather on the part of the pest as compared with its ene-
mies. The cause of seasonal appearance, or more espe-
cially of variations of seasonal appearance, is to be
found in the influence of external factors on the initiation
and velocity of growth and on fecundity and length of
life, in dormancy in various stages in the life histories,
and in the adjustment of the innate rhythm to the annual
climatic cycle.

Tl. THe INFLUENCE or EXTERNAL CONDITIONS ON TIME OF
PEARANCE AND NUMBER OF INDIVIDUALS

1. Differences in Initiation and Velocity of Develop-
ment.—The problem of initiation of development is one
that has attracted much attention of late on account of the
importance of an ability to predict the time when various
insect pests will emerge from hibernation or will reach a
stage of development at which it is necessary to spray,
if such treatment is to prove satisfactory.

In this connection attention has been directed to the
conditions, particularly of temperature, under which
there is no development during periods lying within the
bounds of the ordinary life history of the animal in ques-
tion (Sanderson, '10, Headlee, and Peairs). The limit at
which development does not take place, usually called
physiological zero or zero of development, is better termed
threshold of development. Sandersonhas discussed vari-
ous data and theories relative to the effect of temperature
on development.

The attention of physiologists has been directed toward
the study of the effects of temperature on the rate of
metabolism and development. In general the results
of such study have been interpreted with reference to
Van’t Hoff’s law relative to the increase of reaction
with a rise of temperature of 10 degrees, usually des-

No. 614] PHYSIOLOGICAL PROBLEMS 131

ignated as Qio Vt is the velocity of development at
aor ; V(t+01)

y temperature (+), so that Q,, is the quotient of ier: T
and supposedly is a constant. In fact, the Q, is not a
constant for living phenomena, but usually varies from
2 to 3, being greater for the lower temperatures and
smaller for the higher ones. Snyder has pointed out in
detail that while the temperature coefficient for differ-
ences of 10 degrees varies, the variation is not only for
physiological actions, but also for many chemical reac-
tions; in both cases the variations are in the same direc-
tion. He finds that changes in viscosity with changes in
temperature follow the same rule. He holds the hypothe-
sis that even in the simpler physiological actions we have
to deal with at least two distinct chemical actions whose
fundamental velocities at any given temperature are
different.

Recently Krogh ('14) has calculated the Q,, from Q,,
Q, and the like, at different temperatures for the time
from fertilization of the frog’s egg to the appearance of
the first cleavage plane. He found 53.0 (published as
5.3, which appears to be a error) for the interval be-
tween 3 and 5 degrees, 4.1 for the interval 5 to 10
degrees, 2.0 for the interval 15 to 20 degrees. He raises
the question as to the value of such a variable ‘‘con-
stant.” He calls attention to the fact that the velocity
curve (the reciprocal of the time-temperature curve)
is a straight line within certain limits. This is not
the curve for the reciprocal of Van’t Hoff’s time and
temperature formula. The latter law is valuable only as
evidence that the life process is a combination of chemical
processes. The condition of any environmental factor at
which development does not take place, but immediately
above which development may be initiated, is called the
threshold of development. It is evident that there is a
threshold of development for most species as regards
temperature, moisture, light, oxygen, quantity and qual-
ity of food, and probably other factors. The brief state-

182 THE AMERICAN NATURALIST [Vou. LII

19
it
etd 30 AY ]
18
17 DR
tf GJ
4 \
4c 1 on
“Vv
12 15 Y. f P
19 tJ Pi
N N
14 10 A
HEHH } s
1A R EE
iv Vir LIA
2 sel Eu a
Enis EN
Pgh A ee Be
12 ra erie tig ELL Lil
) 10 t5 20 do 30- 35
i fig. 2
1
+ E p
a
J
Q 1 ^
o .
pd
t A J
6
1
5
4
a
v
9
LA
4 iB
y
1
F i y
0 5 10 15 20 25 30
Fapt
Fic. 1. Showing the time-temperature curve (the longer curve) ooh the ap-
pearance of the first cleavage plane in the = of the frog. The solid curved line
is the actual cu and the broken 1 ensions are theoretical, basea $ ina
hi al constant (time x temperature) om it differs
shorter oblique curve » the SEAR curve, or the a $2; geg cy
s for a by the combined broken and
solid lines of the longer curve. It is a cl sg ine. The of h
pr e continuation of the curve based on Verworn's ('99, p J
irritability curve. on xis of abscissas are rees Centigrade.
figures he ordinates represent 100 minutes for the hyperbola and 100
divided by the time units for the reci (from data by Kro;

procal h).
FIG. ON the time-temperature and velocity curves for the time from

- No. 614] PHYSIOLOGICAL PROBLEMS 133

ments and citations below are in support of this state-
ment.
(a) Temperature Threshold

Very nearly at the time of the publication of Krogh’s
work, Sanderson and Peairs announced that for a large
series of insects the time-temperature curve for develop-
ment is a hyperbola and the velocity of development
curve is a straight line. Peairs concluded further that
the ( cool relative velocity curve which is obtained
by dividing unity (100 to avoid fractions) by the experi-
mentally determined time periods, in days or other units,
and plotting it against the temperature for which the time
was observed, gives points for the different temperatures
which fall in a straight line crossing the axis of tempera-
ture at the zero of the curve, or the theoretical threshold
of development (Fig.1). This theoretical threshold may be
calculated also with two points accurately determined ex-
perimentally. These authors conclude that with the zero
determined, the thermal constant (temperature multiplied

y time—a constant for an hyperbola) can be obtained.
However, they failed to note the deviations from law
which occur at both high and low temperatures and which
require careful attention in practical work.

These entomological workers appear to have over-
looked the work of the fish culturists who have studied
the subject of effects of temperature on development.
Apstein (711), Dannevig (’94), Earll (78), Green (770),
Johansen and Krogh (714), Reibisch (*02), and William-
son (’08) all made contributions of greater or less im-
portance. All called attention to the effect of tempera-

birth to maturity of Toxoptera graminum (from Headlee after Sanderson and
curves for the development of the sil of four species of marine tak yo
figures on the axis of abscissas represent de

grees Centigr:
axis of ordinates are days for the hyperbola and 100 divided by days for its
reciprocal,

134 THE AMERICAN NATURALIST [ Vou. LIT

ture on the rate of fish development, particularly during
the late embryonic stages. Reibisch (’02) showed that
time temperature is a constant, using the hyperbola.
He called the temperature at which development could be
initiated by the slightest increase the ‘‘ threshold tempera-
ture,” which is the same as the zero of development and
physiological zero of other authors. He calculated this
from the hyperbola formula, thus anticipating the work
of Sanderson and Peairs by about eleven years. In fact,
the idea of ineffective temperature below a minimum and
a sum of temperatures which is the product of time X
‘temperature dates from de Candolle’s 1830 article.
Johansen and Krogh worked over the data of Danne-
vig and showed that the velocity is different for different
fishes (Fig. 2, 4, B,C). They note further that tempera-
ture is not absorbed by the organism and that the constant
is only a convenience. They call attention to the fact
that the velocity-of-development curve is a straight line
which, prolonged downward, crosses the axis of abscissas
at a point mathematically corresponding to Reibisch's
threshold temperature. The threshold of development
would be where the velocity curve crosses the axis of ab-
scissas if the straight-line velocity curve held good and
the time-temperature curve were a true hyperbola. Krogh
(714) showed that while 2.7 degrees is the mathematical
threshold of development for cleavage of the frog’s egg,
the first cleavage appeared at this temperature 1,844 min-
utes after fertilization. If the curve were an hyperbola,
at 2.7 degrees the development of the cleavage plane
should have required an indefinitely long time; or, in
other words, it should not have appeared at all. Also, at
4.9 degrees the appearance of the cleavage plane should
have required 1,100 minutes, while the observed time was
approximately 730 minutes. Further, it required 138
minutes for the cleavage furrow to appear at 22.1 degrees,
which is more than at 20.7 degrees, showing a decrease in
velocity at higher temperatures. Thus Krogh points out
that the velocity curve is a straight line only between 7

No. 614] PHYSIOLOGICAL PROBLEMS 135

and 21 degrees, while the limit of development is from
less than 3 to 22.1.

Thus comparing the velocity curves for Headlee’s de-
velopment of Toxoptera (Fig. 1) and for the cleavage of
the frog’s egg, we note that in the case of the frog’s egg
the velocity is too great at the lower temperatures and
falls off at the highest temperature. Also, in the case of
Headlee’s curve for Toxoptera, the development was
much too slow at the higher temperatures. Krogh (’14)
further studied the development of pupe of Tenebrio
molitor, carefully measuring the carbon dioxide given off.
He found that the curve of velocity was a straight line
between 18.5 and 28 degrees, but that it curved upward at
lower temperatures. He tried incubating the pupe at
13.45 degrees, which is the mathematical zero of his curve,
and found that they developed in 1,116 hours, but with
considerable mortality. At approximately 33 degrees
the velocity was less than it should be if the curve were a
true hyperbola. An interesting feature of these curves
is that they approach so nearly to the curve published by
Verworn showing the stimulation effect of heat on activ-
ity. This curve is shown in Fig. 2 by the actual velocity
curve and the dotted extensions which, when compared
with the curve of Krogh for the development of Strongy-
locentrotus, Arbacia, and Tenebrio, indicate the close re-
lation between the amount and rate of activity and that
of general metabolism and growth.

Edwards ('02) made a careful study of the hen’s egg
and established 20-21? C. as the point at which no devel-
opment takes place. There is an optimum temperature
and development is accelerated by slightly higher tem-
peratures and retarded by lower temperatures. Thus
even in a warm-blooded species there is a point at and
below which development does not occur.

(b) Prediction on the Basis of Temperature Laws

Can we predict the time of appearance of any stage in
the life cycle of an animal? Certainly, in so far as we

136 THE AMERICAN NATURALIST [ Vou. LIL

are concerned only with temperature and with tempera-
tures within the straight-line limits of the velocity curve,
we can predict with a high degree of accuracy the time at
which any stage will be reached. Further, within the
2 10

0 12 1 13

“Pig. 3.

Fic. 3. Showing the total temperature curve for the oe of the first
cleavage furrow in the egg of the European frog, as a continuous line from the
data of Krogh. The small cross dashes show the te ies > which e
ments were performed. The actual temperatures are given at the bottom of

gure; at the top the actual degrees above the theoretical threshold of develop-
«degr The at th

ment, which is 2.7 Ç. Saree the left show the total minute-
degr time x temperature, which range from zero 3200. be
righ’ a different degrees of light; figures are added to
illustrate a method hart-making only. 10 units of light are assumed
give quickest development, and both increases and epoca to wt slower
development and hence more e ge sbeebs r

pion
in the same manner as it slowed general at during an entire month
in the experiments of Yung. For further explanation see

No. 614] PHYSIOLOGICAL PROBLEMS 137

straight-line limits the effects of constant and variable
temperatures should be the same. This is due to the fact
that the product of time units X temperature above the
threshold of development is a constant within the straight
line limits. Where it is not a constant, the actual values
may be plotted approximately for any temperature.
Using the data of Krogh (Fig. 3), I have drawn an
approximate total temperature curve for the development
of the first cleavage plane for the egg of the frog. The
number of degree-minutes required for completion of the
cleavage furrow is the same for all temperatures between
7° and 21° C. That is time X temperature is constant
between 7° and 21° C., where it is about 2,475 time-
temperature units or minute-degrees, and the curve is a
straight line. Above 21 degrees the total temperature is
greater than the constant, and below the lower limit of
the constant it is less than the constant. At 2.7 degrees
it should be infinity if the hyperbola held good, but is
actually 1,844 minutes. The time-temperature units are
not expressible at this point, so the actual time is given.
If development takes place below the zero of the hyper-
bola, the time-temperature units may be considered as
having a negative value, but are expressible. From this
curve it is possible to tell how long it takes for the cleav-
age furrow to develop at any temperature shown; for ex-
ample, take 6 degrees (bottom of chart=3.3 degrees at
top). We find from the curve that the total temperature
for this is approximately 2,200 degree-minutes. Thus,
2,200 divided by 6.0 — 2.7 gives 666 minutes. It is true that
the same result could be obtained by reading off the time
on a time-temperature curve (near to hyperbola) with
less labor, but the region in which the total temperature
is a constant cannot be shown on such a curve; and the
time for different temperatures is obtained with less sim-
ple calculations from the reciprocal. The total tempera-
ture curve exaggerates the straight-line limits, and brings
out sharply the fact that high temperatures retard and
low temperatures accelerate as compared with the veloci-

138 THE AMERICAN NATURALIST [ Von. LIL

ties indicated by the reciprocal of the hyperbola to which
the data partially conform.

Factors other than temperature influence the rate of
development. The work of Yung showed that in the case
of the frog light is one of these. Unfortunately the light
was not measured definitely in the work of either Krogh
or Yung. Yung kept one lot of developing frogs in the
dark and one in a window but where the sun actually
never shone on them. Krogh’s work must have been done
in similar light. Yung’s larve were reared under the
light conditions which he used, for a month or two months,
and thus his data are for older stages than those of Krogh,
whose results relate to the appearance of the first cleavage
furrow. Accordingly, any comparison of the two sets of
data is essentially impossible. However, for the purpose
of illustrating a principle which is indicated relative to de-
velopment under the influence of various intensities of
factors other than temperature, I have called the light con-
dition under which Krogh’s work was done 10 units and
have shown it on a scale at the right-hand side of the
graph. It is probable that too strong light will retard
development as well as too weak light. Hence the scale
is shown double, 12-8, 14-6, etc.; either increases or de-
creases in light intensity are assumed to increase the time
required for development. The cross shown on the graph
gives the approximate total temperature for darkness in-
dicated by Yung’s work. This part of the chart is given
merely to indicate a method of chart making—of showing
the way in which variations of one other factor change the
number of time-temperature units required for develop-
ment.

For practical prediction such a curve must be drawn
for the shortest time for development at each tempera-
ture. This will be under optimum light, chemical, etc.,
conditions for the temperature concerned. In establish-
ing such a least-total temperature curve a few careful
determinations within the straight-line limits with other
factors optimum will suffice. Outside these limits the de-

No. 614] PHYSIOLOGICAL PROBLEMS 139

terminations must be more numerous and especial care
must be exercised to have the temperatures constant. In
determining the optimum light for different temperature
much more rapid progress can be made by running ex-
periments under at least three conditions of this factor
for each temperature. Deviations due to factors other
than temperature should be shown on such a chart prob-
ably in a manner indicated by the broken line on Fig. 3.
If the main curve is drawn for shortest time, all devia-
tions in light, ete., will increase the so-called total tem-
perature, and lines may be drawn for these conditions
above the main curve as the facts necessitate.

Much investigation will be necessary to determine the
corrections which must be made in determining mean tem-
peratures which must be derived from conditions in which
the temperature slowly rises and falls during several
hours of each day, within the ranges of temperature
where the velocity curve is not a straight line. Tempera-
tures outside the straight-line limits should not be mixed
with the temperatures of the straight line limits. These
outside temperatures must be considered or estimated in
terms of units sufficiently small to approach accuracy.
In the case of daily temperature fluctuations the tempera-
tures outside the straight-line limits must be considered
by hours, and suitable corrections made before they can
be included in the daily mean. The exact nature of this
correction will have to be determined by careful inves-
tigation.

(d) Humidity Threshold

The workers thus far cited have studied temperature
alone, intending in a general way to keep other factors
constant. There is undoubtedly a threshold of develop-
ment with reference to each factor which influences devel-
opment. Berger found that growth ceased in tenebrionid
larve fed on bran dried at 105 degrees, and that they
lived for months with a loss of weight; doubtless with a
very small increase in moisture they could be maintained _
at the initial weight. More recently Pierce has found —

ES

140 THE AMERICAN NATURALIST [VoL. LIL

that the cotton boll-weevil has a different zero or threshold
of development and’ different design optimum for
each humidity.

(e) Oxygen Threshold

The development of various invertebrates is stopped by
insufficient oxygen (Loeb, ’06, and citations). Johansen
and Krogh found that if the oxygen pressure was reduced
to one half by reducing the air pressure to 380 mm. of
mercury development of plaice eggs was retarded. The
oxygen pressure threshold of development lies below the
amount which will go into solution from air at pressure
of 230 mm. of mercury, but at this concentration much
care was necessary to keep the eggs alive. Shull (711)
determined the oxygen minimum for the germination of
the seeds of Xanthium.

(f) Light Threshold

Loeb (711) states further that light is necessary to the
regeneration of zoids in Eudendrium. Its absence is
further known to slow development in larve of insects
which normally live in the light (Bachmetjew, 692). Smith
found that light accelerates the development of salmon.
Johansen and Krogh found little difference between ma-
rine fishes grown in light or in dark. Davenport (’99)
summarized the literature to that date and showed on the
authority of Yung that moderately strong light increased
growth.

(g) Food Threshold

Recent work has shown that food may be either qualita-

_ tively or quantitatively deficient and cause standstill in

the development of mammals. Thus Osborne and Men-
del (p. 101) show the following methods of producing it:

(1) By under-feeding with rations of suitable qualita-
tive make-up; (2) by the use of diets containing an ade-
quate protein but with inorganic salts supplied in the form
of a mixture of pure chemicals together with sucrose and

_ starch as the carbohydrate component; (3) by restricting

No. 614] PHYSIOLOGICAL PROBLEMS 141

the protein content of the dietary below the minimum re-
quired for growth; (4) by furnishing as the exclusive
source of nitrogenous intake proteins which lack some
amino-acid group indispensable to growth.

Thus the animals were maintained at practically the
same weight and they retained their power to grow long
past the age at which growth normally ceases (335 days)
and for periods equal to half the normal life of the spe-
cies, which is 1,000 days.

Wodsedalek (717) has shown that certain tenebrionid
larve can not only be maintained, but may be reduced
from half-grown to hatching size several times by re-
peated starving and feeding. This seems to leave little
doubt as to the existence of a threshold of development
for food.

(h) Definite Amount of Development

Krogh has shown that the total amount of carbon diox-
ide given off by pupe of Tenebrio molitor is the same for
all temperatures, showing that there is a definite amount
of development to be attained. The rate appears to be
different for different species where no considerable dif-
ference in the total for passing the stage in question is to
be expected, as in the case of fishes (see graphs by
Krogh). Thus, difference in velocity and increase in
velocity at different temperatures and moistures, etc.,
have an important bearing on the variable or unequal sea-
sonal appearance of the different species. The accelera-
tion of development under conditions of factors near the
threshold is a further consideration (for a noteworthy
instance see Bachmetjew’s (’07) retabulation of Mer-
rifeld’s (’90) data) which leads to non-coincident appear-
ance and peculiar modification of normal sequence in ab-
normal seasons.

It appears that the chief reason that there are not more
generations in an annual cycle in the case of spiders or
other animals is that the amount of energy which must
be expended and the velocity of development are such

142 THE AMERICAN NATURALIST [Vote LIL

that the completed sexually mature individual can not
be produced oftener than usually obtains. There is, to
be sute, much evidence that the tendency to hibernate
is not very firmly established in some species and that
under stimulation animals may be induced to repro-
duce nearly continuously, at least for a number of gen-
erations. Cessation of development in any given case is
as much attributable to some factor falling below the
threshold of development as to heredity. The environ-
ment is extremely complex, and the number of factors
which may cause cessation of development and which have
been already established, are so numerous as to indicate
that the number is very much greater than is commonly
supposed, including temperature, moisture, light, oxygen,
evaporation, quantity of food, or absence of any one of
many necessary food constituents. These appear to
operate in accordance with the law of toleration (Shel-
ford, 13) and, with respect to food, in accord with Leibig’s
law of minimum. Where dormant periods are well estab-
lished, their occurrence with reference to the usual sea-
sonal rhythm makes any modification of the usual life his-
tory difficult or impossible.
Variations from the ‘‘normal’’ seasonal weather, and
weather changes are of especial interest as modifying the
usual seasonal succession of adult animals or any area.
In springs with unusually prolonged cool weather, the
various pond species, such, for example, as those noted
on page 146, are crowded together, and reach maturity
much more nearly at the same time than in normal sea-
sons. The same phenomenon has been observed by the
writer in the case of the flowering of early spring plants
of an area near Chicago. The differences in the response
of different species to the same conditions show their dif-
ferent physiological constitutions. This type of varia-
tion indicates that such maladjustments as resulted in the
depletion of the grain crop by the grain aphis in the
southern part of the wheat belt, because the weather fa-
vored them, may occur in undisturbed localities, though

No. 614] PHYSIOLOGICAL PROBLEMS 143

probably not to the same degree. Seasonal succession
and its variation involve, for the pure-science student,
many of the problems which confront the economic
zoologist.

3. Length of Life and Fecundity.—One phenomenon
which has been repeatedly noted in connection with this
study—a matter of common observation—is the variation
in numbers of individuals in different years. The length
of life of individuals may have a pronounced effect on the
population and succession of species on a given area.
Loeb has stated that the great number of individuals in
the plankton of the polar seas in, summer is due to the
longer life of the individual at low temperature. Unless
the low temperature slows the different processes un-
equally this can hardly follow. For example, if a par-
thenogenetic female aphid normally lives a week and pro-
duces 1,000 offspring and then the temperature is lowered
so as to prolong the life to three weeks, unless the differ-
ent functions were unequally affected by the change, there
would be at the end of three weeks but a thousand, while
at the normal rate there would have been a billion possi-
ble individuals. On the other hand, if the rate of repro-
duction remains the same and the length of life of the
individual after the reproductive period is increased, the
results of lower temperature would be very different,
perhaps much as Loeb assumes. Actual observations
along this line are few. In the case of the San José scale,
however, Glenn (*15) found that the number of offspring
is greatest in the individuals breeding in the warmest
weather. Turning to Table I we note (page 146) that
Agelena nevea may live longer in the adult stage than
Argiope aurantia, or the time of appearance may be more
irregular, and hence the question is one for investigation.

The velocity of development of different species is dif-
ferent, and the relative velocity is measurable in some
terms of the angle which the velocity curve makes with
the axis of abscissas (Fig. 2). Thus when we compare
the four species of fish given by Dannevig we note that

144 THE AMERICAN NATURALIST [ Von. LIT

. velocity of development increases more rapidly with in-

creases of temperature for the flounder than for the
plaice; the same difference exists between the whiting
and the cod. Krogh showed that the velocities of the dif-
ferent stages of the frog’s egg, Fig. 1, are the same; but.
the different stages in the life history of the same animal
may differ in velocity at the same temperature.

4. Dormancy.— Dormancy is of much import among
animals inhabiting the same area. Thus the eggs of Eu-
branchipus and Diaptomus stagnalis require both summer
drying and winter freezing before they will hatch. Dor-

mancy is common in the eggs of grasshoppers (Thomas,

79), walking sticks (Trouvelot), ete. Dormant periods
are common, occurring even in deer and armadillo em-
bryos (Palemon), and probably represent hereditary
remnants of impressions made on former generations by
seasonal rhythms.

The causes of these rhythms often are simple. Con-
cerning delayed germination or dormancy of seeds,
Crocker and Davis (’14) have said:

The work to date has shown that delayed germination of seeds is
secured in a variety of ways: by almost absolute exclusion of water
by seed coats (as in the hard-seeded legumes and species of several
other families), by the limiting of the degree of swelling of the embryo,

+ - by reduction of oxygen supply below the minimum for germina-
tion . . .; and finally perhaps by deficiency in salts. To this must be
added delays due to embryo characters.

Dormancy has been overcome by drying in the case of
several species of insects in the writer’s laboratory.

III. SEASONAL SUCCESSION as ÍLLUSTRATED BY THE SPIDERS
OF A SMALL ÁREA OF Grounp

In the spring and summer of 1910, Mr. G. D. Allen un-
dertook the study of the seasonal succession of the fauna
of an area in a vacant lot at Eighty-first Street and Black-
stone Avenue, Chicago, which is a pond in spring and low
prairie in summer, but did not complete the work, though
his collections were extensive and oa extending

No. 614] PHYSIOLOGICAL PROBLEMS ` 145

from the middle of June to November. Miss Katherine
Norcross arranged the records in seasonal order, except
those of the spiders. In the case of the insects which
made up the vast majority of species on such an area, the
question constantly arose as to where the insect had been
previous to its appearance there. During the spring and
summer of 1913, the writer undertook to collect and ob-
serve the spiders of the plot studied by Allen. Spiders
were selected for this study because they do not undergo
a metamorphosis, and may often be found and identified
in a juvenile condition while insects can not. Though in-
complete, the data are adequate for a discussion of the
physiological features of seasonal succession.

The habitat from which the specimens were collected
was about 25 X 50 ft., nearly all of it covered with water
in early spring, usually drying during May, and contain-
ing water thereafter only during and after especially
heavy rains. In July Allen found the vegetation com-
posed chiefly of Eleocharis, Spartina, Carex, Juncus, Lia-
tris, Steironema, Cacalia, and several other composites.
The plants taken together made up what is commonly
called coarse grass and weeds. The writers collections
in 1913 were made on or very near Allen’s dates for 1910.
From these joint sources the data of the following table
were obtained and arranged, but with some gaps where
the spiders were probably too young to identify. The
records marked **C”” are taken from Comstock (*11) and
represent the conditions in which the spiders usually are
at the dates indicated. The spiders were identified by
Banks (*10) and the nomenclature is according to his list.

1. Statement of Succession.—In the spring the area is
a pond in which various Crustacea and worms succeed
each other (see Shelford, 713, pp. 278). Sexually mature
adults appeared in abundance about as follows: Ambly-
stoma tigrinum, March 15; Eubranchipus, April 15; Pla-
naria velata, May 1; Diaptomus stagnalis, May 1. Some
of these animals have been studied sufficiently to show
that they become dormant for the remainder of the year

146 -THE AMERICAN NATURALIST [ Von. LIL

as soon as the pond dries up. Amblystoma tigrinum de-
posits its eggs and then burrows into the mud and re-
mains ten months in estivation and hibernation. Eu-
branchipus deposits eggs that must be both dried and
frozen before hatching. Diaptomus stagnalis is similar
in character. Planaria velata forms cysts which live over
to the following spring.
ABLE I

SHOWING SEASONAL SUCCESSION OF ADULT SPIDERS ON A LOW PRAIRIE Sum-
MER Dry POND

The species are arranged in the order of the seasonal occurrence of adults.

indicates adults; j, young; e, eggs; g, generic identification only. © in-
dicates that the occurrence is As to Comstock and is not based upon
the author’s observation. Dates are given at the heads of the columns in
numerals only.

gielalsis e| oe|3|8 e 3
JERR ERERERRES
| 1330
ore rs ije | 7| 8| o|10/11| “6

Pardosa modica Blck......... AETR | |
Tetr tha laboriosa Htz mr. |. i ae op LAB
Xysticus gulosus Key ........ BA Ras E E JA ela Ple AL coy ant
Dictyna sublata Htz.......... saat Wi: BAE Pat oe A ieee
ina undata Htz........ Me ea oe YFSF oa eae at 300
Attus palustris Peck ......... Fk tans Geen be A Lee Rs ow Veet’
Pardosa canadensis Blek...... EEE A a RO, o a
Lycosa heluo Wal. ........... e OM nay tod Mad ah gt Sr tl
Phidippus podagrosus Htz FA E j 200
Plectana stellata Htz. ........!... |. |ë Ke 25
Epeira trivittata Key......... a es eat LA RE N Baa
“Runcinia aleatoria Htz. ...... De Ta | FN, 160
Mangora gibberosa Htz. ...... Pek RE TE been ten Le pal ceases
Agelena nivéa Wal. .......... weitere £18 STP Le lee ee
Aiweamesoks bhdr Hiss Ou ORC Ps | ea Ae TS
Phidippus audaz Htz......... ¡OC |jO |jC|jOl|jO|jC! * |* \*7C\gGC jc! 80
Argiope aurantia Luc......... SOIC FC LC VIC Ie Pee A TIE es
Argiope trifasciata Forsk...... ii G1 G1 519) * 1) * | > beC i0425

i

At the time these appear, land animals begin to move
about the pond margin, adult and juvenile spiders among

em. The collection and arrangement of the entire
fauna showed the same thing as the spiders, but proved
much less satisfactory in the other cases than was ex-
pected, owing to a lack of knowledge of life histories and
an inability to identify young stages. Turning to Table
I and following out the stars which indicate the occur-

No. 614] PHYSIOLOGICAL PROBLEMS 147

rence of adults, and noting that the species have been ar-
ranged in the seasonal order, starting with Pardosa mod-
ica Beck, which was taken only in April, we end the
season with adult Argiope trifasciata, which appeared as
adults late in the season only.

We note that when the collection proved at all complete
the juvenile individuals follow the adults of the early
spring species, and that they both precede and follow the
species which mature late in the season. The collections
proved to have been made with insufficient detail, and
many young spiders could be identified only to the genus
and are usually omitted entirely. However, the tables
show a sufficient general arrangement of the species
throughout a season to furnish an adequate basis for a
discussion of the problems involved in the phenomenon
of seasonal succession—the problems presented by a com-
parison of the few species whose life histories are known
quite completely.

IV. Discussion

Nearly all species are adjusted to the seasonal rhythm
of the habitat in which they live. Thus Dyctina sublata
appears as adult in May and June, when, as it seems, eggs
must be laid, and juvenile forms characterize the late
summer and autumn. Argiope trifasciata deposit eggs in
October and passes the winter in the juvenile form. Phi-
dippus podagrosus reaches maturity in July, when eggs
must be deposited, and young occur in both fall and
spring. These differences generally represent an innate
adjustment of the life cycle to seasonal rhythm, not read-
ily broken up. It is to be expected, then, that Dictyna
will deposit eggs to better advantage and that the young
hatch better in May than in November, as is the case of
Agelena nivea. It is further to be expected that the
young stages of some spiders will not go on with develop-
ment until cooled for a considerable period. Perhaps one
of the most interesting questions concerning the whole
matter of succession of spiders is to be found in the fact
that from what is known about them, they are all active

148 THE AMERICAN NATURALIST [ Vou. LIT

for about the same period of time; i. e., all life histories
involve about the same period of activity and rest.

An inspection of the table shows that the time of reach-
ing the adult stage varies for the different species, so that
there is a general change of spiders in the adult stage as
the season progresses. This is all that seasonal succes-
sion can mean under any conditions; the fact that the eggs
or other young stages can not be identified or their loca-
tion is unknown does not change the character of phenom-
enon in any locality where the species are resident.

The causes of the succession of species may be roughly
summarized as follows: Species differ in the time in the
annual climatic rhythm at which development begins, in
the time of occurrence of dormancy and in the conditions
necessary to break it up, in threshold of development rela-
tive to several climatic factors, in velocity of development
relative to several climatic factors, food, etc., and in size
and total energy expended. These may be taken up one
at a time.

Considering differences in the time in the annual cycle
at which development begins, as a factor in seasonal suc-
cession, we must notice first that this can be a controlling
factor only where there is no dormancy in the life history
or where the available total of temperature, moisture,
light, etc., above the thresholds is just enough to produce
one generation per year and not to permit of a gradual
moving of the time of appearance to an earlier date each
season, during several successive long seasons. The test
of this would come in the migration of agricultural pests
which are arrivals in localities where the growing season
is longer. There appear to be no easily available facts,
and. for the present this type of maintaining a definite
time of appearance is to be regarded as a theoretical pos-
sibility. The fact that the life histories of various ani-
mals which have been known to migrate extensively into
new territory appear not to be accelerated indicates that
dormancy may control appearance and thus time of be-
ginning development may be a secondary consideration.

No. 614] PHYSIOLOGICAL PROBLEMS 149

Thus we come to the time of occurrence of dormancy
and the conditions necessary to break it up, which result
in the rhythmic tendency of the species fitting into the
rhythm of the climate in which it lives. In many insect
species it appears that drying may be substituted for
freezing. Such species may migrate into climates in
which there is a dry season, instead of a cold one, and
with a longer growing season, and continue with the usual
annual life-history rhythm. Under these conditions in
each growing season the development is stopped by dor-
mancy and proceeds no further until the drying breaks up
dormancy. The development of Eubranchipus, once ini-
tiated, proceeds until the mature individual has produced
eggs. Here dormancy stops all further progress until the
eggs are first dried and then frozen and warmed above
0°: C. Crustacea without dormant periods go on devel-
oping and produce several generations in one summer.
After the conditions necessary for the overcoming of the
dormancy have been fulfilled, or where there is no dor-
mancy, species differ in the threshold conditions for de-
velopment. The thresholds for development are hardly
the same for any two species in which thresholds have
been determined. Thus species will differ in the time at
which development is initiated in the spring. Further,
the increase in velocity with increase in temperature is
different for different species, as indicated by the differ-
ences in the angle which their velocity curves make with
the axis of abscissas (see Krogh, 714, velocity curves
of several species of fish, also Fig. 2). This fact alone
makes it possible for a given set of conditions out of the
ordinary to give a peculiar and irregular occurrence of the
different species of a community.

The total energy as illustrated by the CO, given off by
a species is the same for all conditions in which develop-
ment can occur at all, as shown by Krogh. It is probable,
accordingly, that the total energy expended in develop-
ment is different for each different species. This may
bear some relation to size and weight, though alcoholic

150 THE AMERICAN NATURALIST [ Vou. LIT

specimens of full-grown females of several species of
spiders were weighed and no conclusion could be drawn.
Hither the method of obtaining the data or the fact that
the spiders are all annual is the cause. Krogh found that
the velocity of development is the same at the same tem-
perature in the different stages of the frog, though the
thresholds are different. But there is no reason to as-
sume that this is true of other animals, especially where
there is a metamorphosis.

1. Conclusions.—The preceding pages indicate the in-
tricacy of the problems involved in explaining the sim-
plest life history of annual animals. The physiological
life histories of animals which have two or more genera-
tions per year, and of those whose life cycle extends over
more than one year, are still more difficult to deal with.
The problems involved have of late attracted the interest
of biologists generally, of geneticists, of economic ento-
mologists, of fish culturists, and others, and they consti-
tute a central group of problems for the ecologist. All
these various interests are being focused on the problems
of physiological life histories as the next step in the at-
tempt to advance the science of biology. In all these
lines, the day of the naturalist taxonomist as a central
figure is all but past, and the day of the naturalist physi-
ologist is at hand.

This interest has arisen in the various groups for dif-
ferent causes, but one of them is the variation which oe-
curs in the succession of species and their interaction in
different years, due to peculiar weather conditions. The
green bugs destroyed the wheat crop in 1907 because of
differences in thresholds of development of the aphid
pests and their enemies; the fruit growers do not spray
at the right time in many cases because the insect pests
do not appear at the usual time. This is not to be ered-
ited to the effects of one factor alone; as, for example,
enough work with temperature has been done to show
that, while it is important, the influence of other factors
is sufficient to make prediction on the basis of tempera-
ture alone quite unreliable.

No. 614] PHYSIOLOGICAL PROBLEMS 151

The animal geographer is interested in the same prob-
lems. We note that the animal community illustrated by
the spiders contained animals maturing at every season
of the year. There is a noticeable early spring or vernal
group which the geographer has assumed is montane in
origin (Adams, ’09); and the group of land species which
appears through the summer is traced into different sit-
uations according to specific affinities. It is evident that
successful species are those that fit into the seasonal
rhythm with respect to physical conditions, food, and nu-
merous other relations.

BIBLIOGRAPHY

No attempt to make this list of literature complete has been made; asd
` such papers as were actually used are cited.
Allee, W. C.
1911. Seasonal Succession in Old Forest Ponds. Trans. Ill. State
Acad. Sci., 4: 126-131.

1909. Die Bestimmung des Alters pelagisch lebender Fischeier. Mitt.
d. Deutschen Seefischerei- -Vereins, 25: 364-73.
1911. ao Vertritng der pelagischen Fisheier und larven in der
Beltsee und den angrenzenden Meeresteilen. Wissensch.
Sey AEN N. F., Kiel, 13: 227-280.
Aron, H.
1911. dá ae and Growth. Phillip. Jour. Sci., 6: 1-52.
Babcock, 8.
1912, Meat Water: its Production and Róle in der Phenomena.
arch Bull. 22, Univ. Wis. Agr. Exper.
Bachmetjew, P.
1907. Pipeeimentette Entomologische Studien, Bd. II. Sophia.
Banks, Nathan.
1910. Catalogue of Nearctic Spiders, U. S. N. M. Bull. 72.
Baumberger, J. |
1917. Studies in the Longevity of Insects, Ann, Ent. Soc. Am., T:
323-54. (Good bibliography.)
Berger, B.
1907. Ueber die a oe der Tenebrio Larven gege:
Austrocknung. Arch. te d. Gesammte Phys. (Pfliiger’s Arch. a
118: 607-612.
Bream, F. E
1890. se o ueber die Bryozoen des Wassers Siissen. Bib.
1., Bd. II, Heft 6, 83 ff.
Candolle, A.
1855. Géographie op poms Raisonée. Paris (Cites 1830 paper).

152 THE AMERICAN NATURALIST [VoL. LIT

Child, C. M.
1913. The Asexual Cycle in Planaria velata in y to Senescence
and Rejuvenescence. Biol. Bull., 25: 202.

1915. acid and Rejuvenescence. eae Chapters X and XI.
Comstock, J. H.

1911. The Spider Book. New York.
Crocker, W.

1906. Róle of Seed Coats in Delayed Germination. Bot. Gaz., 42:
265-290.

Crocker, Wm. and Davis, W. E.
1914. Delayed Germination in Seeds of Alisma plantago. Bot. Gaz.,
58: 285-321.

Dannevig, H
1894. The Influence of Temperature on the Development of the Eggs
of Fishes. ee Ann. Rept. of the Fisheries Board for Scot-
land, pp. 147-1
Dareste, C
1892. Earp: sur la production artificielle des monstruosités, ou
s tératogénie expérimentale. Second ed., p. 129, Paris.
oe C.
—*99. p A Morphology, Parts I and II.
ji: wt E.
1878. A Report on the History and Present Conditions of the Shore
Cod Fisheries. U. S. Fish Com. Rept., Pt. IV, 685-731
Edwards, C. L.
Physiological Zero and the Index of hg! ri of the Egg
of the Domestic Fowl. Am. Jour. Phys., : 35l-
bis S. A.
1916. A General Survey of the May-beetles ua a of Illinois.
Twenty-ninth Rep. State Ent., III, pp.
Glenn, P. A,
1915. The San José Scale. Twenty-eighth Rep. State Ent. TL,
87-106.
Green, Seth.
870. Trout Culture. Rochester, N. Y.
Greeley, A. W.
1901. On Analogy between the Effects of Loss of Water and Lower-
ing of Temperature. Am. Jour. Phys., 6, No. 2.
Harvey, L. H.
1908. Floral Succession in the Prairie Grass Formation of South
ota. Bot, Gaz., 46; 81-108, 277-298.
Headlee, T. J.
1917. Some Facts Relative to the Influence of woe Humidity
Moa Insect Metabolism. Jour. Ec. Ent., 10:
an

64, beeen upon Periodic Phenomena in Plants and Ani-
mal:

8, 1851-1859. 36th Cong., Ist Sess., Senate Ex, Docu-
ments, 2: Pt, 1,

No. 614] PHYSIOLOGICAL PROBLEMS 153

Johansen, A. C., and Krogh, A
1914 The ra of Raumpbraturs and Certain Other Factors on
the Rate of Development of Fishes. Conseil Internat. p.
l’expl. de x mer. Pub. de Circumstance, No. 68 (Copen-
en).

hag
Johnstone, J.
1908. Conditions of Life in the Sea. Cambridge.
Krogh, A.

1914, On the Influence of the Temperature on the Rate of Embryonic
Development. Zeit. f. Allg. Phys., 16: 163-177.
1914a. On the Rate of Development and CO, Production of podr
of Tenebrio molitor at Different Temperatures, Zeit. f. Al
Phys., 16: 178-190.
Leeuwenhoek
1719; Epistolae ad societatum. regiam Anglicam et alios illustres viros
seu continuatio mirandorum arcanorum naturae detectorum,
Lugdun Batav. (Fride Verworn.)
Loeb, J.
1906. Dynamics of Living Matter. New Yor
1912. The Mechanistic Conception of Life. Chicago.
Masterman, A. T.
1894, ee beg Rate of Growth of the Marine Food Fishes, 13th
Rept. of the Fisheries Board for Scotland, pp. 289-96.
Mendel, L. B.
1914. Viewpoints in the Study of Growth. Biochem. Bull., 3: 156-
176.
Merrifield, F.
1890. Systematic ie epee Experiments on Some mor in
All their Stages. Trans. Ent. Soc. Lond., 1890, 131-160.
Osborne, T. B., and Mendel, L. Be
1914. py a Growth and the Capacity to Grow. Jour. Biol.

Peairs, L.
1914. pc Relation of Temperature to Insect Development. Jour. Ec.
Ent., 7: 174-179.
Pierce, W. D.
1916. A New miei of the Relationships of Temperature aaa
‘Humidity Insect Development. Jour. Agr. Research, 5:
sprog, o
Reibisch, J.
1902. Ueber den Einfluss der Temperatur auf die entered be
Fischeiern. Wiss, Meeresuntersuch., N. F., Abt, Kiel, 6:

1908. Belntion of Temperature to the Hibernation of Insects, Jour.
nt., 1: 56-6:

é. : >
19084. Distribution of Insects. Jour. Ec. Ent., 1: 245-262
1910. Relation en eae pape to the Growth of Insects. Jod. Ec.
Ent., 3: 113-140.

Sanderson, E. D., end praia L. M.
1913. Relation of Temperature to Insect Life. Tech. Bull. N. H.

154 THE AMERICAN NATURALIST - [ Von. LIT

Shelford, V. E.
1911. Ecological Succession. I. Stream Fishes. Biol. Bull., 21: 8-35,
1913, Animal Communities in Temperate America. Chicago.
1915. Principles and Problems of Ecology as Illustrated by Animals.
(British) Jour. of Ecology, 3: 1-23
Semper, K. i ;
1881. Animal Life, pp. 174-177; 444, New York.
Shull, C. A
1911; ‘The ae Minimum and the Germination of Xanthiwm Seeds.
- Gaz., 52: 453-477.

po C. =
908. A’Comparative Study of the Temperature papas of the
Velocities of Various Physiological Actions. Am. Jou r. Phys.,
9-334,

1911. On the Meaning and Magnitude of Temperature Coefficients of
hysiological Processes. Am. Jour. Phys., 28: 167-175,
Good bibliography appended.

Thomas, ©
1878-79. Influence of Meteorological Conditions on the rey et cng
and Migration of Locusts. Second Rept. U. S. Ent.
P- 16.
Trouvelot, R
1866

(Note on the Eggs of the Walking-stick.) Proc. Boston Soc.,
11: 89.

Yung, E
187 a Histoire de l’influence des milieux physiques sur les êtres
vivants. Arch. de Zoologie Expérimentale et Générale, 7:
251-282
Verworn, Max,
1899. psy Physiology (translation by Lee). New York.
ati
1917, Five pa Starvation of Larve. Science, N. S., XLVI: 366-
367.

Williamson, Chas.
1910. Experiments to Show the Influence of Cold in Retarding the
Development of Eggs of psa Herring (Clupea hagengus L.),
i do (Gadus

Fisheries Seu of Scotland, 27th Rept. for

THE USES OF INSECT GALLS
MARGARET M. FAGAN

BrancH or Forest Insects, BUREAU OF ENTOMOLOGY,
ASHINGTON, D. C.

INTRODUCTION

Tis paper, which is a contribution from the Branch
of Forest Insects, Bureau of Entomology, is a summary of
an extensive study of the literature dealing with the uses
of insect galls. It was made primarily to obtain a his-
tory of the use of the Aleppo gall in the dyeing industry.
In the preparation of this paper I have been assisted by
Mr. S. A. Rohwer under whose direction the research of
the literature was made.

For centuries before the real origin of insect galls was
known, they were noted and given a place, like most other
vegetable substances, among remedies for diseases. The
ignorance of their origin gave rise to queer superstitions
and practises even among scholars, especially in the
Middle Ages, when they were gravely recorded as super-
natural growths and employed as a means of foretelling
the events of the coming year. The gall was supposed to
contain a maggot, a fly, or a spider, each of which be-
tokened some misfortune. If the inhabitant were a mag-
got the coming year would bring famine, if a fly, war, or
if a spider, pestilence. This belief was recorded and
practised for several centuries, even after the time of
Malpighi, who was the first in the Western World to dis-
cover and make known the true origin of insect galls.

The record of the practical use of galls has come down
from the old physicians and naturalists of Greece and
Rome. Their observations were confined chiefly to the
Aleppo gall and the Bedeguar of the rose, but an interest-
ing statement is found concerning two galls which were
used by the Greeks to burn without oil in their lamps.

155

156 THE AMERICAN NATURALIST [Vou. LIT

These were Cynips theophrastea and an undetermined
gall called by Pliny the black gall-nut.

Until very recently in all histories of drugs, tanning,
and dyeing, galls have been considered as of great im-
portance, and at the present time are among the most
valued ingredients of ink. The first use of galls was in
medicine and many besides those discussed below have
been listed as drugs.

Among the Cynipids is the gall of field cirsium pro-
duced by Cynips species (determined by Cuvier) which
was formerly considered, if merely carried in the pocket,
as a very efficacious remedy for hemorrhages. Others
merely mentioned are (Cynips quercus-terminalis) =
Biorhiza pallida, Cynips polycera, Cynips quercus-toze,
Dryophanta quercus-folii, Andricus fecundatrix, and the
undetermined galls called by Guibort, galle corniculée,
galle marmorine, and galle d'Istrie. This last, according
to Trimble, yields 24 per cent. tannic acid.

Besides the Cynipid galls there is a gall on Pistachia
khinjuk, called Gúl-i-pista, produced by Pemphigus pal-
lidus, which enters into the materia medica of India.
Two other Indian galls used in medicine are found. on
Tamarix. One, called Bara-Mai, occurs on Tamariz gal-
ica, and the other, Chota-Mai, occurs on Tamarix orien-
talis. Another Tamarix gall said to have been used by
the Egyptians in medicine is one called by them Cher-
samel, and by the Turks, according to Fockeu, Bazgendge.

To the Cynipids useful in tanning Kieffer has added
Cynips lignicola and Cynips hungarica; and to those used
in dyeing, (Cynips tinctoria-nostra) = Cynips infectoria
Hartig. Cynips quercus-petioli Linneus, according to
the Gardeners’ Chronicle, 1854, is also capable of form-
ing a strong black dye, )

Burton in his journeys in East Africa noted a gall-nut
which was used by the Somali women as the basis of their
tattooing dye. This gall has not been determined, but
the record is of interest as being the only one encoun-
tered of a savage people’s making use of galls.

No. 614] THE USES OF INSECT GALLS 157

Another undetermined gall is that called by Pomet
““Bazdyendge”” and described by him as a reddish gall
on a species of oak in Turkey, which was used with cochi-
neal and tartar to make a very fine scarlet.

So far as can be ascertained no American galls were
ever used for any practical purpose by the Indian (state-
ment of Dr. Hough, U. S. National Museum), and but
few by the white man. No interest appears to have been
taken in this phase of gall history in America until
Trimble, through his interest in the history of tannins,
took up the question of tannic acid in galls and analyzed
a few North American galls. He found that many of
these galls contained relatively large amounts of tannin.
He stated that there are more oaks in the United States
than in Europe which are available for tanning, and that
as the gall partakes of the character of its host-plant then
there must necessarily be more oak galls in this country
suitable for tanning. He also remarks that it is not
known that all species of oak yield the same tannin, there-
fore we may look for a variation in the properties and
composition of the tannins from different species.

Of the Cynipid galls examined by Trimble the richest |
in tannin is one from Texas, on Quercus virens, closely
resembling the Aleppo gall and containing 40 per cent.
tannic acid. This has been identified by Mr. S. A. Rohwer
as Disholcaspis cinerosa. Acraspis erinacei (determined
by L. O. Howard) was found to contain 17.89 per cent.
tannic acid and Disholcaspis globulus Fitch, 3.91 per cent.

_A Dipterous gall on Quercus alba, determined by L. O.
Howard as Cecidomyia or Diplosis species, contains, ac-
cording to Trimble, 9.24 per cent. tannic when air-dried,
and 31.68 per cent. when quickly dried by artificial heat
at 80 degrees.

A gall occurring on Rhus glabra, in many ways the
counterpart of the Chinese gall, was found by Trimble
to yield, when air-dried, 61.70 per cent. tannic acid which
is about 8.3 per cent. less than the Chinese galls yield and
about 3 per cent. less than the Aleppo. This has been

158 THE AMERICAN NATURALIST [ Vou. LII

identified by Mr. A. C. Baker as Melaphis rhois Fitch.

Besides the American galls suggested by Trimble as
being of possible use in the industries, a few have been
recorded as food.

The galls of Disholcaspis weldi (Beutenmiiller) which
occur on Quercus reticulata in Mexico were purchased at
a fruit stand in Mexico City.

Oak-apple galls produced by Cynips spp. are eaten by
school children, and some of them are said to be sweeter
than sugar.

The most important record on the use of American galls
is a note by Dr. A. D. Hopkins on a black oak gall pro-
duced possibly by Callirhytis sp. This gall, because of
its resemblance to wheat, is called ‘‘black oak wheat’?
and ‘‘wheat mass” (typographical error for mast!).
Specimens of this gall were received from Westcott, Mo.,
with the information that they were very abundant intl
had been fed to cattle, hogs, sheep, turkeys and chickens
all of which were fond of them and were getting fat on
them. These galls were also received from Texarkana,
Ark., where they were used to fatten hogs.

The following food analysis and report were made on
these galls in the old Dendro-Chemical Laboratory,
Bureau of Chemistry, U. S. Department of Agriculture,
under the direction of the late Dr. W. H. Krug:

Per Cent.

UR ed saa es be Is vasa ds eek Venue 12.24
AE iE O A ays vee us Sau, 3.37
Crude Eg o O e laces 9.34
PA NI Oo rok 8.56
hii AN 2.89
alabiratal (Siaroh,. CHOY) cien des dde 63.60

Relative food value = 93.43
Nutritive ratio 4

_ The relative food value is high and the nutritive ratio i is wide; showing
that this material is especially adapted for fattening animals.

1 Mast is used for nuts collectively, acorns, chestnuts, beechnuts, ete.,
- especially when used as food for hogs and other animals,

No. 614] THE USES OF INSECT GALLS 159

CyYNIPS GALLA-TINCTORIA Olivier

The gall of Cynips galle-tinctorié Olivier, known in
commerce as the Aleppo gall, Turkey gall, Levant gall,
gall-nut, gall of commerce and ink marble, is found in
eastern Europe, that is, in Hungary, Turkey and Greece,
and in western Asia, on Quercus egilops, Quercus infec-
toria, Quercus pedunculata and possibly Quercus humilis.
This gall as an article of commerce has had the longest
history, having been used from the time of the ancient
Greeks to the present; has been used for the greatest va-
riety of purposes; and has been considered as the richest
of all the galls known to the Western World.

Medicine.—The earliest use of this gall was in medicine
in which capacity it was known to the Greeks and to the
Romans. In Greece it was recorded as of medical value
by Hippocrates in the fifth century B.c. and then by Theo-
phrastus, third century B.c. Its use by the Romans was
treated at some length by Pliny, who stated that twenty-
three remedies were compounded of gall-nuts, and that
among the diseases for which they were used were ulcera-
tions of the mouth, affections of the gums and uvula,
malformed nails, hang nails, etc., and that for the relief
of toothache and burns the inner part of the gall should
be chewed.

From these early days until very recent times authors
of Materia Medica have included this gall-nut as a drug,
designating it as ‘‘the most powerful of vegetable strin-
- gents.” In modern times it has been used in Europe as
a cure for fevers and was especially popular in France
early in the eighteenth century. At that time Poupart in
Mem. Ac. Sci., 1702, made a report on it which proved it
to be of doubtful efficacy. Nevertheless its use was con-
tinued and as late as 1849 Pereira, in London, listed it as
useful for medical purposes, recommending it as a tonic
‘in intermittents, an astringent in hemorrhages, a chemical
antidote, a topical astringent and giving a list of six
medicines concocted from it. At the present time gall

160 THE AMERICAN NATURALIST [ Vou. LIT

products are found in the British Pharmacopeia as as-
tringent ointments and in the U. S. Pharmacopeia, 1916,
ninth revision, the Aleppo gall still appears as the source
of tannic acid and as the principal ingredient in the prep-
aration Unguentum galle. Itis now used only externally.

Ink.—In the manufacture of ink the Aleppo gall was
long considered as a necessary ingredient, especially
where a durable ink was required, as in court records.
In some places the law required that records be made
with ink compounded of gall-nuts.

This use of the gall is not of such ancient origin as the
medical use, for Pliny, who quotes the older authorities
on other matters, has made mention only of the ink com-
pounded of lampblack, which was used also by the
Chinese. Hoefer in his ‘‘Histoire de la Chimie’’ spoke
of an ink used in the third and fourth centuries a.D., com-
pounded of acid and metal solution but failed to say that
this acid was obtained from gall-nuts. The ink made
from gall apple was, however, well known to the monks
of the ninth and tenth centuries, who used it in copying
their manuscripts. An interesting reference to the ink
made from gall-nuts occurs in Scheffel’s ‘‘Ekkehard,’’ a
romance of the tenth century, in which the monk Ekkehard
says *. . . all ink comes from gall apples and all gall
apples from a wicked wasp’s sting.” Of course, this is
of interest only if the knowledge of the origin of the gall
apple were part of the experience of the tenth-century
monk and not supplied from the knowledge of the nine-
teenth-century author. As the search to clear up this
point would be long and arduous and the result of no real
value it has not been made.

From the ninth century down to the present day, gall-
nuts have been included in practically every good ink
recipe for black writing and record inks. The Aleppo
gall is considered as the best for ink-making, but other
important ones are the Morea gall, the Smyrna gall, Mar-
mora gall and Istrian gall, and other good quality galls

a

No. 614] THE USES OF INSECT GALLS ‘161

from France, Hungary, Italy, Senegal and Barbary. The
Chinese and Japanese galls are also sometimes mentioned
in recipes. The Japanese gall has been used in making
school and other cheap inks.

The Massachusetts Record Commission in 1891 made a
Report on Record Inks and Paper in which the superiority
of gall-nut ink was attested, The ink made from gall-
nuts was said to be permanent, if properly made, and to
have the advantage that if the writing should fade it
could be repeatedly restored by a solution of nut-gall or
tannin. Any other coloring matter substituted in whole
or in part for gall-nut and iron solution impairs the qual-
ity of the ink.

In 1912, Oyster in the ‘‘Spatula Ink Formulary”” gives
as the basis of the best black writing and record inks, gall-
nuts. In the recipes for inks used by the United States
Treasury, Bank of England, the German Chancellory,
and the Danish Government the Aleppo gall is specified.

Lehnen also states that nut-gall extract forms an ex-
cellent material for the preparation of ink, especially
where manufacturers can not keep large stocks of the
nut-gall itself. According to the 1917 annual report of
the Oil, Paint and Drug Reporter, large quantities of
gall-nut extracts are imported into the United States.
Of course, all of it may not be used for ink manufacture.

Tanning.—Among tanning materials this gall-nut is the
richest of all in the tanning principle and has been used,
for tanning purposes, in the preparation of hides and
skins, but because of its expense and its value to the tex-
tile colorist it has not been extensively used. Experi-
ments and analyses of these galls were undertaken, how-
ever, with a view to discovering the tanning principle
in vegetable matter. Pliny mentioned the preparation of
leather as being one of the uses of gall-nuts, Bose re-
corded galls as being used in the tanning of hides and
Davis in 1897 spoke of them as being the richest in tannin
of all tanning materials, but made no further mention of

162 THE AMERICAN NATURALIST [ Vou. LII

them in his descriptions of the processes of tanning.
Other more recent writers on tanning materials have also
listed oak galls,

Dyeing.—In the history of the art of dyeing, the Aleppo
gall figures largely from the earliest mention of the art
in literature up to the very present. According to Theo-
phrastus it was used by the Greeks in dyeing wool and
woolen goods and Pliny mentioned it as being used to
stain the hair black and as the best adapted for the prep-
aration of leather and the dyeing of skins. As the an-.
cients could not conceive of a scholar’s taking an active
interest in the technical arts there is no record of how
these galls were used, merely the statements that they
were so used; and it was not until the end of the eighteenth
century that any definite knowledge of these galls was
sought.

It was at that time, when science invaded nearly every
field of endeavor, that the chemists made an earnest effort
to determine the chemical contents and action of gall-
nuts, so as to place the arts of dyeing and tanning on a
firmer and more scientific basis. Déyeux in 1793 was the
first to separate the tannin in the gall-nuts and his ex-
periments were followed up by Scheele, to whom is ac-
credited the discovery of gallic acid. Berthollet and
Fourcroy made more detailed analyses of the gall-nuts
and gave more positive knowledge of the various proper-
ties and their chemical value.

Berthollet in ‘‘The Elements of the Art of Dyeing””
gave perhaps the first scientific account of the art of dye-
ing with full explanations of methods and materials. Ac-
cording to his idea the great value of the Aleppo gall lay
in its astringency and as it is most astringent before the
insect escapes, the immature galls, or as they are called,
the blue galls, are of the most value and are the ones used
in dyeing black, while the white galls, or those from which
the insect has escaped, are used in dyeing light linens.
For the dyeing of black Berthollet considered that no

No. 614] THE USES OF INSECT GALLS 163

other astringent than an infusion of gall-nuts could be
used in the dyeing bath, as too large a quantity of any
other material would be necessary to obtain the same
results.

Bancroft, however, in his ‘‘Philosophy of Permanent
Color,’’ 1813, opposed the idea that the astringency was
the important property of the gall-nut and set forth the
idea that it should be considered merely as a coloring
matter. In defense of his theory he showed that tannin
procured from different vegetable matter and combined
with iron will not produce black, and gallice acid alone will
not blacken solutions of iron, while either tannin or gallic
acid from galls combined with iron forms a black dye
or ink.

At the present time both these theories are known to be
true for the Aleppo gall may be used as a fixing agent in
dyeing or may be used as the basis of a good black dye.
As a dye its use appears to be confined to the dyeing of
leather and of sealskin fur.

In the dyeing of leathers and skins the Aleppo gall is
used in small quantities with other dyeing materials to
obtain the best and most permanent black. That the suc-
cessful dyeing of leathers, however, is not dependent en-
tirely upon a good dye is evident from the following state-
ment on leather dyeing by Bennett, ‘‘Manufacture of
Leather,’’ 1909:

The absorption of the dye by the fiber has been considered a case of
chemical action, of physical action and even as a case of “solid solu-
tion,” but it is ibi probable that more than one type of action comes
into play and that possibly all these theories may be true to a certain
extent. It would, however, appear that with vegetable tannages the
determining factor is the formation of color lakes with the tannin on the
fiber. The tannins are of an acid nature and fix the basic dyes with
great readiness, but the basic chrome-tanned leathers fix the basic dyes
much less readily than the acid dyes, so it is clear that the nature of the
tannage has considerable influence in the matter.

For the dressing and dyeing of sealskin furs, large
quantities of the Aleppo galls were formerly shipped to

164 THE AMERICAN NATURALIST [ Von. LIT

London, where all of our American sealskins were dressed
and dyed for the market. Now, however, this industry
has been established in the United States, and in 1914
gall-nuts worth $17,174 were imported from Bagdad for
this purpose. As the method of dyeing sealskins is a
very jealously guarded trade secret the American firm
engaged in this enterprise has had to work out its own
processes, and according to the Commerce Reports this
has been successfully accomplished and one sale of Amer-
ican sealskins dressed and dyed in America has taken
place, in St. Louis, in September, 1916.

- Analysis.—A: gall so widely known and of such great
value has of course been analyzed many times and is the
standard for the analysis of others. According to Trimble
the most generally accepted analysis is that made by
Guibort, which is as follows:

Per Cent.

AS Te MMR ean i a ds mola bli Uat E

CIMA GORE oe) oss y ee cee ew Sena eos eres 2

Ellagic acid >

Jem Pio das

Chlorophyll and volatile oll .......ocomommomomonso 0.7

Pronn ACTO. MAOT «re. be ey EUs uae + bees 2.5

CE o ev eee een ween 2.5

E sos oe tee ee. VEE T Leet 2

Woods A A on a 10.5

Sugar 7 Q

Albumen

Potassium sulphate

Potassium gallate L............. Sy ere eee ee 1.3

Gallate of lime

Oxalate of lime

Phosphate

DÍLAR as be eee as n.5
100.00

Cynips Insana Westwood
A gall somewhat resembling the Aleppo gall and often
confused with it is that produced by Cynips insana West-
wood. It is better known as the mad apple of Sodom,
Dead Sea fruit, or Mecca or Bussorah gall, and is found
in Palestine, Asia Minor, Albania and Italy on Quercus

No. 614] THE USES OF INSECT GALLS 165

infectoria, Quercus tauricola and Quercus farnetto (con-
ferta). Its use is confined to the locality from which it
takes one of its names, Bussorah or Basra.

This gall has furnished an interesting and somewhat
mystifying theme to poets and has been often discussed
by old writers who have tried to connect this so-called
fruit with some of the unknown fruits mentioned in the
Bible.

In Bussorah or Basra in Asia Minor, probably its native
heath, this gall is used by the inhabitants in dyeing Tur-
key red, and it is more esteemed by them than the Aleppo
gall.

Analysis.—The following analysis of this gall was made
by Bley in 1853:

Per Cent.

Pantie Aoi led lice pias wie ture us ture gkko 26.00
GRE ded or ee A ee 1.60
VAL OU ec, aN Oe ee ee ee 0.60
BOR is AA ee Le IEA ee Oe 3.40
Extractive and sat Ce hel stoke re ee Ee! Bing 2.00
EEE ep ERG o oe Ae O eae aie de 8.40
TWO ORS A pale di ee ches Cue Cs ee 46.00
Montte Ls ie AR ee Se Pei 12.00
100.00

CYNIPS QUERCUS-CALYCIS Burgsdorf

The knoppern or acorn gall, also called the Piedmon-
tese gall, and gall of Hungary, which is produced on
Quercus egilops, Quercus pedunculata and occasionally
on Quercus pubescens and Quercus sessiliflora, occurs in
Austria, Hungary, Slavonia, Bosnia, Serbia, Greece, Asia
Minor and less abundantly in Germany, Holland, France
and Italy. Among the Cynipid galls it ranks next in im-
portance to the Aleppo gall and has been almost as fre-

quently discussed.

In Austria it has been used chiefly by the tanners, but
has also been substituted by dyers for the Aleppo gall. |
This gall, like the Aleppo, is at its best before maturity

‘and should be collected from August to October.

166 THE AMERICAN NATURALIST [ Von. LIT

At its height the Knoppern yields from 45 to 50 per
cent. of tannin.

Cynips KOLLARI Hartig

The Devonshire gall is produced by Cynips kollari on
Quercus avellane-formis, Quercus fastigiata, Quercus
humilis, Quercus ilex, Quercus lusitanica and varieties,
Quercus mirbeckii, Quercus mongolica, Quercus peduncu-
lata and varieties, Quercus pubescens, Quercus pseudoegi-
lops, Quercus rubra, Quercus sessiliflora, Quercus suber
and Quercus toza. It occurs in middle and southern
Europe, North Africa, Asia Minor, and was introduced
into England from the continent early in the nineteenth
century. It attracted much attention because of its rapid
spread, but the interest in it seems to have been confined
to England, as no important reference to it has been found
elsewhere.

Attention was first drawn to this gall in England, when
it became so abundant that the extermination of the oak
forests seemed threatened by it. At that time, about
1858, many notices concerning it appeared in which fear
was expressed that it would do irreparable injury to the
oaks. But the damage done by it was of no great moment
and when the gall was studied it was found to have some
tanning and dyeing properties, and to be useful in making
an excellent ink.

Many analyses were made of this gall in which varying
amounts of tannin were accredited to it. The following
was made in 1869 by Watson Smith:

Per Cent

TIM MU TI ie, 26.71
Hale MOa A ol ea eee ee Traces only

IOS GE A e ea es cos ee 47.88

obaro ek eee A eee ees 20.61

Coloring and TOM... cent cc et 4.80

100.00

RHopITES ROSA Linneus
The Bedeguar of the rose, the gall produced by Rhodites
rose Linnæus, occurs throughout Europe and in western

No. 614] THE USES OF INSECT GALLS 167

Asia on eighteen species of Rosa, and in North America
on Rosa canina only. It was highly esteemed by the an-
cients, but has received very little mention in more mod-
ern times as being of any particular use to man.

This gall was mentioned by Pliny as being among the
most successful applications for the restoration of hair.
For this purpose it had to be powdered and mixed with
honey. In Italy it has been used, when powdered and
laid on the affected parts, to cure the bite of venomous
creatures. This use by the Italians may have grown out
of the story related by Pliny that the mother of one of
the pretorian guard had a dream, after her son had been
bitten by a mad dog, in which she was directed to procure
the little round balls at the root of the wild rose and apply
them to the affected part. Cuvier has recorded the Bede-
guar of the rose as among the remedies successfully used
against diarrhea, dysentery, and cases of stones, scurvy
and worms, and as late as 1868 the farmers near Harro-
gate were known to gather the mossy galls of the rose to
make an infusion for diarrhea in cows, for which they
claimed to find it very successful.

AuzLax sp. Riibsaamen

The gall of the sage or ‘‘ Pomme de sauge”’ is produced
by Aulas sp. Riibsaamen on Salvia pomifera and other
species of Salvia in the Isle of Crete.

The earliest available record of the use of these galls
is that by Belon in 1558 in which he described them as
being large as galls, covered with hair and sweet and
pleasant to the taste. They were collected at the be-
ginning of May and sold by the people of Candie to neigh-
boring villagers. Olivier stated that ‘‘they are esteemed
in the Levant for their aromatic and acid flavor, espe-
cially when prepared with honey and sugar, and form a
considerable article of commerce from Scio to Constanti-
nople, where they are regularly exposed in the market.’’

168 THE AMERICAN NATURALIST [VoL. LII

Fockeu in 1897 mentions having found these galls in
the East but states that to-day the old common name,
Baisonge, is unknown and that the people of the country
when questioned concerning them said that they had never
noticed their existence and expressed doubt of their ever
having been used for food, or in making confections.

This name ‘‘ Baisonge’’ was not used by either Belon or
Olivier for the gall of the sage, but has been used by
Cuvier to designate a terebinthe gall from Syria.

AULAX GLECHOMZ Linnzeus

Another Cynipid gall which has been used as food is
the gall of the ground ivy made by 4Aulax glechome on
Glechoma hederacea L. It occurs in Lorraine and
Sweden.

This gall was used in France as food and is said to have
an agreeable taste and the sweet odor of the host-plant.

CHINESE Oak GALL

An unidentified oak gall, said to closely resemble the
European gall, is one which was recorded in Pen T*Sau
as Woo-shih-tsze.

The following translation of the note concerning it has
been published by Pereira (Pharm. Journ., Vol. 3, 1844,
pp. 384-7):

Woo-shih-tsze also comes from the West, and from India. The tree
is said to be sixty or seventy eubits high and eight or nine cubits in
circumference, and grows in sandy and stony places. It is compared

. to the camphire tree. It flowers in the third moon; the flower is
white: and rather red in the center. The bud formed is round like a
ball; at first green—when ripe, yellowish. An insect eats into it and
forms a hole in it. They say that the tree one year produces the Woo-
shih-tsze, and another year produces something which resembles a
ehestnut.

Another name is Whi-ztsip-tsze. It has various medicinal properties.
‘It is used with some other ingredients for dyeing beards black.

The taste of the Woo-pei-tsze is, according to them, a sour, saltish
taste—of the Woo-shih-tsze, a bitter taste.

No. 614] THE USES OF INSECT GALLS 169

In the Materia Medica of China (Smith, 1871, p. 100)
it is called ‘‘food for the foodless’’ and is recommended
for medicinal use. It is said to differ little from those of
the European market and to have been used formerly in
making ink and in dyeing hair.

As this gall is described by the Chinese as coming from
the ‘‘ West’’ could it possibly be the Aleppo gall, the dis-
tribution of which is eastern Europe and western Asia?

SCHLECHTENDALIA CHINENSIS (Bell)

Besides the Cynipid galls many others have been re-
corded as of use to man. Most of them are merely in-
cluded in the native Materia Medica of China and India,
but a few have had other uses.

The most important of these galls is the Chinese gall or
Woo-pei-tsze, produced by Schlechtendalia chinensis on
Rhus semialata, in northern India, China and Japan. It
has been known and used by the Chinese for many cen-
turies, perhaps even longer than the Aleppo gall has been
known in the West. It rivals the Aleppo gall in im-
portance and like the latter is still an important article
of commerce.

The Chinese gall has been used in medicine, tanning,
and dyeing, and is now imported into Germany and the
United States for the manufacture of tannic acid, of
which it yields about 70 per cent. As this gall has been
fully treated in a paper by A. C. Baker, which has been
submitted for publication it is unnecessary to give details
here.

Cuermes sp. (Baker)

The gall identified by Kirby and Spence as that of
Aphis pini has been identified by Mr. A. C. Baker as
Chermes sp., possibly Chermes lapponicus Chol, possibly
some other. It occurs on spruce-fir in Lapland.

Linneus states that this gall was used as food, and

170 THE AMERICAN NATURALIST [ Vor. LIL

ha $ »
Kirby and Spence suggested it as a possible dyeing ma-
terial. Linnweus's description of it is as follows:
The extremities of the branches of the spruce-fir bear small yellow

cones. . . . When arrived at maturity they burst asunder and discharge
an orange-colored powder which stains the clothes of those who ap-

` proach the tree. I conceive these exerescences to be caused by some

minute insects. The common people eat them raw as a dainty, like
berries.

It was probably the reference to the orange-colored
powder staining the clothes which led Kirby and Spence
to suggest that this gall might be placed among dyeing
materials.

PEMPHIGUS CORNICULARIS

The gall of Pemphigus cornicularis, called in India,
Kakra-Singhee, in Syria, Baizonge, and in Europe, gall
of the terebinth, occurs in southern Europe and Turkey,
in Spain, Syria, China and India.

In India this gall is used in medicine by the natives who
assign to it great astringent and tonic properties.

The Hindus have also used it, to a limited extent, in
dyeing.

In Thrace and Macedonia Belon recorded it as being
collected at the end of June, while still immature, and
sold at high prices to the inhabitants of Bource, who used
it in coloring fine silks. In Spain, Syria and China it
was used as an ingredient in making scarlet dye.

ALDACAY GALLS

Galls called Aldacay or Caducay galls were recorded
by Roxburgh in 1805 as occurring on the leaves of Mimosa
arabica on the coast of Coromandel. Kirby and, Spence
in speaking of this gall called the host-plant Terminalia
citrina.

These galls were said to have been among the most
valued of the native dyeing materials and to have been
sold in every market. The natives dyed their best and

No. 614] THE USES OF INSECT GALLS 171

most durable yellow with them, and they were also used
by the chintz painters for their yellows. When mixed
with ferruginous mud a strong black dye was obtained.

The astringent properties of these galls were evidently
stronger than those of the fruit of the tree, as an ink made
from the galls resisted the weather longer than that
made from the fruit. —

Roxburgh did not identify these galls, but suggested
that they might be the ‘‘Faba bengalensis” of the old
Materia Medica writers. The ‘‘Faba bengalensis’’ ac-
cording to Bosc is the fruit of the Myrobolan citrin altered
in its form by the puncture of an insect, but no dyeing
properties are ascribed to it. As no further reference
to these galls has been found they are still undetermined.

SuMMARY

The important uses of galls have been in medicine, the
manufacture of ink, tanning and dyeing, with a few ref-
erences to their use as food, and one to their use as fuel.

In medicine the following galls have been used: An-
dricus fecundatrix Hartig, Cynips sp. Cuvier on field
cirsium, Cynips galle-tinctorie Olivier, Cynips polycera
Giraud, Cynips quercus-foliíi Linneus (Cynips quercus-
terminalis) = Biorhiza pallida Olivier, Cynips quercus-
toze Bose, Pemphigus cornicularis, Pemphigus pallidus,
Rhodites rose Linneus, Schlechtendalia chinensis (Bell)
and the undetermined ones: Chinese oak gall, Istrian gall,
Marmora gall, galle corniculée, and Bazgendge (Fockeu)
or Chersamel on Tamariz.

In the manufacture of ink the galls used are: Cynips
galle-tinctorie Olivier, Cynips kollari Hartig, Schlechten-
dalia chinensis (Bell), the Aldacay or Caducay galls, the
Istrian, Marmora, Morea and Smyrna galls and others
from France, Italy, Hungary, Senegal and Barbary.

For tanning the following have been found useful:
Cynips galle-tinctoria Olivier, Cynips hungarica Hartig,
Cynips insana Westwood, Cynips kollari Hartig, Cynips

172 THE AMERICAN NATURALIST [ Vou. LIT

lignicola Hartig, Cynips quercus-calycis Burgsdorf and
Schlechtendalia chinensis (Bell).

For use in dyeing have been recorded: Cynips galle-
tinctorie Olivier (Cymips tinctoria-nostra) = Cynips in-
fectoria Hartig, Cynips insana Westwood, Cynips kollari
Hartig, Cynips quercus-calycis Burgsdorf, Cynips quer-
cus-petioli Linneus, Pemphigus cornicularis, Schlechten-
dalia chinensis, possibly Chermes sp. Baker, and the un-
determined galls, Aldacay or Caducay galls, the gall-nut
used by the Somali women for a tattooing dye, Baizonge
Cuvier and Bazdyendge Pomet.

As food, only a few galls have been used: Aulax sp.
Rúbsaamen or Baisonge Fockeu, Aulas glechome, Cynips
spp. Girault, Disholcaspis weldi (Beutenmiiller), Cal-
lirhytis sp.? Hopkins, Chermes sp. Baker and Schlechten-
dalia chinensis. In the case of the last named the gall
itself is not eaten but the powder found on the outside is
used for seasoning soup.

As fuel for lamps the Greeks used Cynips theophrastea
and an undetermined gall called by Pliny the eae
nut.

Common Names or Insect GaLLs

In the following list of the common names of the insect
galls which have been of any practical use, will be found
a number of names for the gall of Cynips galle-tinctorie
and several for that of Schlechtendalia chinensis. To
avoid confusion I would suggest that for the former the
name Aleppo gall be adopted, and for the latter the
name Chinese gall, as it is under these names that they
are designated in the commerce reports, in some of the
trade journals and in the technical works on dyeing, tan-
ning and ink manufacture.

eoru alt, eo Cynips quercus-calycis Burgsdorf
Aloppo WAU a cla. cok ce nN Cynips galle-tinctoria Olivier
Baisonge Foekeu ..... avi.. o.. Aulaz sp. Rübsaamen

Haizonge OCavier 0 apiu.. Pemphigus cornicularis

Bara Mai ... : Hindu name of a gall on Tamariz gallica

t

No. 614] THE USES OF INSECT GALLS 173

Bazdyendge Pomet ...........ooo Turkish name for a gall on oak
Bazgendge Fockeu ............. ide name for a gall on Tamariz
entalis
Bedequar of the rose ........... odias rose nS ame
Black gall-nut Pliny ............ Undeter
DCR ORR WHERE n.n. RESTA, Canivnytie ts Hopkins
DUDA gal: pes i ae Cynips insana Westwood
Reread Sy ee ee. os. a Egyptian name for nese, jo (q.v.)
Chines palace A Schlechtendalia chinensis Bel
goleo (pear gallo ¿Ei oi pea pao ball
Gho MAP. A du a gall on Tamariz ori-
ntalis
WOME TOR SPO MOS ON, cee Culpa de insana Westwood
Devonshire... gal evi dd ssw spans Cynips kollari Hartig
Fruits for the foodless .......... A Chinese oak
Faba Dengatensta® cock. Es A gall on Terminalia chebula
prob pia ve Cynips galle-tinctorie Olivier
Gall a field eirsium ....o...o.oo. Cynips - Cuvier
Gall of wage A ee AO Aulax sp. Riibsaamen
Gall = Mimosa arabica Roxburgh . sae schol is
Gallapo -o i ova ks Rare ee are s galle- eia Olivier
Gai-nnt Borton oaeee Pc ithe e abl
Galle PLATO 00% Fa ok Cynips kollari
Palle’ OW APURADO S00 fo ee ndricus fecundatriz Hartig
ENA corniculée Guibort ........ Undeterminable
PRR oes ic eae bea ves ow emphigus pallidus
hort A A Cynips quercus-calycis ae age
TOE MANOS Ee ees Cynips mas gaara Olivi
IA UL a a SS ndeter
Japanese pall ii ri Schle chtendalia chinensis Bell
togha PO oo ee emphigus culari
eet ECT es Cue ee aes nips rcus- ous calycia ag
Levant palb o.oo. eH e Cynips galle-tinct
Mad pla Or Sodom La. o. Cynips insana Westwoo
Marmora ME E o Undetermina
ecca a ower ys na Westwood
NUM Fes ue ck, eed en ee ue es Cynips galle-tinctorie Olivier
Nntgal ERA PATS Schlechtendalia chinensis Bell
Oak aprio, uc rr Andricus quercus-californicus
Dulcapple galls i. ccs .Cynips spp. Girault
Oriental Gall A Cynips galle-tinctorie Olivier
Piedmontese gall ............... Cynips quercus-calycis Burgsdorf
omme de chene Guibort ........ Undeterminable
Small crown gall of Aleppo ...... Cynips polycera Giraud
Bomrät-úlanl Lu Arabic name for Chota Mai (q.v.)
a A ne name for Bara Mai (q.v.)
Turkey, gall ... ces coe beeen ene Cyn alle-tinctorie Olivi
Wheat mass [Mast] ...........- cary sp.? Hopkins

[Mast]
Whip-ztsip-tze and Woo-shih-tsze .. A Chinese gall on
Wü-peitssó ads bees eee okas chinensis (Bell)

174 THE AMERICAN NATURALIST [ Von. LIT

BIBLIOGRAPHY
The information embodied in the foregoing paper has been compiled only

from the sources listed below.

Bancroft, Edward. Experimental UE . » . Philosophy of Permanent
Colours. . . . London, 1813, Vol

Belon, Pierre. Les rabos ap. o Singularitez et Choses Me-
morables. . . . Paris, 1588

Bennett, Hugh G The Manufacture of Leather. London,

Bertholett, C. L., and A. B. Elements of the Art of Dyeing oo. by
Andrew ies), 2d edition, London, 1824

eae Wm. Two New Species of H: olcaspis from Mexico. Psyche,
Vol. 18, 1911, p. 86.

Buignet, e Sur les Noix de Galles Anglaises, g ra Vinen. Jour. Pharm.
et Chimie, 3d Ser., Vol. 30. Paris, 1856, pp. 290-291.

: ¿uegsdort, FLO. L. Phystkalisch-Gkondinis che re von den Verschiedenen

oppern. Schrift. Berlin Ges. naturf. Freunde, Vol. 4, 1783, pp. 1-12.

p
Burton, R. F. First Footsteps in East Africa. London, 1856, p. 27 (foot-

note).

Connold, E. T. British Oak Gall. London, 1908,

Cosens, A. The Founding of the Science of Cecidology. 46th Ann. Rept.
Ent. Soc. Ontario for 1915, pp. 88-93.

we Frank. Curious Facts in the History of Insects. London, 1865, pp.

146.
oe hee A Jagstiso of the Materia Medica. Philadelphia, 1808, Vol.
2,
Cuvier, a Régne A itsiiar’ sar, p. 411.
Cuvier, G. The Animal Kingd m (with IEA additions by E.
riffith). London, 1832, el 15, pp. 423-32.
Davis, a T. The Manufacture of Leather. Philadelphia, 1897, p. 29, and

dela Ww. Physico-Theology. London, 1798, Vol. 2, pp. 325-32.
Deyeux, C. Mémoire sur la noix de Galle. Ann. Chimie, Paris, 1793, pp.

Dupont, Torin, L’industrie des Matières Colorantes. ... Paris, 1902,

pp. 327-8.
/Eneyelopedia Britannica, 9th edition, 1882. Article on galls.

Fockeu, H. Etude sur Quelques ie sig de Syrie. Rev. Biol. Nord de la
Frane
eee H. Sur Quel anne Sri AA Rev. Gén. Bot., Paris, Vol.

Freema ie and Chandler. a. ale ’s Commercial Products. Boston, p. 351.
Gardeners’ Chronicle, eet eles p- 742,
Geoffroy, seg sees ca A T: ase! Pomel, Vegetable and Animai Sub-
stances ... in Pip panes from MS. copy by G. Douglas).
eae 1736, pp. 368-9,
, A. A. Fragments of North American Insects, III. Ent. News,
ve 24, 1913, p. 60.

No. 614] THE USES OF INSECT GALLS 175

ence M. On the Galls of Terebinthus and Pistacia. Pharm. Journ. ai
. 3, London, 1844, pp. 377-382.
car M. Histoire Naturelle des Drogues Simple. Paris, 1849, 5th ed.,

Hill, aon J. r Decade of Curious ea geese 1778, p. 15:

Hoefer, F. Histoire de la chimie. Paris, 1866, p.

Hopkins, A. D. Proc. Ent. Soc. Wash., Vor 5, ae p. 151.

Hummel, J. J. Dyeing of Textile Fabrics. London, 1909.

The Insects of Commerce. Leisure Hour, London, Vol. "4 1854, pp. 437-40.
Johnson’s Encyclopedia. New York, 1893.

Kakra-Singhee. Science Gossip. London, 1865, p

Kieffer, T. J. Die Insekten Mitteleuropas. . ae 1914, Vol. 3,

pp. 3
Kirby and Se. noe to Entomology. ee 1846, p. 215,
Knoppern Galls. Science London, 1865,
Linné, C. Lachesis eps ‘(crane Fig arene MS. and pub-.

1882,
/ a or Bussorah Galls. Pharm. Journ., Vol. z a 1848, pp. 422-24.
Modern American Tanning. Chicago, 1910, Vo
Muller, Albert. In Memoriam Wilson ee y . Zool., 2d ser., Vol.
3, London, 1868, pp. 1196-12
Nouveau Dictionnaire d’Histoire Natorétie. Paris, 1817. Article Galle.
Oil, Paint and Drug Reporter, April 16, 1917, N. Y., and Ann. Rept., 1917.
Olivier, i A. Voyage dans l’Empire Othoman. ... Vol. 2, 1801-1807,

P-

hal London, 1813. Article on Galls.

Pereira, J. Observations on the Chinese Gall, called ‘‘ Woo-Pei-Tsze’’ and
on Sa Gall of Bokhara termed aont tiaia. 22 Pharm. Journ., Vol. 3,
London, 1844, pp. 384-87, 3 figs.

Pereira, J. The Elements of Materia Medica and Therapeutics. Vol. 2,
London, 1850, pp. 1224-1233.

Pliny. Natural PEE translated by Bostock and Riley, 1856. Book 24
and 25 and

per A ate History of Drugs. . . . 3d edition, London, 1737, p.

Ukien +. 5 ae 1844, Vol.

Roxburgh, Wm. Trans. Soc. Arts, Manufacture ate nein Vol. 23,
London, 1805, pp. 407

Riibsaamen, Ew. H. Mittheilung úber die von Herrn J. Bor unta i
Oriente PERE e Zoocecidien. Zool. Jahrb., Jena, Bd. 16, 1902,

Smith, F. P. Contributidas towards the Materia Medica and Natural His-
tory of China. Shanghai, 1871, p. 100.

Stephenson, J., and Churchill, J. M. Medical Botany. London, Vol. 3, pl.
152.

Supplement to Commerce Reports, 1915, U. S.
Toothaker, C. R. Commercial Raw Materials. Philadelphia, 1905.
Trimble, Henry. Some American Galls. Am. Journ. Pharm., 1890, p. 563.

176 THE AMERICAN NATURALIST [VoL. LIT

ip Henry. The Tannins. Philadelphia, 1892.
U.S armacopoeia, 9th Revision, 1916.
Dai English Ink Galls. Pharm. Journ., 2d ser., Vol. 4, London,
862-3, 20.
Virey, J. J. Histoire Naturelle des galles des végétaux, et des insectes qui
les produisent. Journ. Pharm. et Sci. Access., Vol. 6, Paris, 1820, pp
161-169,

Vogl, A. Ueber Tamarisken-Gallen. Lotos, Prag, Vol. 25, 1875, pp. 133-
136.

Waring, E. J. Pharmacopeia of India. London, 1868, pp. 209-213; Ap-
pendix, p. 463.
Waring, E. J. Remarks on Some of the Use . Bazaar Medicines and
Common Medical Plants of India. London, 1874, p. 51.

THE
AMERICAN NATURALIST

Vor. LIT. April-May, 1918 Nos. 616-617

CONTINUOUS AND DISCONTINUOUS VARIA-
TIONS AND THEIR INHERITANCE IN
PEROMYSCUS

DR. F. B. SUMNER

SCRIPPS INSTITUTION, La JOLLA, CALIF.

I. [INTRODUCTION

Many of the views which we are now accustomed to
associate with the names of Weismann, Bateson, DeVries,
Nilsson-Ehle and others were either foreshadowed or
clearly formulated by Francis Galton, many years earlier.
Galton’s polygon, by which he illustrated the difference
between continuous and discontinuous variations, is
doubtless known to most readers; as is also his distine-
tion between ‘‘blended’’ and ‘‘particulate’’ inheritance.
It is less familiar, perhaps, that Galton regarded all in-
heritance as ‘‘largely, if not wholly, ‘particulate.’ ’’ Even
skin color, the classic example of blended inheritance in
man, is presumably ‘‘none the less ‘particulate’ in its
origin, but the result may be regarded as a fine mosaic
too minute for its elements to be distinguished in a gen-
eral view.” Again, ‘‘the blending in stature is due to
its being the aggregate of the quasi-independent inheri-
tances of many separate parts’’ (1889, p. 139).

Galton did not deny all heritability to those variations
which were represented by the minor oscillations of his
polygon, although he refers to such variations as ‘‘un-
stable.’’

With the modern revival of Mendel’s principles of

177

178 THE AMERICAN NATURALIST [ Vou. LIT

heredity and the definite formulation of a ‘‘mutation
theory” of evolution, some of Galton’s more or less tenta-
tive views have crystallized into dogmas. Along with the
two just mentioned, there has been incorporated the prin-
ciple of the ‘‘continuity of the germ-plasm,’’ a conception
which was likewise first clearly formulated by the great
English geneticist, though its modern expression we owe
to Weismann.

These various hypotheses have been woven together
into a single fabric and made to reinforce one another.
It will hardly be denied that some rather flimsy reason-
ing has been employed at times by those here concerned.
Thus one familiar syllogism runs somewhat as follows:
Somatic modifications are not inherited; fluctuating varia-
tions are not inherited; therefore fluctuating variations
are somatic modifications. Indeed, ‘‘somatie’’ and ‘‘non-
hereditary’? have come to be used interchangeably by
many writers. Whether or not somatic modifications
ever become germinal is a matter to be settled by evi-
dence. But I must confess that I have never regarded
as self-evident the contention that because characters are
found to be ‘‘non-hereditary’’ they are, ipso facto, ‘‘so-
matic’’ in origin.

A certain sanctity and inviolability has come to be at-
tached to the units of heredity or ‘‘genes,’’ according to
the neo-Mendelian creed. Not only do these units refrain
from any degree of blending, but—save for occasional
mysterious ‘‘mutations’’—they are quantitatively and
qualitatively unchangeable. Thus, the only differences
upon which selection, natural or artificial, can act are dif-
ferences due to the presence or absence of different
genetic factors. ‘‘We know,” say the Hagedoorns, in an
article (1917) which is typical of much of the recent lit-
erature of heredity, ‘‘that all the different genes, all the
different inherited factors ... are each in themselves
invariable. . . . Liability to chasse by selection is synon-
ymous with genotypic variability, and this true variability
o synonymous with impurity.” <

No. 615] INHERITANCE IN PEROMYSCUS 179

Much dialectic skill has been displayed in maintaining
this set of opinions against the many facts which seem
directly to refute them. Indeed, it must be conceded that -
a fairly consistent and logical edifice has been erected
upon these foundations. Strictly logical, though often-
times improbable interpretations have been given to each
new volley of hostile data, until the fortress has begun
to seem impregnable—at least to a frontal attack.

But perhaps, of late, another metaphor has come to suit
the situation better—that of the two knights fighting on
opposite sides of the same shield. The Mendelians have~
recently had recourse to more and more minute factorial
differences in explaining certain lesser gradations of
color in some of their material, until at length the dis-
tinction between their opponent’s s ““contimuity” and their
own ‘discontinuity ”’ is more imaginary than real.
Water is a continuous medium for all the ordinary pur-
poses of life, and solutions of different substances may
be completely ‘‘blended’’ therein. Its resolution into
hypothetical molecules, atoms, electrons and the like does
not in the least affect these fundamental facts.

The publication of the data which I offer in the present
paper confessedly does not constitute a ‘‘frontal attack”
upon the multiple factor hypothesis. My results belong
to a class of facts which have already figured extensively
in this controversy, and which have been met by ingeni-
ous and plausible counter-arguments. As I have stated
elsewhere, I am led to doubt very seriously whether any
concewable evidence could be brought forward which
would be admitted by the more extreme neo-Mendelians
to be really damaging to their position. As in so many
other cases, the victory is to be won, if at all, through
a process of ‘‘attrition.’’ Positions are gradually aban-
doned which are never disproved in a logical sense. In-
deed, as hinted, above, there are clear signs that the
defenders of the ‘‘multiple factor’’ explanation of selec-
tion and blended inheritance are already retiring from
their main positions.

‘ome

180 THE AMERICAN NATURALIST [VoL. LIT

Il. The DISTRIBUTION OF SUBSPECIES

The term subspecies, as here employed, is nearly equiva-
` lent to geographic race. These subdivisions of a species
occupy different, though often contiguous areas. When
contiguous, they are said to intergrade completely with
one another along the boundaries of their respective
territories; and in any case, their ranges of variation
overlap broadly. It is this fact, indeed, which leads to
their being ranked as subspecies, rather than as distinct
species, since the differences between some of the more
widely separated among them would be quite sufficient.
to give them specific rank were there no connecting forms.
In such reports as those of Osgood on Peromyscus
(1909), Nelson on the rabbits (1909), or Goldman on Neo-
toma (1910), the geographic ranges of certain species are
seen to be divided up into what look like quite arbitrary
subdivisions, corresponding to the ranges of the compo-
nent subspecies. The boundaries between these subdi-
visions oftentimes follow certain natural barriers, but in
some instances this does not appear to be true. And, in
any case, it is doubtful whether any geographic barrier,
save a continuous body of water or a lofty and unbroken
range of mountains could prevent the free diffusion of such
rodents. These minor areas, furthermore, frequently com-
prise territory having a very wide diversity of physical
conditions. For example, Peromyscus maniculatus gam-
beli is represented as ranging from the foggy coastal area
of central and southern California across the hot, semi-
arid San Joaquin Valley to the snowy heights of the
. Sierra Nevada. And in latitude, its range is said to ex-
- tend roughly from the 31st to the 48th parallel.
According to Osgood,
Specimens from Monterey, the type locality, are absolutely identical
with those from San Diego and the northeast coast of Lower California,
and the intervening region is inhabited exactly the same form.
These, moreover, are like specimens from . . . the west slope of the
Sierra (p. 69).

We might well be puzzled to discover any common ele-

=

No. 615] INHERITANCE IN PEROMYSCUS 181

ments of the physical environment which were responsi-
ble for the presence of the same subspecies under such
widely divergent conditions of life. Particularly is this
true when the environmental differences, as in the present
case, far exceed those between the habitats of certain
quite distinct subspecies.

Nor does the contention seem justified that- such ex-
tensity in the distribution of a single subspecies is fully
accounted for by the absence of any insurmountable bar-
riers to its dispersal. So far as geographic features are
concerned, the barriers between the range of gambeli and
the ranges of certain neighboring subspecies seem to be
no greater than some of those which traverse the terri-
tory of gambeli itself. Looking at the distribution maps |
in such publications as those just mentioned, one is im-
pressed by a seeming analogy between the boundaries of
these various subspecific ranges and those of the political
subdivisions of the earth’s surface. In considerable de-
gree these last are bounded by geographic features, but to
a large extent, also, the lines of demarcation seem to be
drawn quite arbitrarily—the territories merely bound
one another.

While great weight must be given to the findings of
these taxonomic experts, I think it is our duty at present
to accept certain of their conclusions with considerable
reservation. This is particularly true of assertions as
to the absolute identity of the characters of specimens
from widely different parts of a given range. The pub-
lished data make it plain that the authors are in no posi-
tion to detect minor differences of a statistical nature. A
small number of specimens from each locality are com-
monly compared, the measurements ‘‘in the flesh’’ of the
various specimens necessarily having been made by a
number of different collectors. It will be evident from
the ensuing pages that the differences with which we are
dealing are often of such a nature as to be revealed only
by the comparison of large numbers of individuals, meas-
ured according to uniform standards. As regards the

182 THE AMERICAN NATURALIST [ Vou. LIT

latter point, tests which I have made of the standards of
measurement employed by several competent collectors
show clearly that the differences due to ‘‘personal equa-
tion’’ are sometimes at least as great as those which char-
acterize quite distinct local races. _

Accordingly, we might feel justified a priori in enter-
taining some skepticism as to the homogeneity of these
races of animals throughout such great areas. Further-
more, I already have a certain amount of direct evidence
which renders this contention improbable. Such evidence
will be considered later.

-An extremely desirable undertaking would be to run a
series of trapping stations through the territories of two

A B
= E 7
E a

Fie. 1.

adjacent subspecies, at right angles to the supposed line
of demarcation. This the author hopes to do in the
course of time, though the task is not as simple as might
perhaps be anticipated. Theoretically, a number of pos-
sible conditions might be revealed by such an investi-
gation.

| In the first place, it might be found (Fig. 1) that each
of the two races was, in reality, ‘‘absolutely identical’’

A B

Sars a A a

Fie. 2.

throughout its own range, while the transition between
the two might be fairly abrupt.

| Secondly, there might be an unbroken intergradation,
in respect to the differential characters, throughout the

No. 615] INHERITANCE IN PEROMYSCUS 183

entire ranges of both the races (Fig. 2). In this case
there would be no real boundary between the two groups,
and indeed the recognition of two subspecies, rather than
one or three, would be quite an arbitrary procedure.

Finally, there might be a condition, less easy to repre- |
sent by diagrams, in which neither race was completely
homogeneous, each being subject to considerable local
variation within its own territory. Such local differences
might or might not tend to be graduated as indicated in
Fig. 2. Or, there might be some degree of gradation with
respect to certain characters (e. g., pigmentation), but
not with respect to others (e. g., length of appendages).
In such circumstances, the recognition of two ‘‘subspe-
cies’’ would depend upon the fact that the population of
each of the respective territories was relatively uniform,
and the changes encountered at the boundary relatively
abrupt. á

I am not yet in a position to say with certainty which of
these possibilities is realized in the case of the species with
which I am dealing (Peromyscus maniculatus), but I
already have some strong evidence that the third one most
nearly represents the actual state of arairs. As regards
depth of pigmentation, we certainly find something -ap-
proaching a graded series as we pass from the interior
desert regions of California toward the coast, or as we
pass from the coast of southern California, northward
into successively more humid regions, as far as Alaska.
But here we are dealing with a number of ‘‘subspecies.”’’
I have grounds for believing, however, that similar gra-
dations occur within areas conventionally assigned to
single subspecies.

Other questions of high theoretic importance relate to
the nature of the animals inhabiting the so-called ‘‘areas
of intergradation.’’ Does this intermediate population
manifest a complete blending of all the subspecific char-
acters, or does it consist of a mixture of individuals, sev-
erally exhibiting the respective racial characters in a
fairly pure state, or may there be a mosaic condition more

184 THE AMERICAN NATURALIST [ Vor. LIT

directly suggestive of Mendelian segregation? <A defi-
nite answer to these questions I am likewise obliged to
defer for the present.

Truly representative collections have been made by me
thus far at only four stations within the State of Cali-
fornia, though various other points have been visited
and considerable numbers of the mice have been trapped
there. My four principal collecting stations are located
near Eureka, Berkeley, La Jolla and Victorville.! Me-
teorological records were kept for about two years at
each of these points. These records have not thus far
been carefully analyzed, however, so that their publica-
tion must be postponed. A preliminary comparison of
climatic conditions at these four points has already been
made (Sumner, 1915a). It will suffice, for present pur-
poses, to state that, as regards both atmospheric humid-
ity and rainfall, these stations rank (from highest to low-
est) in the order given above, t. e., Eureka, Berkeley, La
Jolla and Victorville; while as regards mean annual tem-
perature the reverse order holds.

The distribution of the three subspecies of Peromyscus
maniculatus, recognized by Osgood as occurring within
the limits of California, is represented in Fig. 3. It will
be seen that one of my stations (Eureka) lies within the
range of rubidus, another (Victorville) within the range
of sonoriensis, while the other two (Berkeley and La
Jolla) lie within the range attributed to gambeli.

In the ensuing pages, I am not in the least concerned
with characterizing and defining those taxonomic groups
which have been called Peromyscus maniculatus gambeli,
rubidus and sonoriensis. I shall merely discuss the dif-
ferences between (and within) four representative collec-
tions taken by me in widely separated and elimatically
different regions of the state. The question as to what
“subspecies”? a given mouse ‘‘belongs to’’ is pior my pur-
poses a distinctly minor consideration.

1 Four other stations have been added since the present paper was writ-
ten, but the data derived from these can not be included here.

No. 615] INHERITANCE IN PEROMYSCUS 185

DISTRIBUTION MAP
MUSEUM OF VERTEBRATE ZOOLOGY

sam
ing that of poser the Nentes st that of s onortensie. Supposed areas of inter-
gradation between two races are indicated > dotted lines.

186 THE AMERICAN NATURALIST [ Von. LIT

III. DIFFERENCES BETWEEN THE Four Loca Races UNDER
CONSIDERATION?

These differences may be divided, for the sake of con-
venience, into pigmental and structural ones. Since the
former are the most obvious, they will be discussed first.

1. Pigmental Differences

The pigmental differences relate to (1) the hair, (2)
the skin.

Hair.—Like the other members of the genus Peromys-
cus, the mice of the present group are covered with pig-
mented hairs upon the dorsal and lateral surfaces, while
the ventral surface and to a large extent the feet are cov-
ered with white hair. Upon the trunk these white hairs
are, to be sure, devoid of pigment only at the distal ends.
Parting the pelage at any point, dorsal, ventral or lateral,
_ reveals the presence of a slate-colored basal zone in each
hair.

The most obvious differences between the races under
consideration relate to the dorsal coat color (Fig. 4).
This is darkest in the animals from the humid redwood
district (Eureka), palest in those from the Mojave Desert
(Victorville), and of an intermediate hue in the collec-
tions from Berkeley and La Jolla. These last two races
likewise differ from one another, the former being darker
than the latter. Thus we have a series of four grada-
tions, which are correlated directly with gradations in
the rainfall and atmospheric humidity of their respective
habitats.

It is important to notice, however, that these differences
of shade relate rather to averages than to individual cases.
All of the Eureka mice are not darker than all of the
Berkeley mice. Nor are all of the Berkeley mice darker
than all of the La Jolla mice, nor all of the latter darker
than all of those from Victorville. In comparing repre-
2] here use the word “race”? as being a non-committal term, elastic
enough to cover r two collections of individuals which show significant

x differences of type

Wie. 4, Skins of male adult wild specimens of the Eureka, Berkeley, La Jolla and Victorville races
‘named. The skins have been selected with a view to showing the average shade of each series

of Peromyscus maniculatus, in

order

[ST9 ‘ON

SAISAWOUAd NI AONVITIAAHNI

LST

188 THE AMERICAN NATURALIST [ Von. LIL

sentative collections of any two adjacent races belonging
to this series, there is found to be a broad zone of over-
lapping. That is to say, there are many individuals in
each set which, so far as color goes, could be equally well
placed in either. I have, for example, laid out in par-
allel rows considerable numbers of sonoriensis and the
La Jolla form of gambeli, and found that the darker half
of the former set completely overlapped the paler half of
the latter. While no confusion would be possible between
the paler sonoriensis and the darker gambeli, there were
a large number of specimens which could only arbitrarily
be assigned to either ‘‘subspecies.’’ Indeed, it is freely
admitted by systematists that in many cases they can
assign a given specimen to its proper subspecies only if
they know the locality at which it was trapped. No such
confusion would be possible, however, between the more
divergent races of our series, e. g., those from Eureka
and the desert. I have never seen a rubidus which could
not, by color alone, be readily distinguished from sonori-
ensis and vice versa.

Any attempt to give verbal equivalents for these color
differences is highly unsatisfactory. In a later report I
expect to undertake the analysis of these shades by means
of a color wheel. For the present I will content myself
— with a very brief statement. The dorsal darker stripe
of the Eureka mice is of a shade lying somewhere be-
tween Ridgway’s ‘‘sepia’’ and black, the paler lateral
region lying between ‘‘Saccardo’s umber’’ and ‘‘sepia.’’*

3 The ensuing remarks apply only to the mature pelage. These mice pass
through three distinct pelage phases: (1) the juvenal, which, in all races, is
neutral gray in hue, and considerably darker than the adult shade; (2) the
post-juvenal or adolescent, commonly paler and yellower than the last; (3)
the mature or adult pelage, which is still more highly colored and frequently
of still paler shade. The first molt occurs some time durin ng fee second
month after birth, the second some time between the age of six months and
a year. The various races of mice here considered, and even ng mutants,
are probably as clearly distinguishable in the immature pelages as they are

4 See o. 1912, ‘‘Dresden brown’’ and ‘‘mummy brown’’ perhaps
approximate the shades in question as well as the last two mentioned.

No. 615] INHERITANCE IN PEROMYSCUS 189

Since the coat color is at no point homogeneous, any such
comparison with nd tinted paper is of course very
crude.

The desert mice are of a hue which can not even ap-
proximately be represented by reference to Ridgway’s
‘<color standards.’’ The effect is probably not far from
that which would result from a mixture of fine streaks of
black and of ‘‘ochraceous buff’’ or ‘‘cinnamon buff,’’ so
proportioned as to approximate the general hue of the
barren soil. As in all of these races, the mid-dorsal pel-
age is commonly darker than the lateral. All that need
be said of the two collections of ‘‘gambeli’’ is that they
are intermediate between the extreme types just re-
ferred to.

As a special case of the general hair color of the body,
though not entirely correlated with this, is to be men-
tioned the color of the ‘‘ankle’’ region. The latter, par-
ticularly on its ectal surface, is covered with very short
pigmented hairs, whose depth of shade affords another
feature distinguishing the average condition of these four
races.

I have given considerable attention to a microscopical
examination of the hairs of these various mice. In posi-
tion the pigment is of two different sorts, axial and super-
ficial, located in the medulla and cortex respectively.
A series of more or less disc-shaped, black pigment
bodies extend from the base of each hair throughout
the whole, or a considerable part, of its length. In the
stouter hairs, there are, in the expanded region, two to
four longitudinal rows of these bodies. In all cases, they
alternate regularly with air spaces.

Tn the all-black hairs, the black pigment extends very
nearly to the extreme tip. In the banded hairs, a region
of varying length occurs in the distal half, in which the
black pigment gives place to yellow. The dark pigment
does not end abruptly, however. The dense black bodies
become fragmented into their component rounded gran-
ules, as we pass from one segment to another, first giving

190 THE AMERICAN NATURALIST [ Vor. LIT

way to scattered collections of these granules (which are
dark brown when seen singly) and later disappearing
altogether. In the transitional region, black and yellow
pigment may frequently be found in the same segment.
In most hairs, the dark bodies again replace the yellow
ones as we pass toward the tip; occasionally the yellow
continues as far as any axial region is distinguishable.

The yellow pigment seems to be restricted to the axial
part of the hair. To some extent, it occurs in the form
of granules, but, unlike the black, it is largely present in
a diffuse condition. This pigment is not all of the same
tint, but varies in shade from a pale yellow to an orange
or even a very pale brown.

For a varying length, on the distal, tapering ends of
nearly all the hairs of the colored parts of the body, there
is a very dark, granular pigment, lying close beneath the
surface of the hair. This overlies and reinforces the ax-
ial pigment, so that the distal end is frequently darker
than any other part of the hair. The superficial pigment,
where dense, commonly looks almost black, but when seen
in a thin layer the single granules appear brown. As
already stated, this is likewise true, though in lesser
degree, of the ‘‘black’’ axial pigment. In one of the
““mutants,”” to be described later, this distal dark zone is
nearly or quite lacking, and the same is true of certain ex-
ceptional samples of hair taken from normal individuals.

The yellow pigment is readily soluble in even fairly
dilute potassium hydrate solutions, whereas the dark pig-
ment is very much more resistant to this reagent, and
may remain unchanged, even after the complete disinte-
gration of the hair.”

Most students of this subject seem to follow Miss Dur-
ham (Bateson, 1903) in recognizing three pigments in the
hair of Mus musculus—the black, the brown or chocolate,
and the yellow. After considerable examination of the

5 I have, however, observed preparations in which even the densest black

pigment bodies assumed a reddish brown color, especially near the margin
of the cover-glass.

No. 615] “INHERITANCE IN PEROMYSCUS 191

hair both of Mus and Peromyscus, I can not feel sure of
any sharp distinction between the black and the brown
pigments. It is true that the axial pigment bodies of the
basal portions of the hair are nearly dead black, while
most of the superficial pigment at the distal ends is dis-
tinctly brown. But all gradations occur in the axial pig-
ment of the transitional zones, and these gradations ap-
pear to be due not merely to differences in the density of
the clusters of granules, but to gradations in the depth
of color of the individual granules themselves. Without
having made any careful chemical tests, I am disposed to.
believe that black and brown, in the hair of mice, are due
merely to different degrees of aggregation of a single
pigment. On the other hand, this dark pigment seems to
differ, chemically and otherwise, from the various shades
of yellow.

The differences in the color of mice of different sub-
species and of different parts of the pelage of a single
individual appear to be due to two chief causes: (1) the
relative length of the pale zone, in relation to the rest of
the hair; and (2) the proportionate numbers of the all-
dark and of the banded hairs; probably also to (3) the
depth of shade of the yellow pigment in the pale zones,
and (4) the degree of concentration of the superficial pig-
ment at the distal ends. In some of the ‘‘mutants,’’ as
will be pointed out below, certain other factors contribute
to the differences shown.

Of importance for our general viewpoint is the fact that
no one of the geographic races which has been examined
possesses any type of hair which is wholly lacking in any
other race. It would be impossible from a single hair,
or even a small group of hairs, to say from what sort of
mouse they were taken.

When viewed on the ventral side, these four races of
mice likewise present characteristic differences. They
form a graded series in respect to the whiteness of the
pelage, which is purest in the desert race and least so in
that from the redwoods. The differences are found to

192 THE AMERICAN NATURALIST [ Vou. LIL

result from the relative length of the terminal pigment-
less zone which is present in these hairs. The ventral
hair of the desert race also appears to be somewhat longer,
or at least of a softer texture, than that of the others.

In the ease of the ventral surface, like that of the dorsal,
these differences relate to averages rather than to indi-
viduals. Likewise, it is of interest to note that within
each race there is little or no correlation between the
dorsal and the ventral shade. I have frequently graded
a considerable row of mice of a single race in respect to
the shade of the dorsal pelage, and found, on turning the
animals over upon their backs,* that the order of arrange-
ment did not correspond with the ventral gradations of
shade.

Another differential character of these races is the de-
gree of lateral extension of the ventral white area of the
body, or, conversely stated, the ventral extension of the
dorso-lateral pigmented area. The colored and uncol-
ored regions of the body come together abruptly along an
irregular lateral line extending from the snout to the tip
of the tail. In the desert race, more of the white ventral
region is usually to be seen in side view than is seen in
the darker races. The gradation of the other three races
among themselves is less obvious.

This degree of extension of the colored area relates not
merely to the body but to the appendages. In the darker
races an elongated tongue commonly extends down upon
the fore-limb, in some cases even to the hand, while in
sonoriensis ‘such a ventral projection is usually little de-
veloped. The graduation of our four races in regard to
this character corresponds to that noted in respect to
shade. Similar conditions are observable on the hind
limbs, particularly upon the ankle, where the pigmented
hair may extend as far as the heel, or may fall short of
this in varying degrees. The case of the tail will be dis-
cussed separately.

A hair character which seems to be peculiar to sonori-
ensis, among the races here considered, is the presence of

6 Fresh specimens, not skins, are used for most of these comparisons.

No. 615] INHERITANCE IN PEROMYSCUS 193

small clusters of white-tipped hairs near the anterior in-
sertions of the ears. But even this feature is not evident
in all individuals.**

Many species of Peromyscus, including the maniculatus
series, have what is known as a ““bicolored”” tail. The
hairs throughout a longitudinal stripe of varying width,
upon the dorsal surface of this member are dark brown or
black, while those of the ventral side are white. Now a
casual inspection serves to show that this caudal stripe
is broader and darker in the Eureka mice than in the
desert ones, while a more careful comparison shows that
the ‘‘gambeli’’ individuals are, on the whole, intermediate
between the other two.

Fortunately, the breadth of this stripe is a character
which may be subjected to fairly accurate measurement.
It is my practice to slit the skin of the tail along the mid-
ventral line, strip it off, and press the inner, damp sur-
face firmly against a strip of black cardboard. The total
width of this skin (= circumference of tail) is then taken
at the mid-point of its length; likewise the width of the
tail stripe. The ratio between the two readings is next
determined, the width of the dorsal stripe being expressed
as a percentage of the circumference of the tail. The fol-
lowing are the figures for the four races and the two
sexes, the figure in parenthesis representing the number
of animals measured :?

TABLE I
fapides, A (QO) A Se, ees 42.51 + 0.45
Mites S thy ae 41.96 + 0.53
Berkeley gambeli, 3 (24) ................ 36.08 + 0.80
Berkeley gambelé, 9 (28) .iocoiooo.o.o.o.. 35.50 + 0.56
La Jola gumbel, d (85) ¿cd 32.08 + 0.33
La Jolla gamba 9 TOY 1... 32.43 + 0.49
sonoriensis ah (TI) 6456. a ts 27.49 + 0.32
PORRO A a. 28.92 + 0.36

6a This condition I have recently found to occur in occasional specimens
of rubidus trapped near Carlotta, California.

7Since this character was not measured when these studies were first
commenced, the number of individuals included in the present table falls
far short of those measured for some other characters.

194 THE AMERICAN NATURALIST [ Vou. LII

While the statistical certainty of these four types can
not be doubted, it must again be insisted that the differ-
ences relate to averages rather than to individual ani-
mals. The frequency distributions of the various widths,
as represented by the histograms (Fig. 5), show this
point clearly. There is a certain amount of overlapping,
even between the most divergent races.

Skin Pigmentation.—Certain regions of the skin are
colored by dark pigment. The regions showing skin pig-
mentation most clearly are the ears, tip of snout, soles of
feet, and, in the males, the scrotum.

Frequent comparisons of considerable numbers of
freshly killed specimens have made it plain that, in re-
spect to the pigmentation of the ears, our four races can
be arranged in the same graded series as was found to
hold for coat color. As regards the other three skin char-
acters, I have never compared more than two races at a
time, but I feel little doubt that all four could be arranged
in the same order. No exact measurements are here pos-
sible, as in the case of the tail stripe. In a few instances
I have, however, graded a given character, according to
an arbitrary scale, and have thus been able to express the
differences between two races in a roughly quantitative
way. The following comparison between 42 sonoriensis
and 38 La Jolla gambeli with respect to the pigmentation
of the scrotum will illustrate this point.

TABLE II
sonoriensis gambeli
Numb f Percentage Numb t Percentage

ar AA ys OS EA MERES 1 2.4 3 7.9
Moderate 0.0 tE. 3 14 2 5.3
WAG aia O 5 11.9 7 18.4
Nery Ae. ee E 2.4 4 10.5
a e a 32 76.2 22 57.9
Total.. 42 100.0 38 100.0

A similar tabulation was made in another case, com-
paring two lots of specimens of these same races in re-
spect to the pigmentation of the foot.

No. 615] INHERITANCE IN PEROMYSCUS 195

It might readily be contended that all these various pig-
mental differences, which have thus far been considered,
_are merely manifestations of some general tendency tow-
ard a given degree of pigmentation of the body as a

>>

rubidus
ng Cases. Meana 42.28

-Nu $ un 5.0

ego beli
ean= ~ Y

<a OF DH A

/

/

I
üj

+ Mean =32-21

La Jolla gambe
31

I
sonoriensis !

13 cases .Mean= 28.12
+

I

120 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 33 40 4) 42 43 49 45 46 47 98 93 50 SI SZ 53 S4 55 36 ST”

Fic. 5. istograms, showing the frequen stributions of the percentage
values for the width of the tail es (ratio re a in ja four races
here considered (sexes combined).

whole. This tendency perhaps manifests itself in a

greater or less intensity or extensity, or both. In the
absence of exact quantitative standards, it is impossible

196 THE AMERICAN NATURALIST [ Vou, LII

to determine whether or not these various pigment char-
acters are correlated. If they are, the correlation is cer-
tainly not a close one, as frequent observations have
shown. For example, the grade of foot-pigmentation was
determined for the paler and darker halves of a series of
sonoriensis and also for a series of gambeli. In both
cases the average grade for the foot was slightly greater
for the darker half than for the paler; but the difference
was so small that I am not sure of its significance. Again,
in the lot of 38 gambeli comprised in Table II the darkest
individual (dorsally) and the one with the darkest feet
were both devoid of visible pigment on the scrotum.
Similar entries are frequent among my notes.

2. Structural Differences

The structural features which I have subjected to quan-
titative determination are (1) weight, (2) body length,
(3) tail length, (4) foot length, (5) ear length, (6) number
of tail vertebre; together with several other skeletal char-
acters which I shall not discuss in the present paper.
The methods employed throughout these studies will be
described more fully in a later report. A brief statement
will suffice for the present. Body length, as here em-
ployed, is the total length, minus the length of the tail.
In taking the total length, a special contrivance is em-
ployed, the body being stretched slightly and to a uniform
extent. A constant procedure is likewise employed in
measuring the tail length. The figure recorded for the
latter represents the distance from the first free caudal
vertebra to the tip of the tail, under a uniform degree of
tension. The ear length here used is that from the sum-
mit of the ‘‘notch’’ to the tip of the ear. Foot length is
the distance from the heel to the tip of the claw of the
longest toe, the foot being pinned, sole downward, to a
blackened board.

The statistical methods employed in analyzing these
data have been rather fully discussed in a former paper
(1915), to which the reader is referred.

No. 615] INHERITANCE IN PEROMYSCUS 197

(1) Weight and (2) body length are not dealt with di-
rectly in the present comparisons. The former is an in-
dex of metabolic condition as well as of size (1. e., length).
Captive mice, for example, are commonly fat in compari-
son with wild ones. Body length is of little significance in
comparing two groups of mice, unless we know, either
that the animals are all of the same age, or that the limit
of growth has been reached by all of them. These things
are frequently impossible to determine.

(3) Tail length is dealt with, both as an absolute value
and as a percentage of the body length. If absolute tail
lengths are to be compared in two groups of animals, the
comparison can only be made between animals of approxi-
mately the same body length. My practise is to divide
each series into a number of size-groups, differing by
only two millimeters of body length. Group 80-81 of
one series is then compared with group 80-81 of the
other, group 82-83 with group 82-83, ete. The graphs
(Figs. 6 and 7 and 9-12) are based upon this procedure.
Each ‘‘curve’’ connects the means of the size-groups of
each series of animals, the abscissas representing body
length, the ordinates the character under comparison.
In order to eliminate very young mice, groups having
body lengths of less than 80 mm. are omitted. Even so,
it is likely that most of the animals in the lower groups
of the series are immature, but this fact in no way affects
the validity of the comparisons.

It will be seen, from an inspection of the figures (6 and
7), for both males and females, that, as regards tail length,
the Eureka mice (rubidus) stand in a class by themselves.
In comparison with the wide interval between this long-
tailed race and the other three races here considered, the
latter differ but slightly among themselves. It is evi-
dent, none the less, that the La Jolla animals have some-
what longer tails than do those from Berkeley or Victor-
ville, while the last two agree fairly closely in their mean
condition.

Relative tail length, i. e., the length of the tail expressed

—
e —

ES
E O

79
/ 14
78 e ;
qe
77 es tag A
ri ‘
76 bs !
75 2/ i
6 i
16 i
74 In EN '
ve fo EN !
+ `, y Y
rae rs EA y / NWA IG i
ra J 1 1] “os
1 Yv,
3
7i
I
a oe
69
/
68 z
41
es rubidus
65 2an La Berkeley tambel

————— Ladolla gambeli
sonoriensis

80 8i 62 83 8+ 85 86 87 88 89 J0 91 92 93 J4 35 96 97 38

Fic. 6. Graphs for i gg of the absolute tail lengths in the four races
(males). Abscissas denote ength; ordinates denote tail length; the figures
along the “ curves ” indicate the Ases of individuals in the various size groups.

67 Fa op aomen anmann onan rubidus

66 A == Berkeley gambeli
i amm La Jolla gambeli

ahd ‘igh

a e sonoriensis

80 8) 82 83 84 85 86 87 88 89 30 91 92 93 9495 36 37 38 99 100 101 102

Fic. 7. Comparison of the absolute tail lengths in the four races (females).

200 THE AMERICAN NATURALIST [ Vou. LIT

as a percentage of that of the body, is to a considerable
extent independent of the size of the animal. Larger
mice, it is true, have relatively somewhat shorter tails
than do smaller ones. But the differences are so slight
that they may be overlooked, unless the mean size of the
two groups under comparison differs considerably. The
relative tail lengths of our four races of mice may be com-
pared in the following table. This shows the same rela-
tions as were portrayed by the graphs. It also shows
that there are no significant differences between the sexes
as regards the length of this member.

TABLE III
sabuk CBE ei a ee ce cele 104.45 + 0.38
PUDONG, OCT A es eas ee 103.37 + 0.52
Berkeley gambeli, ji (26) .............. 81.69 + 0.55
Berkeley gambeli, Y (21) .............- 81.76 + 0.65
La Jolla jambe ft (99) cir 84.36 + 0.35
Ea Jolla. gambeli, Q (45) .... 2. «0... 83.04 + 0.43
Maoriene RIG) A Sere as as a d 81.29 + 0.44
nora e (DL. cs ee ee 81.30 > 0.45

The distribution frequencies for these various lengths
are represented by the histograms (Fig. 8). From these
it is evident that only an oceasional Eureka mouse has as
short a tail as the longest tailed members of any of the
other three races. The latter, however, differ from one
another but slightly.

(4) In respect to foot length likewise (Figs. 9, 10) the
Eureka mouse is very distinct from the other three races,
while the latter show no significant differences among
themselves. It is of interest, however, that in all four of
these races the female has, on the average, a slightly
shorter foot than the male. If any one still entertains the

8 Owing to a slight «change in the manner of measurement, which w:
made after these studies were commenced, the tail lengths of the Cater
animals of my series have been rejected from the computations, This proce-
dure has affected particularly the numbers of the Berkeley series.

No. 615] INHERITANCE IN PEROMYSCUS 201

8
7
é
x rubidus
hi ter Cases.Mea n= 104.01
3
2
i
1 | | O
4
17 Pk
16 Pa
15 he Á
4 r
13 7
2 Pes
n Va
10 A É
3 g”
: re
?
z La Jolla gambeli A
» 144 Cases. Meon = 83.95
3 4
2
'
7
I
7 I
6
5
7 +
y pen e eli
i S. Mean = 61.72
ELL M
I
is |
% l
i3 |
2 4
n '
bag (j
3
a
7
6
3 >.
+ orie 13
s “6 cases. Mean =8119
2 I
1 i
66 67 68 69 70 7} 12 73 74 75 76 77 78 79 80 81 82 83 84 85 06 87 BH 89 90 31 32 33 34 35 J6 97 38 3) e m n2 103 ne ns 16 117 MS us Ro

Fic. 8. Histograms, showing the frequency distributions of the paige ao
values for the length of the tail (ratio to body length), in the four races (sex
combined

notion that small feet, along with other feminine charms
in mankind, are due to ‘‘sexual selection,”” the situation
in Peromyscus ought to give him pause.® The mean dif-

® This same difference was found by me to hold for white mice, at least
for full grown individuals (1915, pp. 358, 367).

202 THE AMERICAN NATURALIST [ Vou. LIT

ferences between males and females, computed according
to a method earlier described by me (1915, pp. 345, 346),
are:

PRISE St hf Rs ow Sie op og ee ES 0.31 mm. + .08
gambeli (Berkeley) ............-..... 0,29 mm. + .05
gambeli (La Jolla) ............-...-+* 0.09 mm. + .07
HONOR 00 cs ee a ee ae ae 0.38 mm. + .05

(5) In respect to ear length, we find a quite different
set of relations.. It is the La Jolla mouse in which
these appendages are the longest, the Berkeley mouse in
which they are the shortest, while the redwood and the
desert animals occupy an approximately intermediate
position and scarcely differ significantly from one an-
other. It is here to be noted that the two extremes of
the series, in respect to this character, have been placed
by the systematists in the same ““subspecies”” (gambeli).

(6) The counting of the tail vertebre, like the other
measurements of skeletal characters, has not yet been
completed. I have, however, determined the number in
25 specimens each of the Eureka, La Jolla and Victor-
ville races. The fifth vertebra, counting from the most
anterior one in the sacrum, has been regarded as the first
of the caudal series. The averages and the frequency
distributions are indicated in the following table.

TABLE IV
| 23 | 24 | 25 | 26| 27 | 28 | 29 | 30 | 31 | average
ey OS | |2| 9/6 5/11/12 | 28.0
gambels (La Jola) occ o e 4015 13) 3 | 26.6
CONO AS 112 si7l6l1 r 25.7

The significance of these differences seems highly prob-
able, despite the small numbers. That between rubidus
and sonoriensis can hardly be questioned. It seems plain,
however, that the differences in tail length between these
various races is not accounted for by the differences in
the number of the vertebre. Thus the Eureka mouse
has a mean tail length which is 28 per cent. (of the smaller
- number) longer than that of the desert mouse. The pre-

No. 615] INHERITANCE IN PEROMYSCUS 203

S
de
2
22.0
7
4 a
2 3
Y E
21.0 /
9 À
fs E
E Se al
s El
4 y s 7
3 SS
7 y
20.0 14 x
9 ag ss
4 e —_—_—_ Berkeley gambeli
‘+

————- La Jolla gambeli
ee okie a in sonoriensis

80 8! 82 83 84 85 86 87 88 89 90 91 92 93 9495 96 97 98 99 100
FIG Comparison of foot-lengths in the four races (males).

pedo Nevo ht ane

Sa

A

rubidus

Fubu»vo-rbruny mo

2
`
` .
9.0 & ; a Berkeley gambél
wo. /
N ‘ es o e Ladoila gambeli
Ax 1
s +
A TAO T A B Se a AoA sonoriensis
` ‘
yt
v E

80 8l 82 83 8+ 85 86 87 88 89 90 JI 32 93 J+ 95 36 37 38 J9 100 101 102
Frc. 10. Comparison of foot-lengths in the four races (females).

204 THE AMERICAN NATURALIST [ Vou. LIT

ponderance in the number of vertebra is only 9 per cent.
The differences in the length of this appendage are there-
fore due partly to the number of vertebra, but chiefly to
the length of the individual vertebre.

Résumé of Racial Differences.—In relation to the vari-

&FaaN

e
o-h

j
]
l
l
l
!
I
Lae
!
}

si

O-N&RODANDL

5

3
hb pUBNOL ORE TU in N d to

eae

-f I6

F

RETES

maa TRACE

Berkeley gambeli
em pu Lavolia ¿ambeli
escanea SONOTIENSIS

80 8l 82 83 84 85 86 87 68 89 30 91 92 33 9495 36 97 38 33 100

Fic, 11. Comparison of ear-lengths in the four races (males).

ous pigmental differences, those both of intensity and ex-
tensity, the four races under consideration were found to
present the following graduated series:
Eureka > Berkeley > La Jolla > Victorville.
As regards the length to the tail, the series became:

Eureka > La Jolla > { Victorville

No. 615] INHERITANCE IN PEROMYSCUS 205

When the number of caudal vertebre was considered,
we had the same arrangement as the last for the three
races for which determinations had been made, viz.:

Eureka > La Jolla > Victorville.
In respect to foot length, the following order held:

3 ‘A
oT

4
w%

UAne@veo—-NeFUONWOEWYO=HYOEFUDRNOVO~NHLPUDBNSY oo py pi
E

-—
~

(ins PUDPAUS

z

Berkeley gambeli
ss LA Jolla gambeli

A Sanari ensis

30 81 82 83 84 85 86 87 86 89 90 91 32 93 94 95 36 37 38 33 100 101 102

Fic. 12. Comparison of ear-lengths in the four races (females).

La Jolla,
Eureka > E Berkeley,
Victorville.
Finally, as regards ear length, we had a quite different
alignment, viz. :
La Jolla > Es Victo eat > Berkeley.

It is plain that these ““subspecies”” have diverged from
one another in respect to characters which have varied

cd

206 THE AMERICAN NATURALIST [Vor LIT

quite independently. There is no single graded series
for all the characters, which would lead us to suppose that
they are in some way correlated or ‘‘linked’’ together.

When pigment characters alone are considered, the
Berkeley mice are certainly intermediate between the La
Jolla and the Eureka ones, and to that extent may be said
to ‘‘approach rubidus.’"° But this is not true of the
length of the tail, the foot or the ear. Indeed, as regards
the first of these appendages, the Berkeley race diverges
even farther from the Eureka race than does that of La
Jolla.!*

The question whether any of these various character
differences may be physiologically or genetically linked
together, so as to exhibit concomitant variations, is an
interesting one, which I hope, in time, to treat rather
fully. But I have already computed coefficients of cor-
relation between two pairs of characters, viz.: between
tail length and width of tail stripe, and between tail length:
and foot length.

In obtaining the former, I have based the coefficients
upon the deviations from the mean relative tail length of
each race and each sex, taken separately. Of these coeffi-
cients, three are positive and five negative. They range
from — 0.23 to + 0.09, the mean being — 0.03. Thus, it
is plain that there is no appreciable correlation, within a
single race, between the width of the tail stripe and the
length of the tail, despite the fact that these characters
seem to be associated, when certain darker races of the
northwest coast are compared with more southward rang-
ing forms.

There is, however, a quite marked correlation between
the length of the tail and that of the foot. I do not here
refer to the obvious fact that larger animals have larger

10 Osgood, 1909, p. 69. This author likewise states that Berkeley speci-
mens are iongan than typical gambeli.’ ’

11 This conclusion is strengthened by consideration of an even larger
series of Berkeley mice which were not included in Table III. The two sets
were trapped in two different localities in the Berke i

No. 615] INHERITANCE IN PEROMYSCUS 207

tails and likewise larger feet than smaller animals. My
figures show that, even when animals of the same body
length are considered, those with longer tails tend, on the
whole, to have longer feet, and vice versa. To obtain
these results, I have computed the coefficients separately
for each size-group, containing ten or more individuals.'?
All but 5 of these 21 figures are positive, the mean being
+ 0.27. Thus the greater tail and foot length of the Eu-
reka race may have arisen simultaneously, both being the
expression of a single constitutional change.

One further word regarding the nature of these racial
differences, before we pass to a consideration of their
heredity. It is plain that, with a single possible excep-
tion, all of the differences thus far considered are ‘‘sub-
stantive,’’ rather than ‘‘meristic,’’ to follow Bateson’s'®
terminology, or ‘‘proportional,’’ rather than ‘‘numeri-
eal,’’ to use terms recently employed by Osborn.'* In no
case are they of the nature of ““presence-and-absence””
differences, such as figure so widely in Mendelian litera-
ture. Whether or not, on ultimate analysis, they can be
resolved into the latter category, will be discussed later.

The differences without exception relate to means and
modes, as was illustrated above by histograms con-
structed for two of the characters (Figs. 5 and 8). The
frequency polygons commonly overlap broadly, when ad-
jacent members of the series are compared. We find an
approach to discontinuity only in a comparison of the
most widely divergent races.

The single difference of a meristic or numerical char-
acter is that relating to the number of caudal vertebre.
But even here the difference is one of averages, for no
single race seems to be characterized by the unvarying
presence of any particular number of vertebre, as certain
larger taxonomic groups are characterized by a definite
number of teeth or mamme. Itis worth mention also that

12 Cf, — 1915, pp. 349-350, 409-415.

13 1894, pp. 22, 23. ,

14 1915, p. 199. In the paper referred to, Osborn has given some atten-
tion to the case of Peromyscus.

208 THE AMERICAN NATURALIST [ Vou, LIT

the last one or two caudal vertebre are commonly rudi-
mentary, so much so that it is not always easy to deter-
mine their exact number. It is scarcely more fitting to
apply the term ‘‘meristic variation’’ here than it would
be to apply this term to such variations in the number of
cells as distinguish a larger from a smaller foot or ear.

(To be concluded.)

INTERNAL FACTORS INFLUENCING EGG PRO-
DUCTION IN THE RHODE ISLAND RED
BREED OF DOMESTIC FOWL II.

DR. H. D. GOODALE :

MASSACHUSETTS AGRICULTURAL EXPERIMENT STATION, AMHERST, Mass.

Cycles.—By cycles of egg production are understood
the existence of periods of egg production alternating
with periods either of decreased egg production or entire
cessation of egg production. These cycles may be either
long or short. The long-term cycles may have a period
of ayear. Shorter cycles exist with a period of three or
four months, i. @., winter, spring and summer and fall.
There are still shorter cycles with periods measured in
weeks, while one may also recognize irregular cycles. A
litter, as defined by Miss Curtis (714), is a short period
of egg production alternating with a non-productive
period and is well illustrated by broody birds, though it
may occur in non-broody individuals. A ‘‘elutch,’’ ac-
cording to Miss Curtis, is the set of eggs produced on
consecutive days. Its termination is marked by the ap-
` pearance of a blank day.

The form of the yearly cycle depends to a consider-
able degree upon some of the internal factors under dis-
cussion, that is, it varies in different individuals. Egg
production, as a rule, begins in the fall and winter, and
continues at a fairly constant rate in most individuals
until spring, when the rate rises somewhat in many indi-
viduals. Those that have been laying at a relatively high
rate do not show this acceleration, at least not in as
marked a degree. Sooner or later in a broody race
broody periods appear, which interrupt egg production at
fairly constant intervals. The rate, however, during the
nonbroody periods shows no slackening, but on the con-
trary a very slight acceleration may be demonstrated, so

209

210 THE AMERICAN NATURALIST [ Vou. LII

that the lower egg production noted after the first broody
period is due solely to the interruption of production.
During the summer, the rate of egg production slackens,
due almost entirely to the broody periods. The data for
the 1913-14 flocks show that after June the rate of pro-
duction .is about constant for the next three months,
largely because the first of July marks the point at which
practically every individual in the flock has entered on its
broody portion of the year. Some time in the late sum-
mer or during the fall, the various individuals stop lay-
ing and moult, some at one time, some at another, but
usually at the end of a broody period. After the rest
period in the fall, the birds gradually begin to lay again
in mid-winter, somewhat as they did as pullets, except
that the rate is slower asarule. Except for this feature,
the character of the second year’s production is much
the same as the first.

The winter. cycle is regarded by the workers at the
Maine Station (Pearl, 712) as the most important of all the
cycles, at least from the standpoint of the investigation of
the inheritance of egg production. They have found
that it represents a definite period in the life history of
the individual, among their Barred Plymouth Rocks.
Furthermore, during this period, the greatest differences
are to be observed in the egg production among indi-
viduals. They also find that high winter egg production
is correlated with annual egg production, as would be ex-
pected ‘except in the event that high egg production early
in life tends to lower production in later life. In other
words, a bird that is a good layer during the winter is
probably a good layer at all times. There are other
reasons, mostly of a practical nature for the use of the
winter cycle as a measure of fecundity.

Taking the year as a basis the workers at the Maine
Station recognize as its first characteristic the winter
eycle beginning with the first egg of the pullet and ex-
tending to March 1. This date is taken as a convenient
working point that falls near the biological division point.

No. 615] EGG PRODUCTION 211

During this period, flock production rises from zero to a
maximum and then slows down somewhat toward its
close. This slackening is due to a cessation of produc-
tion on the part of most individuals while nearly all show
at least a slackening of egg production towards the close
of the winter cycle. The exact date at which the cycle
ends varies with the individual, and may occur at almost
any point during the winter months including March.
Miss Curtis (714), in another connection, has published
the monthly records of a few hens that show this cycle.
They are shown in Table VII. With one exception, No.

TABLE VII

A PORTION OF TABLE XXV FROM CurTIS (’14) SHOWING THE WINTER Ece
PRODUCTION OF A NUMBER OF BARRED PLYMOUTH ROCK PULLETS. THE
DECREA

| Pullet Number

Month E AA | | | | PORE ES | |

| | | | | heed iet
a E N EE ET S Bhs | 10309 paces EM
November............ vee of 15119117 1518 5113| 6/12} 8} 1] 9
Dechert... i Sik oe: 125 18| 27 J Be 16/24) 8/14/15) 0 | 18

1911 |

Pas cn «Sy eo ine 1319 3 13| 4 alig 1 6 5 0 |14
February. ecesna eat |13 110/15 | 6/1 | | 811 5| 0 | 16
Winter totale 3 ear | 66 | 69 65 (57. | 56 | 48 | | 38 | Partes las 33 | 1 57

236, the birds all laid over 30 eggs. The evidence for a
winter cycle is shown by the depressed egg production in
January and February and is very clear. The records
published by Gowell, ’02, *03, also show this point.
Pearl and Surface, *11, describe the other periods as
follows:
hide next period (March, April and May) is the natural laying sea-
It corresponds to the egg-laying part of the natural reproductive
ads eri by the wild Gallus. . . . Naturally, therefore, a high
and a low variability in production are exactly what we find char-
acterizing the laying in each of the months of this period.
The third period (June, July and August) is characterized by a
gradually falling mean production and a variability gradually inereas-

212 THE AMERICAN NATURALIST [ Vou. LII

ing. . . . This is the period in which the rearing of the chickens nat-
urally occurs and it also represents an extension of the breeding season.

The fourth period (September and October) is not easily separated
from the third in respect to laying, but in general it is the period of
moulting. . . . It is characterized by reduced laying and marked in-
ereased variability.

It is not clear from their accounts whether or not they
consider that these periods extend through the second
year or whether they are to be considered as characteris-
tic solely of the pullet year. :

Just how far the data on Barred Plymouth Rocks are
applicable to our Rhode Island Reds is somewhat uncer-
tain. At the outset it is evident that the small per-
centage of birds that show an interruption in their winter
laying because of the presence of a broody period afford
no evidence either for or against the existence of a winter
cycle. Of the birds that do not go broody during the
winter two classes can be distinguished, viz., those that
show an interruption in their winter laying and those that
do not. In the 1913-14 flock and in the 1915-16 flock
from the original source, a large percentage of the birds
show no interruption in production, not even a slump in
the rate of production. Such birds lay at an approxi-
mately constant rate through the late fall and winter
months, and on through the spring. Among the records
of the main portion of the 1915-16 flock are many that
show an interruption or else a slackening of production.
Of these birds it can be said that they have a winter cycle.
But there are two points about these records that make
it difficult to interpret the interruption in production as
an index of a cycle. First, the interruption may occur
at almost any time during the winter followed by a re-
sumption of production in mid-winter, and second, some
individuals show more than one period of nonproduction. .
While there is definite evidence that a winter cycle exists
in some but not all Rhode Island Reds, the possibility
that some at least of these interruptions of production
may be due to environmental factors must be fully recog-
nized.

No. 615] EGG PRODUCTION 213

As already noted, there are many Rhode Island Reds
which show no sign of a winter cycle. Whether this
means that such birds do not have a winter cycle, or
whether it means that some other factor covers up an un-
derlying cycle is uncertain. In many instances these
birds are very much like the Barred Plymouth Rocks
noted by Pearl and Surface, *11, who state:

Many birds of course have no proper winter cycle at all. They begin
to lay for the first time in January or February and keep on laying
without any large break straight through the spring cyele.

But there are many other Rhode Island Reds that begin
to lay in October, November and December and lay through
the winter and spring without any breaks whatsoever.
Moreover, those birds that begin to lay in January or
February for the first time very rarely show any break
whatsoever. For these instances where the birds begin
to lay late in the winter it is conceivable that the winter
cycle might extend well into March but that the compara-
tively mild weather at that season of the year would tend
to eliminate the rest period and thus conceal the winter
cycle. But this argument cannot be applied to those in-
stances in which the laying is continuous from its start in
October, November or December right through the spring.
There seems no reason to speak of a winter cycle for this
class of Rhode Island Reds.*

The spring cycle, in Rhode Island Reds, in so far as it
can be differentiated from the winter period, differs
chiefly from that of the Barred Rocks in extending nearly
through June, since the end of June marks the point at.
which practically every bird that will go broody has be-
come broody at least once. Egg production is at its maxi-
mum at the beginning of this period due to active laying
on the part of practically all individuals but falls sharply

4 Later work on this point demonstrates, beyond doubt, that the presence
of a definite winter cycle is not characteristic of our strain of Rhode
Island Reds as a whole. There is some evidence, moreover, that the winter
cycle is a Mendelian recessive, the dominant allelomorph being continuous
winter production. For details see Goodale, ’18.

214 THE AMERICAN NATURALIST [ Vou. LII

after its middle, the rate of decline being much greater
than for the Barred Plymouth Rocks.

The summer period may be considered to be July,
August and September. Practically all the birds of the
flock are laying in broody cycles with egg production re-
maining at approximately a constant level, while consid-
erable partial moulting is going on. It passes gradually
into the fall period which is characterized chiefly by the
cessation of egg production—usually coinciding with a
broody period—and the onset of the fall moult together
with some slowing in rate of production of those birds
that are laying. Biologically the fall period overlaps the
calendar year since it may extend into December. The
question of egg production in the fall, at the end of the
pullet year, is on rather a different basis from that of the
other seasons of the year. The egg producing mechanism
of the hen seems to be in a peculiarly unstable condition
and unless great care is exercised may cease functioning
in response to slight adversities in environment. Some
hens, however, and this is really the important point, con-
tinue to lay throughout the fall months with the same
regularity they exhibited in the spring. This affords us
an opportunity to build up a strain of birds that will be
persistent layers throughout the year.

Thus, persistency in egg production through the fall
months enters in as a factor in determining the total egg
production as has also been emphasized by Rice (714).
Some birds cease laying relatively early in the fall, say
late in August or September; others, however, of the
‘same age, breeding, and under the same conditions, con-
tinue to lay all through the fall, at approximately the
same rate of production as during the summer months,
although the rate may fall off slightly. Now, these birds
will have quite a different total record from those that
stop early in the season and if one examines his records
he finds that while many persistent layers are also good
producers early in the season, nevertheless a great many
of the birds that were good layers during the winter are

No. 615] EGG PRODUCTION 215

not persistent layers during the fall, while some birds
with low winter records are good fall layers.

Other kinds of cycles, which are described in the fol-
lowing paragraphs, have been noted in the Rhode Island
Reds. In broody individuals, where the practice is fol-
lowed of ‘‘breaking up’’ the hen, a series of cycles is in-
troduced, marked by broody periods alternating with an
egg production period. If the natural course of events is
not interfered with in the case of broody hens, the broody
period lasts until the chicks have hatched. Then the
period of rearing takes place. Toward the end of this
period the hen begins to lay and the cycle is repeated.

There is also a short time cycle of one or two weeks
which needs little discussion at present. It is shown by
an acceleration in rate of egg production followed by a
decline somewhat as follows: Egg—blank—egg—egg—
blank — egg — egg — egg — blank — egg—egg—egg—egg—
blank — egg — egg — egg — egg — blank — egg —egg—egg—
blank— egg—egg—blank—egg—blank— egg—blank—egg
—blank—and repeat. Although one often finds records
that approximate closely the above scheme, their rarity
suggests either that the fluctuations in rate are due di-
rectly to some extraneous circumstance or, if there is a
fundamental rhythm of this sort, that it is subject to dis-
turbance from the environment. Whatever may be the
cause, at present it is doubtful if it is indicative of an
internal factor.

The next type of cycle is that exhibited by certain hens
which lay at a relatively high rate for a time and then
stop. This period may correspond to the term often used
by poultrymen when speaking of a hen’s clutch. Here
again we are confronted by doubt as to the causation
= of these blanks, for many hens do not have such pauses
in production. Once they begin laying they continue
without any considerable vacant period (not exceeding
three or four days) until the onset of the first broody
period.

Stamina.—A strong bird is readily distinguished from
a weak one, but it is difficult to separate the birds per-

216 THE AMERICAN NATURALIST [ Vou. LIT

manently according to a definite standard since it is im-
possible to secure a constant environment. A fairly uni-
form environment in the sense that all birds are exposed
to the same external conditions at any one moment is
fairly easily secured; but since the external conditions,
particularly weather conditions, are very variable and
follow no definite course, and since a bird’s vigor is a
resultant of its own inherent strength of resistance
against the environment, it is clear that the objective
vitality observed in each member of a flock may be un-
equally affected by the surroundings.

The evidence available on the relation of vitality to
fecundity thus far points in two more or less opposite
directions. Many birds of low vitality have made not
only excellent but even high records. On the other hand,
birds of strong vitality may make low records. At the
same time there is a point at which the vitality becomes
too low for good egg production. In the fall of 1913,
thirty-eight birds graded early in the fall, before laying
commenced, as ‘‘poor’’ in respect to vigor, were put in
the laying houses. They had an average record for the
winter period of only 20 eggs against an average of 38
for the entire flock, including the poor birds. Low vital-
ity evidently depressed egg production in this instance;
mainly, through retardation of the time of commence-
ment of laying rather than by slow production. The in-
fluence of lack of vigor on winter egg production is shown
in Fig. 9, where the curve of winter egg production for
the entire flock is represented by the continuous line and
that for the ‘‘poor’’ birds by the dotted line.

Occasionally, birds of low vitality may make excellent
egg records. In one family in particular, the birds were
of distinctly mediocre quality, as evidenced by their
weight, activity, hatching quality of eggs and viability of
chicks and yet they were able to make high records, the
average for the family of seven individuals being 63.3—
ranging from 33 to 81—for the winter period, with a
yearly average of 192.4 for the five birds that survived
throughout the year, and with a range of 154 to 210.

No. 615] EGG PRODUCTION 217

Moult.—Moulting exercises some influence on the num-
ber of eggs produced, since birds that are moulting often
do not lay, particularly during the fall moult. Moulting
itself may be induced at certain seasons of the year by
changes in management, especially those changes that
tend to stop egg production. Such changes apparently
change the course of the metabolism of the bird. Brood-
iness in late summer and early fall appears to be a com-
mon cause of the onset of a moult and consequent cessa-

PERCENT

ornata SSRS

0-4 59 tx 2509
EGGS

Fic. 9. The effect of low vitality on winter egg production. The graph
shows the percentage of the flock of 1915-16 laying the specified number of eggs.
The continuous line is for the entire flock. The dotted line is for that portion
of the flock graded “ poor” or of low vitality, before being placed in the laying

tion of egg production. At least one might draw this
conclusion from the fact that egg production, as the usual
rule, ceases with a broody period, for in most instances
the last egg laid in late summer or early fall coincides
with the beginning of a broody period. It is not clear,
however, whether the moult starts because the bird has
reached the limit of her production period, or whether
the moult begins because of the interruption to egg pro-
duction due to the onset of the broody period.

In the Rhode Island Reds we have observed a partial
moult that begins in the early part of the summer and as
a rule seems to affect egg production very little. In the
autumn this partial moult is followed by a more extensive
(often complete) moult attended by cessation of produc-
tion. It is possible that the summer moult has been in-

218 THE AMERICAN NATURALIST [ Vou. LIT

duced by broodiness, but that the hens are broken up so
quickly and the impulse toward resumption of egg pro-
duction at this season is so strong that it inhibits the
moult at various stages.

Pullets that begin to lay very early in the fall very
often undergo a moult during the latter part of the same
fall. It is not clear that early production of itself tends
to induce the moult so long as the birds affected are not
hatched too early in the season. The moult is more com-
monly observed in pullets that are hatched very early in
the season and which begin.to lay in August and Septem-
ber. Such birds rarely make a continuous record but in-
stead stop laying after producing a variable number of
eggs and moult much like birds from fifteen to eighteen
months of age. In the flocks with which we have been
dealing the variability in maturity has induced some com-
plications in handling the flocks. If an attempt is made
to hatch birds sufficiently early in the season, so that a
good share of them will begin laying in November, some
begin too early, lay a while, and then moult.

Rate and Rhythm of Production.—The percentage rate
of production at any time and for any period may be
taken as the number of eggs times 100, divided by the

length of the period involved, measured in days. Rate

is an important factor in determining egg production,
but in the Rhode Island Reds is of quite secondary im-
portance as compared to date of first egg (and of course
age at first egg).

The distribution of the percentage oe calculated for
the winter period (between the first egg of each pullet
and March 1) for the flock of 1913-14 has been deter-
mined and the graph shown in Fig. 10 plotted. The
curve shows a considerable homogeneity of rate in the
flock. Since the general trend of events, such as acci-
dents, temporary ailments, etc., is of such a nature that
some birds do not attain their natural inherent rate, the
curve shows a very gradual slope at the left-hand side up
to about 30 per cent., a somewhat more rapid rise be-

No. 615] _ EGG PRODUCTION 219

tween 30 and 50 per cent. and a much sharper rise beyond
50 per cent. The mode comes at the 61 to 70 per cent.
group, while the maximum rate does not exceed 90 per
cent. It is interesting to note that 8.5 per cent. of the

30

INDIVIDUALS IN PERCENTS
hs bs
== S

a

0
1-10 41-20 41 -J0 yl -40 41- $0 s1- 60 él - 70 7I - 30 81 - 90
RATE IN PERCENTS.

Fic. 10. Percentage rate of production for the winter period. The per-
centage of the flock laying at the specified rate is shown by the ordinates, the
rate by the abscisse. Flock of 1913-14. M=62.50, S. D. = LV. + 26.23.

flock laid at a rate exceeding 80 per cent. for the entire
winter period.

The effect of variability in rate on total production for
a definite period is shown by the records given in Fig. 11.

25

S 5 8

INDIVIDUALS IN PERCENTS

a

0
110 1-20 21-80 31-40 41-50 51-60 61-70 71-80 81-90 91-100
RATE IN PERCENTS
_ FIG, 10a. Percentage rate of production for the winter period. The per-
centage of the flock laying at the specified rate is shown by the ordinates, the
rate by the abscisse. Flock of 1915-16. M=54.59, S. D. = 18.52, C. V. = 33.93.

[ Vou. LIT

THE AMERICAN NATURALIST

220

‘aSUas §,[1vaq Up Joonpoad aad0]paut Y oq 01 WS PINOM IS8U] SIUL “97va
MOIS Y 38 pue A[}U9}}{UII0} UT PIV] Jey} PNA V ST OSOG “ON IMYA ‘uoljonpoad JO 9381 MOIS JOYIVA V JV Ápsnonupquos pre] 3803 39 ¡nd e Jo eq}
SI JGLP ‘ON ‘OBI qu e 38 Apsnonurjuos prer WU} PPNA V JO 3811 SI Gp “ON ‘a8u OBLIGAR əy} 189U paanjuU siap¡nd 99191 [Ty “UMOUS OSTE
aie [dy pue IIe 107 spiod94 oy} y3noyppe “uorjonpold 339 id aq} 0) oduslajor peppads YIM spiod9l 339 Apep Jo sədAL ‘It “DIM

AAA AAA
2000000 1a na al Papas ana DORIA”
El a, Ade rro en lo e dp
er CPA ee ee eee Lode ieee st
er] | et te ee ae ee T o e
Jas LA dada rr
seo | Te os |e dE 260 ss y 0606161000 lst alo ra jas ajn oro el JE j i les
es SAV] 9€% “DIH LSU LV Ay “GIGI ‘TZ HOUVIA ATHOLVA : L6LE (ON
"nm i T ly ili SAA O A ase
sedal lla el ri lo
“NN, E PLA dl Pi Di PP APN <A
A SAME ADO ELO E a LLAT] LIN l Litli: *
a eee eS A AO ee. SET ee eee! C o d
ROn 202r pele oe AON
soL | te | 0816383 | 43 196! se |v3 8633 1203/61 81/21 9t/ct | FT) eT) St] It/or] 6/8) ¿[9 8) F)e) se) T mosto: a :

SAVA SPZ ‘OOM ISU LV GY ‘IGT ‘9T HOUVIA AAHOLVH 6h ON

No. 615] EGG PRODUCTION 221

Number 49 laid at an exceptionally high rate. On the
other hand No. 4797 laid at a rather low rate for a bird
that produced eggs continually throughout the winter.
Although she began laying several days earlier in the
year than No. 49, she produced 24 eggs less, due to her
slow but steady rate of production. An entirely different
type of rate is shown by No. 5080. This bird is a true
mediocre producer in so far as may be judged from her
record. She laid her first egg at a fairly early date and
fairly early in the season but her rate of production was
very low. The eggs, too, were produced at haphazard in-
tervals. Her record is to be compared with that of No.
4568 (Fig. 4), which laid the same number of eggs but all
in the last part of February. Records similar to those of
No. 5080 are shown by Nos. 274 and 280 (Fig. 12), both
Barred Plymouth Rocks in one of the contest pens at the
Essex County Agricultural School. Another type of
rate is shown by No. 4815 (also Fig. 12), which exhibits
also a well-defined winter cycle. This bird matured early,
began laying early in the season, and laid well for about
six weeks. Then she stopped entirely and did not lay
at all again until late in February.

While rate may be considered independently of the
rhythm, i. e., rhythm may be ignored; rhythm can not be
considered apart from rate. A hen may lay twelve eggs
in a month and the rate be described as such or as 40 per
cent. without paying any attention to the sequence in
which the eggs appear, but if the rhythm be considered,
attention must be paid to the sequence between eggs.
Thus, two blank days may repeatedly alternate with one
blank day between eggs, producing a regular rhythm, or
the twelve eggs may be laid in groups of two, three, or
more eggs on successive days and then a considerable
period of blank days intervene, the rhythm in this in-
stance being irregular. A certain degree of regularity
of rhythm is closely associated with high rate of produc-
tion, though some birds of relatively low rate lay in a
regular rhythm (Fig. 11). Most low-record birds, how-

[ Vou. LII

THE AMERICAN NATURALIST

222

e “9 % ‘ported 1891 19JU]Mm JUO, Y SMOYS ‘poy PULISI IPOH Y ‘eTSF JO “o * “pio991 PAYA IUL
he Cssey_) Ájuno;) xessq ay] Aq pajonpuo0d 4sejU0D IYJ UF (PRG PUT PLZ SON) SHPON yous] POLIVL OM} JO spaooaa AIMA ‘GL “PLA

"saoonpold aad0;paur 918 LJL

“91949 197U]M

*1900498 penya

| /

l

dy

o A

oa

st
|

pie JY

Q

Et ORCT

0

an

|
f
|
ab

vas

1) sister
E

NMONUN() ‘DD ISU LV ADY ‘GIGI AAHOLVH

G popmouop)

"TT ‘SIA

PL3 ‘ON

/

pes “dy

r]
EN

Sn.

/ A

o

po oe r

~

Tal

..... "AON

s]830.L

Te ale

El
A
|
|
|
E
si
96! 96
i

vz

|

it SS

Ez TG | 06 pe

se
o
l
|
eel
i

ia | el | ct

o

| | |

N

1 | 91=st61

awa

SAYA FFZ ‘DOW LSUIJ LY apy

15

‘GIGI ‘8% HOUYI CHHOLVH

0809 ʻ'ON

EGG PRODUCTION

No. 615]

Cpapnjou0p) "Zt 'ÐIA

ii We TT Tin] IE j ae: Pay
ot} 7] lr lla ll [sl lr lrlal Trial] [el in] es Plis
6£ B g | |
9 i / I / J / -
a | ena” pur
od: E | l / IIRI |! El! / / RE SR
o /|/|/|/ (IN HAT a hid Fee HL A ON
|
s1840, | TE | OF hy 23193 |93 +3 |€e3 23 (13103 |61181/21 191 | St!) FI) et] st! 111011 6) 8 1 pr ay
SAV 23% 'DOH ISU LY ADV “GIGI ‘IZ HOUVIA GAHOLVH CIS ‘ON
FULLE : / BREE. T, / o aN
er | | / Pit i |] BE! Peak M pi PER
Z
4 F E / l S AR
w A E ETE Bac eee hy, / ao aes
€ / / / = ae
0 te ON
S1830. | TE | O8 | 68 | 8Z | 23 | 98 | SZ | FZ | ET 33 |12 [0% 61 81/21 91/91 VI ¡€T[Z1 [11 01 6/8 I oe
NMONMUN/) ‘ODT LSUIJ LY 49V ‘GIGI CAHOLVH P8SZ ‘ON

224 THE AMERICAN NATURALIST [ Vor. LII

ever, have an irregular rhythm. While the observed
rhythm is rarely entirely regular, yet each hen tends to
produce eggs according to a rhythm that is character-
istic to a certain limited degree for that individual.
Superimposed on the daily rhythm are evidences of other
rhythms having a beat measured by months or years.
Although from another standpoint egg production is a
more or less continuous process, at least in so far as the
growth of the yolk is concerned, even though the fluc-
tuations in the activity of the albumen and shell glands
may be more pronounced, the rhythm may be considered
as it appears on the record sheets, 1. e., the rhythm
shown in the deposition of the eggs. Various types of
rhythm have been observed. If, as a working basis, it
be assumed that an egg a day represents a standard
rhythm, although this high rhythm is rarely reached for
extended periods, it will be found that some hens lay
every other day, or we may say a one half rhythm, others
two thirds, i. e., two days out of three, others three
fourths and so on. The three fourths or four fifths
rhythm is common among most good layers. Occasion-
ally the series may be repeated without the intervention
of a zero day as is shown by the time of day the eggs are
laid, for if the time of day at which the eggs are collected
from the trapnests is recorded, the rhythm is shown even
better than by the daily records. It has been our prac-
tise to visit the nests about every hour and a half and to
record the time the eggs were gathered in units of a half
hour. Thus the approximate time that each hen drops
her egg is known. Rhythm, as shown by the time of day
a hen lays her eggs, is an index of the essential con-
tinuity of the activities of the ovary in the growth of the
ova, rather than of any rhythm in its activity, since the
interval between eggs is more uniform than when the
daily record only is used. One may perhaps infer that
there is some rhythm in the activities of the oviduct since
it is known that the stimuli for its activity comes from

the presence of the yolk in its lumen. However this may

No. 615] EGG PRODUCTION 225

be, the several types of rhythm shown by the daily records
are found to depend very much on the time of day that
the eggs are laid.

None of the various types of rhythm, i. e., one half,
three fourths, etc., are characteristic of any one hen, al-
though many individuals seem to center about a par-
ticular rhythm, e. g., two thirds. A bird with this rhythm
may fall to the one half type but does not often, except
in the spring, exceed the three fourths type. While little
stress can be laid on this point, it is interesting to note
this tendency toward a definite rhythm in some individ-
uals. But aside from these considerations, birds of the
same age, which begin to lay at approximately the same
time and which do not become broody, do not lay with
the same rhythm: Thus, of two full sisters, hatched the
same day, one laid only about every other day, while the
second laid about five days out of six. The rhythm, then,
is quite an important factor in determining the number
of eggs laid.

Various causes may interfere with the normal rhythm,
such as causes that interfere with the formation and
growth of the egg, and other causes such as environ-
ment, season, method of management and internal factors
such as broodiness. In many birds evidences of a rhythm
with a period of some length may be noted as shown on
page 215. The example given is of course idealized but
actual records of nearly the same type may be observed.

In this connection the question of the nonproductive
periods, usually of short duration, that occur in some
records and which produce irregularities in the rhythm
may be discussed. As will be pointed out later, brood-
iness is responsible for some of these periods. A similar
period, that may be noted in some hens’ records during
the winter, may be taken as an index of the existence of a
winter cycle. But other individuals may have two or
more such periods or may have a single period fairly
early in the winter (No. 4529, Fig. 13). Some of these
periods may be inherent in the individual’s makeup but

Harcuep Marcu 7, 1915. Ace ar First Eca, 214 Days

| |
alada e167 |B | 8 | 10) 11 | 12 | 18 | 14) 15) 16 | 17) 18 | 19 | 20 ae 24 | 25 | 26 aa 29 | 30 | 31 | tals
MES : O ER Dae A O ee ee. nes
Oe ee E A A A ee LE
(761759 | 9 1105) 9 | 9 | 9/95] 1 | [7.517.5|7.5| 9 |45|- 1105] o
| | | Qe E AS A ASA ee ASR O
| 2 o Teo ae E ==> ae oes oes

me lsfo lo mx pon lo} ad. 75 [105] 3 |,N, 105 3
7.5 [10.5 3. 7.5 |10.5| 3 ie a | de 9 | 9 pom 3 | | E 105 7519 105 9 105 f Pp a.
| og, LA A A
: N N | N wal N
10 [10.5 Ni N1 N1 z | | | | No| N3 MR kale T e aa NF] [No 9 [19
8 E [2 |N N9 | N9 | N9 9 ba 9 5405) 3 | (7.5) 9 | 10.5|10.5) 9 105 9 [10.5 10.5| N3

e ae Marcu 28, 1915. Aem at Frrst EGG —

Da | | | | | | | | | | | o-
Me- 1 2.14 415 6 7 810 +10] 11/1211 ae eo ce 20 | 21 aoe 24 ag 27 | 28 Pie oe
o (ages ACA IES
MS a aida al A
ae A a A ara
Nk Net N. EN

ee eet eee PA ABC
Me OI ile MA a e alo
— o 0011213 | E PAR E & FAAR AERA 3 | |75 qe Se alo loss E Mo

Types of Records.

Fie. No, 4529 is oe of an early maturing pullet that laid discontinuously but ohne made a fairly good
winter reco ord, Nos, 5032 and 246 are nesters, i.
lected or the hen ready to leave the nest is otal by the numerals,

birds that visit a ne pao regularly but do not lay.
ep” used for the half-hour period.

e at which the egg was col-

Hatcuep Marcu 30, 1913. Aem at First EGG —

saan eae |
Date | - To-
1913-14 | 1 | 2} 3 | 4) 5) 6/7 | 8 | 9 | 10} 11 | 12] 13 | 14] 15 | 16 | 17| 18|19|20|21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | ais
Nov... = | EEE EE mo
j Dec. PE err ee E | | | | | 0
Is N NIN N | N N | N NIN 0
S 3 11.5 9.5 | 3 95| 1 11.5 3.5 | 1.5 158
N N NN | | | N
Feb 41 1 11.5 1.5 913 | | | 4 ES A
M A NN |N N N|N|N N|N|N UN N N NIN |,
e a aed N2 8 |11.5|3.5 10.5] 2 whats 9 | 1.6/4.5 (9.5 [11.5 | 2.5 9 |10.5| 3.5 8.5 | 10.5
ac N N oa N N > N NINININ NIN at g | o
2 10.5 [Na $ 3-5 10.5 11.5 | 3 5 10.5|10.5| 2 8.5] 11 | 2 7.5| 10 |115| 11| 2 | 2 7.5} 10/11] 2 (Si $
M NÍNININ N N NIN N N1 NININININ NINININ nININ o!
a 85|10|10| 2 | 2 8.5 2 8.5 | 10 yo | N3 8.5 | 10 (11.5/11.5| 3.5 9.5|10| 2 | 2 8 (115| 2 |
de N N NiNIN N N N N | Nin |N N |N o
ne 8 | 10 |11.5| 1.5 10 | 10 | 10 (11.5) 1.5 1.5 2 8.5 110.5; 2 10 |11.5) 2 | 8.5 | ¿q |115] 1.5 8 | 10 |
Jul N |[N |N N 5 N ol
oe 10.5 /11.5| 2 | 3 8.5 110.5| 2 7.5 Na 1.5 |
Noo ee | E
re 12 N|N N N N N N N N N N | N d
o 8.5 | N li| 4 8.5 | 11 | 2.5 85|10| 4 85/12 | 2 185) 11 [10.511.5| 4 85/11 115 | 10 | 11.5) 3.5
2.5 |
NIN N N N N 85|N |N NIN N | | 0
9 |11.5,3.5 10 | 1.5 10 |2.5| 4 851121 3 10 | 4 Nj 1 [35 9.5 | 1.5 11.5] 4 pa
11.5
N, Pr N | | | 0
11 Ni 11.5 1.5 |

Fig. 13, (Concluded.) Years' Total 0

228 THE AMERICAN NATURALIST [ Vor. LII

others are probably the result of the environment, since
it is well known that nonproductive periods can be in-
duced by artificial means.

One of the most interesting things in connection with
the rhythm of egg production as observed by Pearl (712)
is the existence of hens which never lay an egg, but the
record of whose visits to the nests shows a very definite
rhythm. A number of such hens have appeared in our
flocks. (Nos. 5032 and 246, Fig. 13.) The hour of their
visits to the nest exhibits exactly the same sort of rhythm
as normal hens. These facts point strongly to the ex-
istence of some mechanism other than the formation and
deposition of an egg which controls the extrusion of the
egg. It is interesting to note that if one of these hens is
removed from the nest before she is ready to leave, she
returns and persists in doing so until, shall we say, she
thinks she has laid her egg. Autopsies of several of
these hens show that they fall into two classes; viz., those
that are producing yolks or eggs but depositing them in
the abdominal cavity, and those in which a tumor of the
reproductive system is involved.

Laying hens often visit the nest at the proper day and
hour but fail to lay. Such hens (No. 4529, Fig. 13) usu-
ally lay the day previous and the day after in regular
routine, though at times they may pay two or more such
nonproductive visits in succession.

A study of these latter records shows that some hens
have indications of a potential capacity greater than the
actual capacity revealed in the records. Very many hens
pay an occasional visit to the trap nest without laying
(note the n’s in the various records), while a few pay
such visits more or less regularly, at various periods of
their lives. The striking feature of these visits is that
they are made at the hours one would expect if an egg
were actually laid (No. 4529, Fig. 13), though the nature
of the stimulus that causes such visits is uncertain.

Broodiness.—Broodiness, from the commercial as well
as biological standpoint, is one of the most important of

No. 615] EGG PRODUCTION 229

the factors influencing egg production. In general, with
the onset of the first broody period, the monthly pro-
duction falls off 40 per cent. of its former rate. Brood-
iness, however, as met with in the laying hen, is to some
extent an artificial condition. In a free state a hen be-
comes broody after she has laid a clutch of eggs, incubates
them, and rears a brood of chickens. Altogether she is
not producing eggs for a period of some ten weeks or
more. After this she may again lay a clutch and repeat
the process. Egg production under such conditions re-
mains at a relatively low ebb. It is a matter of common
knowledge among poultry keepers, however, that by
various means the broody hen can be ‘‘broken up.” That
is, she can be induced to discontinue manifestations of
broodiness and after a period varying from a few days
to several weeks will begin to lay again. As a rule, how-
ever, only a few—ten or twelve—eggs are laid before a
hen goes broody again. The process may be repeated
indefinitely.

There is a considerable variation in the number of
times a hen goes broody in a year, the length of the broody
periods, the trouble required to break her up and other
characteristics of broodiness.

The age at which the first broody period occurs de-
pends in part upon the time a hen begins to lay. In the
vast majority of instances egg production precedes brood-
iness. At least 15 to 20 eggs are laid before a hen be-
comes broody, though it may be many times that num-
ber. Age incidence in the first place depends upon the
age at which the hen lays her first egg but after that it
depends upon other circumstances, which have not been
determined. Thus, the age of a bird at her first broody
period may vary from eight months up to the end of the
second year. Usually, however, the first broody period
comes on when the bird is from 11 to 15 months of age.
After the first broody period, the periods recur about
once a month, if the hen is promptly broken up. There
are records in our files of a few Rhode Island Red hens

230 THE AMERICAN NATURALIST [ Vou. LIT

that have not been broody for from one to three laying
years.®

The number of broody periods per year, then, depends
upon the date of the first period and in the second place
upon the cessation of production in the fall. Egg pro-
duction usually, but not always, ceases with a broody
period.

A broody period has two phases. First, the period of
manifestations of broodiness such as clucking, ruffling
of feathers and cessation of production. This period is
variable, some hens being easy to ‘‘break up’’ while
others are very difficult. Manifestations of broodiness
sometimes begin several days before egg production
ceases, and may rarely continue without cessation of egg
production or without hanging to the nest. I do not re-
call a case when egg production began before the cessa-
tion of the manifestations of broodiness.

The second phase begins with the disappearance of the
manifestations of broodiness, and extends up to the time
egg laying recommences. Its chief characteristic is non-
productiveness and its length varies considerably.

Broody periods coming during the height of produc-
tion, March and April, are usually of short duration, but
gradually lengthen as the summer advances, until they
sometimes last for a month or more. During the winter
months, the periods are longer than those occurring
during the spring and often lead to the cessation of egg
production for several months.

Egg production is affected by broodiness chiefly
through the number of broody periods. Hens that go
broody many times during the year have a much lower
production than others that go broody only two or three
times, other things being equal. It is of particular in-
terest to note the abrupt way in which the monthly egg
production usually decreases with the onset of brood-
iness, regardless of the time of the year. Thus, a hen

5 The statements in this section are based on an intensive study of broodi-
ness, the data on which will be published in due course of time.

No. 615] EGG PRODUCTION 231

may be laying at a 75 per cent. rate before going broody,
but with the appearance of the first broody period pro-
duction falls off 40 per cent. of its former rate. In gen-
eral it has been found that for each hen the rate for the
broody part of the year is only about 60 per cent. of the
rate of the nonbroody part.

It is not known whether the intense development of
broodiness in the summer months is directly due to the
weather conditions as such or whether it is due to some
internal cause or is part of the annual cycle. At any rate
it is evident that it operates to decrease the egg produc-
tion very considerably. Further discussion will be post-
poned until the study of broodiness now in progress is
ready for publication.

Types of Winter Records.—The various factors de-
scribed in the foregoing pages combine in many ways and
produce as a result different types of records, several of
which may now be discussed in more detail. For the
present we may divide the various types of records into
high (over 30 eggs), mediocre (under 30 eggs) and zero
producers.

High producers (over 30 eggs) may be divided into
several subclasses. First, the early maturing, nonbroody
high that lays continuously at a high rate and makes a
very high record (No. 4846, Fig. 3). Second, the late
maturing nonbroody high that lays continuously at a high
rate but makes a lower record than the first in direct pro-
portion to the difference in maturity. (See Figs. 3 and
4.) Third, the broody, early maturing high that lays at
a high or fairly high rate'during the laying periods (not
shown). Such a bird’s record is cut very materially by
the broody periods. Individuals of this type are not very
numerous during the winter period. Fourth, there is the
high bird that exhibits a pronounced winter cycle or
period of good production during the early part of the
winter, but which stops after a time and may not lay at
all for several weeks. This type is closely related to the
bird that lays her eggs in clutches, but because of her

232 THE AMERICAN NATURALIST [ Von. LIT

early start makes a comparatively high record. Finally
there is the type of bird shown by No. 4797, Fig. 11, that
matures early, lays steadily and does not go broody but
lays at a comparatively slow rate. Such birds may make
high records, but they never make the highest ones.

Mediocre producers (under thirty eggs) may result
from any one of the various types previously described
for high producers through the failure of one or more fac-
tors. Thus, delayed maturity will cause a bird to fali
below the dividing line at 30 eggs to a degree that will
vary inversely with the age at first egg, due allowances
being made for the date at which the individual was
hatched (Fig. 4). Or a bird may fall below the required
number of eggs through a slow rate of production (No.
5080, Fig. 11, or Nos. 224 or 284, Fig. 12). The former
type of bird (Fig. 4) would seem to be a late maturing
high, since it is clear that its record results directly from
the variability in time of first egg. Hence this type of
mediocre producer is to be regarded as an artificial class
rather than a real class as in the case of the birds typified
by No. 5080.

Zero producers, by definition, are birds that do not
lay until after March 1, and need no further discussion,
except to note that some of them clearly result from the
combined effects of date of hatch and age at first egg
rather than from an inherent inability to lay during the
winter (i. e., from a lack of the winter cycle).

There are, then, numerous types of records resulting
from the interaction of the various components described
in the earlier part of the paper. Although the records
described are winter records, the observations apply
equally to annual egg production. High egg production
results only from a combination of the right set of fac-
tors. If any one of several of these factors fail, egg pro-
duction is lowered.

(To be concluded.)

BACTERIAL PHYLOGENY AS INDICATED BY
MODERN TYPES!

DR. R. E. BUCHANAN

Iowa STATE COLLEGE

THE importance of the group we call bacteria in any
theories concerning the origin and evolution of life on our
planet is well shown by several recent writers on the sub-
ject, notably Jensen (1909), Osborn (1916), and Kligler
(1917). Tn each case, however, there are certain misin-
terpretations of our knowledge of the modern bacteria
and their function which go far to invalidate, or at least
to weaken, the specific conelusions which they reach with
reference to the types of primitive bacteria.

In our search for hints as to ancestral types by in-
vestigation of modern species of bacteria we must hold in
mind that although present-day bacteria approach most
closely to what we conceive must have been primitive life,
nevertheless and for this very reason the group of modern
bacteria must have the longest evolutionary history of
any existing group. That any modern species closely
resembles the original type is therefore not extremely
probable. It can not be too often emphasized that in
speculations concerning evolutionary history based upon
modern forms with no adequate fossil ancestral types we
must deal only with the tips of the ultimate twigs of the
branches of the evolutionary tree. By a careful com-
parison of the surviving forms we may gain a knowledge
of their probable relationships, but it should be remem-
bered that in no case this relationship is that of parents
and offspring, but that of brothers and cousins. Perhaps
we may speculate upon the probable arrangement of the
branches of the evolutionary tree that have disappeared

1 From the Bacteriological Laboratories.

233

234 THE AMERICAN NATURALIST [ Vou. LII

in past geologic ages by study of the survivors. but it is
self-evident that there should be a perfect knowledge of
these survivors, morphological and physiological, so that
we may not be led astray by superficial resemblances
when there exist, in fact, deep-seated and fundamental
distinctions.

The geologic evidence that has been adduced as to the
character of the primitive bacteria i is of but the slightest
value. Speculation as to the primitiveness of nitrogen
fixers, for example, based upon the geologic evidence
introduced is scarcely convincing.

It should also be noted that it is quite possible that the

bacteria do not constitute an homogeneous group in the
sense that all are descended from a single primitive type
of bacterium. It may be that there have been included
in the group bacteria forms which have assumed similar
morphological or physiological characters without having
a common ancestry. Botanists, for example, at the
present day are by no means convinced that seed plants
have all had a common origin; in other words, the ability
to produce seeds may have arisen independently in two
or more groups of the fern plants. It is possible that
some of the forms we term bacteria have been derived
from the fungi, others from the blue-green alge or pos-
sibly some even from the protozoa. In short, it may be
that the actual relationships existing between various
bacteria may be very distant.
- A study of these modern bacteria will reveal relation-
ships such as those just indicated. The possibility that
the bacteria are a derived group must be constantly held
in mind. To prove that they are primitive it must be
shown that no group from which they might have sprung
or to which they seem to be related can be regarded as
more primitive. This has not been satisfactorily accom-
plished in certain cases.

Modern systematic bacteriologists are in fair agree-
ment that there should be recognized five or six distinct
groups or orders of bacteria. Of these, the Eubacteriales,

No. 615] BACTERIAL PHYLOGENY 235

or true bacteria, are generally regarded as the least
specialized and possibly the most primitive. It is pos-
sible that the great group of the sulphur bacteria, the
order Thiobacteriales, is equally primitive, though the
genera and species have been less studied and are not as
well known. There are unquestionably many intergrad-
ing forms between the Eubacteriales and the Thiobactert-
ales, as shown by close parallel séries of morphological
types. It seems equally clear that there are intergrading
forms between the Thiobacteriales and the blue-green
alge. Morphologically, too, one may find every grada-
tion between the typical colorless, sulphur-containing
Beggiatoa, through species of this genus showing bac-
teriopurpurin, through the faintly colored, slender Oscil-
latoria to the thick, deeply pigmented forms. If these
intergradations and indicated relationships are real, it is
apparent that the true bacteria may have come from the
blue-green alge through the sulphur forms, or the blue-
greens may have come from the true bacteria, or the
sulphur forms may be closer than either of the other
groups to the primitive types from which all three groups
have been derived. While there is no definite proof ap-
parently possible at the present time, it is not at all im-
probable that the last assumption is the true one. A
relationship quite certainly exists between the group of
sheathed filamentous bacteria (the Chlamydobacteriales)
and the blue-green alge. The resemblance is so well
marked that certain species of the iron bacteria are quite
commonly included by algologists among the alge. The
relationship to the Eubacteriales is not quite so clear.
Possibly the genus Spherotilus (Cladothrix) may be re-
garded as a link, for this organism consists of rod-shaped
cells occurring in chains, all embedded in a gelatinous
sheath. Motile cells (gonidia) with polar flagella some-
what resembling Pseudomonas types may be developed.

- The fungi apparently are related to certain of the bac-
teria through the Actinomycetales. This latter group
has some resemblance to certain of the true bacteria such

236 THE AMERICAN NATURALIST [ Von. LII

as the Lactobacillus of sour milk and the diphtheria types.
It is possible that these organisms together with a few
genera from the Eubacteriales represent an entirely dis-
tinct series. The Spirochetales apparently constitute a
group showing combinations of characters which relate
them to the Eubacteriales and the Thiobacteriales on the
one hand, and the Protozoa on the other. The group
Myzxobacteriales is apparently related to the true bac-
teria, but not to higher groups of plants or animals, unless
there may be some as yet undiscovered relationship be-
tween these forms and the slime molds as suggested by a
superficial study of their fruiting forms.

The interrelationships just discussed among the vari-
ous great groups of bacteria may be illustrated by the
following diagram in which the connecting lines are in-
tended to show relationship, but not necessarily deriva-
tion.

NO BLUÉ=GREEN ALGAE /

FUNG! PROTOZOA

Actinomycerales )
re a

7 SUNE MOLDS qa

Fic. 1. CHART ILLUSTRATING THE PROB/ BLE INTERRELATIONSHIPS OF THE
GREAT GROUPS OF BACTERIA AND THEIR RELATIONSHIPS TO OTHER GROUPS, as
Fungi, Blue-green Algae and Protozoa

From the standpoint of the student of evolution the
order Eubacteriales (possibly also Thiobacteriales) is of
special interest, for within it is probably to be found
greater variation in physiological activity than in any
other group of plants or animals. A comparison of the
modern forms belonging to this group may well give some
hint as to their evolution. Too much can not be expected,
however, without getting far into the realms of specula-
tion.

After rather careful consideration a committee of the
Society of American Bacteriologists has proposed a list

No. 615] BACTERIAL PHYLOGENY 237

of names to be recognized as valid for the genera of this
order. They have also suggested that these genera be
grouped in seven families. Altogether twenty-eight
genera are recognized. It should be possible, if adequate
knowledge is at hand, and the Eubacteriales constitute
an homogeneous group, so to arrange these genera as to
show their probable and their possible relationships, and
perhaps gain some knowledge thereby of their origin and
evolution.

From the standpoint of the evolution of bacteria we
are much interested in the organisms which can live and
grow in the total absence of organic matter, those which
utilize inorganic substances exclusively in the manufac-
ture of their own food, in short, those bacteria which are
strictly prototrophic.

Let us consider the possible sources of the various ele-
ments needed in the building up of the primitive bacterial
protoplasm. We have no reason to suppose that such
primitive bacterial protoplasm differed in any marked
degree from the protoplasm of modern forms. Such or-
ganisms must have available carbon, hydrogen, nitrogen,
oxygen, sulphur, phosphorus and iron, with small quanti-
ties of a few other elements. Upon the earth before the
advent of other plant life, the carbon necessary for-
growth would probably be secured from carbon dioxide,
or possibly from methane or carbon monoxide; the hydro-
gen was undoubtedly present in abundance in water,
perhaps traces also of the free element, of methane, or ,
of ammonia may have been available; the nitrogen was -
probably present in sufficient quantities either as ammonia
or as nitrates, and of course in the form of less available,
relatively inert, gaseous nitrogen; the sulfur probably
existed as sulfids, sulfates, and free sulfur; the phos-
phorus was probably found in phosphates and the iron in
both ferrous and ferric condition. It is evident that ele-
ments and compounds were present in abundance and
variety, but not in the form of organic compounds. All
modern living organisms are divided into those which

238 THE AMERICAN NATURALIST [ Von. LIL

require their food to be ready manufactured for their use,
and those which can manufacture their own food (proto-
trophic). It is apparent that the primitive organism was
probably prototrophie.

The manufacture of food from inorganic materials re-
quires the expenditure of energy. We must account, if
possible, for the sources of such energy for the proto-
trophic forms. Among modern organisms the energy
needed is secured always from one of the two sources,
light rays or chemical oxidation. While other types of
energy are known, apparently plants have not been
adapted to their utilization. If light rays were first used
as an energy source, the primitive organism was prob-
ably provided with some pigment which was of signifi-
cance in the absorption of the light and in its conversion
into chemical energy. Among modern forms which may
have resembled such primitive organism may be cited the
simpler types of the blue-green alge and the phototactic
sulphur bacteria containing the pigment bacteriopur-
purin. If the Chamberlin planetesimal hypothesis of
earth origin is accepted, such might very possibly have
been the primitive types. However, primitive conditions
may have been such that light energy was not available.

-Organisms developing under such conditions must have
been directly dependent upon chemical energy. Such
energy might be secured by the oxidation of ferrous iron,
of free sulphur or of the sulfids (particularly hydrogen

_sulfid), methane, hydrogen, carbon monoxid and am-

_monia. Organisms among modern species are known
which can utilize each of these methods of securing
energy. There is no reason, therefore, why any one of
these should not be a method used by a primitive form.
The modern types of organisms which oxidize ferrous to
ferric iron are in many respects among the most highly
differentiated of the filamentous bacteria and show many
points of resemblance to the blue-green alge. They show
few primitive characters, and are probably to be regarded
as not closely related to the primitive bacteria.

No. 615] BACTERIAL PHYLOGENY 239

A study of the organisms which at the present time are
known to secure energy by the oxidation of H,S, CH,, H»,
CO and NH, show that they possess certain character-
isties in common. For the most part the organisms are
cocci or rods, occasionally spiral, in some cases motile,
and then always with polar flagella. While there are
some exceptions to the rule, the organisms for the most
part do not thrive in a medium containing much organic
matter. It is not improbable that the primitive organism
had characters not greatly unlike these enumerated.
Just what type of oxidation is most primitive it is diffi-
cult if not impossible to determine, although certain con-
jectures may not be out of place. Probably one of the
most common of the easily oxidized substances of the
primitive earth was hydrogen sulfid. It undoubtedly
was a common constituent of thermal springs. The
modern representatives of the groups which thrive in
water containing hydrogen sulfid are abundant both in
numbers and in species. By means of the energy which
they secure from the oxidation of H,S and S they probably
take up CO, and transform it into food and protoplasm.
Apparently all of the forms which have been investigated
are motionless cocci or rods or spirals motile by means
of polar flagella. No modern form is known which pro-
duces spores. Many of the species contain a pigment
bacteriopurpurin and swim or grow toward light. show-
ing positive chemotaxis or chemotropism. We may find
every gradation between the modern representatives of
these forms and the blue-green alge, on the one hand, and
the true bacteria, on the other. Many of the blue-green
alge contain a purple coloring material in addition to the

blue and green pigments. From the standpoint of evolu-
tionary requirements, therefore, it is evident that some
primitive organism having much the same type of metab-
olism as the modern sulphur bacteria would be a satis-
factory starting form. y

Before additional stress is laid upon a sulfur bac-
terium as a possible progenitor of modern forms, we

240 THE AMERICAN NATURALIST [Vou. LII

should examine carefully other possibilities. It is con-
ceivable (though scarcely probable) that hydrogen may
have constituted a larger percentage of the atmosphere
in times past than now. Several species of modern bac-
teria have been described which in the presence of hydro-
gen and oxygen may secure their growth energy by com-
bining these elements directly or indirectly. These
species are motile rods with polar flagella. These
modern members of the genus Hydrogenomonas, how-
ever, are very far from primitive because under ordinary
conditions they are pantotrophous growing well on ordi-
nary laboratory media. Thus far no organism strictly
prototrophic capable of utilizing hydrogen has been
found. This does not prove that such organism has not
existed, but throws the burden of proof upon any one
who would urge hydrogen oxidation as a primitive method
of securing growth energy. The results of Kaserer
(1906) seem to indicate that the organism catalytically
causes the transformation in the presence of hydrogen
of carbonic acid into formaldehyde, the cell then using the
formaldehyde as food.

Methane and carbon monoxide are also oxidized by cer-
tain of our modern bacteria, the organisms securing their
growth energy in this manner. These organisms accord-
ing to the descriptions are autotrophic and do not thrive
in the presence of organic material. It is possible that
these represent primitive characters. The organisms are
rods, motile or non-motile, when motile with polar
flagella. If either methane or carbon monoxide were
common in the atmosphere of the early earth, forms of
this general type may have flourished. ‘That these gases
were sufficintly abundant does not seem probable, but the
possibility must be admitted. |

Several types of modern bacteria are known which
oxidize ammonia to nitrites and nitrites to nitrates, utiliz-
ing the energy thus secured for chemosynthesis of food
from inorganic materials. At least one species of the
nitrifying bacteria is a coccus, others are rods, motile by

No. 615] BACTERIAL PHYLOGENY 241

means of polar flagella. It is not at all improbable that
ammonia may have been abundant enough on the primi-
tive earth to have constituted an adequate energy source
for the primitive bacteria.

Which of these modern types most closely resembles
the primitive organism living on autotrophic existence?
It is perhaps impossible to say. The modern representa-
tives of the nitrifiers and the methane and carbon
monoxid oxidizers are apparently rather isolated groups
without numerous species and apparently not closely
related to other forms. The sulfur oxidizers, on the
other hand, are abundant, of many types, and show many
intergradations with other bacteria and the blue-green
alge. Possibly a somewhat better case can be made out
for them. However, it should be noted that all of these
forms have certain characters in common, they are all
autotrophic, all are aerobic, and when motile are elongate
cells with polar flagella. It is perhaps a fair inference
that the aerobiosis and the polar flagellation are primitive
characters. We may well conclude with Jensen that al!
of these organisms discussed are related and may be

= LES...)
( Thiomonas ? )
T
S Girepsemenas) ( Nitrosorronas ) ( a NTN )
| Mitrosococcus ) (Hydrogenomonas ) genomonas E

Citrobacter )
EUBACTERIALES

Fic. 2. PROBABLE RELATIONSHIP OF CERTAIN MODERN GENERA OF BACTERIA TO
HE PRIMITIVE ORGANISM AND THEIR RELATIONSHIPS TO EACH OTHER,

placed in a single group. Expressed in terms of modern
representatives of the primitive types, the following dia-
gram might express the idea.

We may next concern ourselves with possible and
probable relationship of these various forms to other
members of the Eubacteriales, disregarding the Thio-
bacteriales. The remainder of the Eubacteriales differ
from the autotrophic forms thus far discussed in that in

242 THE AMERICAN NATURALIST [ Vou. LIT

every case they require the presence of organic carbon
compounds in the substrate in which they grow. These
compounds may be of the greatest diversity of types, but
none of the bacteria are capable of manufacturing their
own carbon food. It is possible that other types of bac-
teria than the prototrophic may not have developed upon
the earth until after the evolution of higher plants, such
as the alge, upon which they could depend for food.
Possibly there may have been some start made, however,
in the utilization by one type of organism of the dead
bacterial protoplasm of another type.

How may we detect relationships of modern meta-
trophic bacteria to these more primitive types? Possibly
by a study of intergrading forms. The genus Hydro-
genomonas apparently is either autotrophic or meta-
trophic according to the conditions of the environment.
Some primitive organism may have acquired properties
similar to those of the modern Hydrogenomonas and con-
stituted the progenitors of the modern forms. Possibly
this type of differentiation may have arisen in several
groups. It is conceivable, for example, that some organ-
ism having characters such as Nitrosococcus might have
given rise to an independent branch, possibly to forms
like Micrococcus. This, of course, is pure speculation.

Among the metatrophic bacteria we are probably justi-
fied in placing the genus Pseudomonas as most closely
related to the forms discussed because of its close morpho-
logic resemblance, with rod-shaped cell and polar flagella,
to the autotrophic forms; then too, there is the evidence
of the intergrading Hydrogenomonas. Somewhat less
diversified in nitrogen metabolism are the related genera
Azotabacter and Rhizobium, both usually with polar fla-
gella, rod-shaped bodies, primitive nitrogen requirements,
and marked capacity to utilize carbohydrates, oxidizing
them quite completely to CO, and H,O. The supply of
energy is so abundant to these organisms that in the absence
of sufficient combined nitrogen in the substrate they can fix
atmospheric nitrogen, and build it into their protoplasm.

No. 615] BACTERIAL PHYLOGENY 243

This nitrogen fixation must be carefully differentiated
from the nitrification previously discussed. Probably
the non-motile group Mycoderma which resembles the
other organisms in ability to oxidize sugars (preferably
ethyl aleohol), but is non-motile, should be placed here.
These three genera are obligate aerobes and secure their
growth energy by relatively complete oxidation of carbo-
hydrates, alcohol or even acetic acid. They apparently
constitute a natural group related to Pseudomonas. It
should be recalled that a statement of relationship does
not imply derivation, but simply common ancestry.

We have now considered all the bacteria which show
the primitive characters of polar flagellation and obligate
aerobic utilization of carbonaceous foods. In the genus
Pseudomonas we find evidences of differentiation in me-
tabolism, particularly ability to bring about proteolysis.
In some species we have evidences of adaptation to
anaerobic conditions, among the so-called denitrifiers.
Some members of this group are capable of taking oxygen
from nitrite and nitrates under anaerobic conditions, with
evolution of free nitrogen. Other forms are known that,
can reduce sulfates to sulfids. Such facultative an-
aerobes, securing oxygen from an easily reduced com-
pound, evidently make use of the oxygen in the same
manner as though growing under aerobic conditions for
the oxidation of carbon compounds. The next step in the
development of anaerobiosis was probably the utilization
of carbon compounds, securing growth energy by intra-
molecular oxidations; in such forms fermentative capacity
becomes well EN

The close relationship im morphology and physiology
existing between the short spiral Vibrio and Pseudo-
monas indicates that the family Spirillacee has come from
an ancestry having much in common with Pseudomonas.

The other bacteria belonging to the\Eubacteriales are
more specialized in general morphology. and in physiol-
ogy than the forms thus far mentioned. 7 ven motile the
cells are peritrichous rather than with pblar flagella.

244 THE AMERICAN NATURALIST [ Vou. LI

Some forms have developed the ability to produce endo-
spores (family Bacillacee) and seem to comprise a
closely related group of genera whose relationship to the
more primitive types is somewhat problematic. Another
well-marked group of bacteria includes the large series
of (usually) gram-negative bacteria that produce no
spores. These may be included in a family Bacteriacez.
With the exception of polar flagella, there is no very
marked difference between the Pseudomonas forms and
the Proteus types. It is quite possible that they are
closely related. The cocci apparently form another
homogeneous group, the Coccacee. The affinities of the
group may be sought in several places. For example,
there is apparently very close resemblance culturally and
physiologically between the chromogenic cocci and the
chromogenic rods closely related to the genus Bacterium;
the organism usually termed Bacillus prodigiosus (Ser-
ratia marcescens) is remarkably near certain red cocci as
Rhodococcus roseus.. The possibility that there is a rela- .
tionship between the Nitrosococcus and Micrococcus has
already been pointed out. Then there is a decided rela-
tionship evident between the aciduric bacilli and the genus
Streptococcus. All of these origins are possible; if all
these relationships are true, the group Coccacee must be
regarded as heterogeneous, that is, polyphyletic.

The group containing the tubercle bacillus (Mycobac-
terium) and diphtheria bacillus (Corynebacterium)
shows undoubted relationships to the order Actinomy-
cetales. If they have no common origin with other
genera of the Eubacteriales they should be included in the
order Actinomycetales. However, there is decided evi-
dence of relationship through Leptotrichia and perhaps
Erysipelothrix to the lactic acid bacteria. If this is a
valid relationship it would indicate that the Actinomy-
cetales are an offshoot of the Eubacteriales, or at least
have a common ancestry.

The various relationships illustrating the probable
phylogeny of the class Bacteria is illustrated in the ap-

BACTERIAL PHYLOGENY 245

No. 615]

‘VINGLOVG JO SUAGNQ ANV SAITINVA ‘VUANAH AHL 40 AIHSNOLLVIA J GULSADDAS ‘g "DIA
= CITWIYJILIVIOXA LI Y SAIVLIDAWONILOV
| ž
SVIOVINILOVa

avaovTioVva JVJ YI YIJLVJ ODAN

csJav.
-143.L29v9N3
IVIIDVWIAVNOLIOUNISA + re TN EELE IVWI2VW2202
IVIIVMUIYIAS atada O DI L7OGOLOZY $1920904 AIHE (5n33079 9 >
Taca, } naoz Y WAWVINALIVGOY.LIN 5737070] Kydd \W 3015009797 )
i WY fp htidg SOLIOLIOUIIDO sp. 7 ALDO GOHAN SADIODOSOM IAS $N 220207 /V DITIGSION
/ OTAGA SOLIOLLIOPANOG JOJ DUOI O UOA PL SDUOLHOSOLLIRI LETIA CE
7 SIAL PANTENE ' > - A EP TO APDL
STIM (Li PITASOPOYY O/SG/AOCPOYY ) E>uotomL) $073070141 ) YoSoy Cc3I1WVIY3LI9VW9Y
-3VWH2OYIAS |f- SITVIYVYILIVIOIHL -OQAWYIHI

246 THE AMERICAN NATURALIST [ Von. LIT

pended diagram. Relationships which have appeared
probable to the writer have been indicated by solid lines
connecting genera, possible relationships have been indi-
cated by dotted lines. The genera comprising the fami-
lies recognized by the Committee on Nomenclature as
belonging to a single family are enclosed by a heavy line.
Genera not recognized by the committee are enclosed in
dotted lines.

The foregoing analysis would seem to indicate that the
grouping of genera by the Committee on Classification,
with some slight modifications possibly, represents fairly
well true phylogenetic relationships of the bacteria. The
exact boundaries of the families are of course of little im
portance providing the scheme of classification tends to
show relationships. |

DISPROOF OF A CERTAIN TYPE OF THEORIES
OF CROSSING OVER BETWEEN
CHROMOSOMES!

PROFESSOR H. S. JENNINGS

JOHNS HOPKINS UNIVERSITY

I

Two types of relations have been proposed to account
for the facts of “crossing over?” between pairs of char-
acters that follow the same pair of chromosomes. One is
a varying relation between the substances forming the
factors belonging to diverse pairs in the same chromo-
some; the other a varying relation between the two mem-
bers of the same pair, in the two paired chromosomes.

‘The former type is represented by the ‘‘chiasmatype”’
theory, held by Morgan and his associates, in which the
diverse relations are held to be, or depend upon, the ac-
tual diverse distances apart of the factors extended along
the linear chromosome. When the chromosome breaks,
for any cause, it is more likely to separate two factors far
apart than two close together; on this depends the vary-
ing cross-over ratios.

The second type is that brought to notice recently by
Goldschmidt (1917), and commonly called the ‘‘variable
force” theory. It is conceived that the two members of
a given pair, as A and a, in the two paired chromosomes,
may be held or drawn to their places by a pair of varying
forces, which allow them to exchange places on the aver-
age in a certain proportion of cases; while B and b are
held by a different pair of forces, which allows these two
to interchange in a different proportion of cases; C and e
by a still different pair, ete. The result would be diverse

1 This paper arose and took shape during discussions on theories of
erossing over in the Seminary on Genetics at the Johns Hopkins University.

247

248 THE AMERICAN NATURALIST [ Vou. LII

ratios of crossing over when diverse pairs are compared;
the cross-over ratio between A-a and B-b would be dif-
ferent from that between A-a and C-c, and so on. It is
this theory that I propose to examine. I do not under-
stand that Goldschmidt commits himself to any form of
this theory, or to any theory that is exclusively of this
type, so that this discussion is not presented as a com-
mentary on his views, but on this type of theory for its
own sake. Is it possible to explain the observed ratios of
crossing over by any theory of this type?

To grasp the matter clearly, it will help to have an
example before us. Let the following twelve groups of
letters represent twelve pairs of chromosomes in twelve
cells, each chromosome bearing two factors, which we
will call A-B and a-b. The upper two letters in each
pair show a single chromosome containing the factors
A-B, the lower two the mated chromosome containing the
factors a-b.

I AB AB AB AB AB AB AB AB AB AB AB AB
" ab ab ab ab ab ab ab ab ab ab ab ab

Now suppose that the forces holding A and a to their
chromosomes are such that A and a exchange in one
fourth of all cases, while B and b exchange in one third of
‘all cases. That is, A and a exchange places in every
fourth chromosome pair, B and b exchange places in every
third pair. The letters that are thus to exchange with
their mates are italicized in the pairs indicated above.
The exchanges will evidently give the following result:
TT AB AB Ab aB AB Ab AB aB Ab AB AB ab

; ab ab aB Ab ab aB ab Ab aB ab ab AB

my es Tr

By a cross-over is meant the fact that two factors of
diverse pairs, as A and B, which in the germ cells that
formed the parent were following the same chromosome
(as in I, above), are found in the germ cells from those
parents (II, above) to be following diverse chromosomes
(as in the third pair of II, above); while conversely the
two factors A and b, which were following diverse

No. 615] THEORIES OF CROSSING OVER 249

chromosomes, are now following the same one. The
cross-overs in II, above, are those indicated by the +
sign; there are five of these. The ratio of the number of
these new combinations (5) to the total number of germ
cells (12) is the cross-over ratio; in this case the cross-
over ratio'is %2, or .417.

Examination of this case will illustrate an important
fact. A cross-over is produced only when one of the two
pairs exchanges while the other does not. In the last pair
to the right, in the example given above, the members of
both pairs exchange places, but this does not give a cross-
over—since A and B are still together, as they were
before the double exchange.

Now if the number of exchanges for each pair of cells
is different from that in the example given above, the
resulting cross-over ratio will be different. By suppos-
ing each pair of factors, A-a, B-b, C-c, D-d, ete., to have
its own characteristically diverse frequence of inter-
change of its members, all sorts of cross-over ratios could
be obtained, varying from 0 to 1; that is, from no cross-
overs to all cross-overs. The question in which we are
interested is, could the observed cross-over ratios in such
an organism as Drosophila be accounted for in this way?

Tt is to be noted that the problem as we take it up is
independent of the nature of the forces that hold A and a
(and the other factors) in their places, and that permit
them to exchange in a certain proportion of cases. These
forces may be utterly heterogeneous in the different
cases; they may turn out to be of any kind whatever, so
far as this examination goes. We ask merely whether,
if the forces, whatever they are, give a constant average
proportion of interchanges characteristic for each pair,
they can yield the cross-over ratios actually observed.

:

Tt is evident that on this theory there are two kinds of
ratios to be dealt with: the ratio of the number of inter-
changes of A and a (characteristic for each pair), and the
ratio of the number of cross-overs, between two pairs A-a

250 THE AMERICAN NATURALIST [ Vou. LII

and B-b; the latter of these ratios depends on the former.
We shall call the former the exchange ratio; the latter is
commonly known as the cross-over ratio, which we will
designate by the letter C.

The exchange ratio signifies the ratio of the number of
exchanges between A and a to the total number of germ
cells:

Exchanges
— Total Number

The eross-over ratio (C) signifies, of course (following
Morgan and the general usage), the ratio of the number
of cross-overs to the total number of germ cells or
progeny:

Exchange Ratio

Cross-overs
ae Total Number

Goldschmidt (1917, page 90) has given a formula for
the cross-over ratio resulting from any two exchange
ratios, and has computed the resulting cross-over ratios
from certain assumed exchange ratios. We shall give the
formula a simpler expression than Goldschmidt has done;
one that will enable us to determine its properties and
limits of performance.

In cross-over ratios we deal with two pairs of char-
acters, which we may designate A-a and B-b. Let zx
signify the exchange ratio for one of the pairs; and let y
signify the exchange ratio for the other pair. Thus, if
A and a interchange in one third of all cases, this pair’s
exchange ratio x will be one third (or .33%); while if B
and b interchange in two fifths of all cases, its ratio, y,
will be two fifths (or .4). For convenience we will always
choose x and y in such a way that if there is any differ-
ence, x will designate the smaller ratio. That is, x will
always be equal to or less than y.

Now, suppose that originally the first chromosome of
the pairs bears the two factors A and B, the second a and
b (as in I, above). After crossing over in the proportion

No. 615] THEORIES OF CROSSING OVER 251

x, we shall have, in these first chromosomes of the pair,
A and a in the following proportions:

xa
(1—x)A
Similarly, in this same chromosome we shall find B
and b distributed in the following proportions:

yb
(1—y)B

(Thus, if A and a interchange in two fifths of all cases,
then after interchange we shall, in the first chromosome,
find a in two fifths of the cases, A in three fifths; and
similarly for B.)

What will then be the proportions of the various com-
binations of the two pairs of factors? It will evidently be

xa + (1—x)A, multiplied by

yor ts ye
= xyab + x(1— y)aB + y(1— x) Ab

+ Us x) (17) 4B

The cross-overs are aB and Ab, the proportion of which
is evidently:

x(l—y) +y(1—x) =x+y-— 2xy

The same result will be reached if we consider the
second chromosome of each pair (that which originally
contained a and b); so that the same proportion holds for
both together. This, therefore, gives us our formula for
the cross-over ratio in terms of the exchange ratios of
the two pairs. It is essentially the same formula em-
ployed by Goldschmidt (1917), giving the same results,
but written in more perspicuous form. |

Let us therefore recapitulate in algebraic form the es-
sential points. :

252 THE AMERICAN NATURALIST [ Vor, LIT

x = exchange ratio of one pair,
y=exchange ratio of other pair
(so selected that x =y).

Then for the cross-over ratio (C) of the two pairs, the
formula is
r C=x+y— 2y.

An example or two will make the use of this formula
clear. Suppose that the exchange ratio of pair A-a is
%; of B-b it is %. Then

X= %; y=%
l C =% +% — 2%.) = %s = 486
Again, let
Kids ye aoe
C= .31-+ .34 — 2(.31 x .34) =.439

(It is customary to express the results as percentages;
thus the last example would give a cross-over ratio of
43.9 per cent. For our purposes it is more convenient to -
leave them as decimals.)

Now this formula has certain characteristics and limita-
tions that allow us to bring the theory on which it is based
to a test. The theory is that each pair has its character-
istic exchange ratio; if that be the case, this formula
holds.

We shall set forth certain of the important relations
between cross-over ratio and exchange ratios, revealed
by this formula; then show how these provide a test for
the theory which the formula expresses. To aid in the
comprehension of these relations, we give a table showing
all cross-over ratios for two pairs of characters, resulting
from the combinations of exchange ratios varying by
tenths from 0 (no exchange) to 1 (all exchange). The
table illustrates all the relations to be deduced from the
formula.

No. 615] THEORIES OF CROSSING OVER 253

Exchange Ratio for One Pair (A-a).
€i a USADO e AN LO

0 0 .10 .20 .30 .40|.50|.60 .70 .80 .90 1.00
.11 .10 .18 .26 .34 .42|.50|.58 .66 .74 82 .90
21 .20 .26 .32 .38 .44|.50|.56 .62 .68 .74 80
30 .34 .38 .42 .46| .50| .54 .58 .62 .66 .70
40 .42 .44 46 .48| .50| 52 .54 56 58 -60

.50 .50 .50 .50 .50/ .50| .50 .50 .50 .50 .50

.60 .58 .56 .54 .52|.50|.48 .46 44 42 40

80 .74 .68 .62 .56|.50 |.44 .38 .32 .26 .20

.90 .82 .74 .66 .58 | .50 | .42 .34 .26 .18 .10

Exchange Ratio for the Other Pair (B-b).

3
4
5
6
71 .70 .66 .62 .58 .54|.50 | .46 .42 .38 .34 .30
8
9
0

1.00 .90 .80 .70 .60|.50|.40 .30 .20 .10 0
A qQ—_— _

EXPLANATION OF THE TABLE
Table of the values of the eross-over ratios resulting from. combina-
tions of different = ratios, from 0 to 1, of two pairs of factors.
Based on the formu
(0=x4+y — oxy
Where x=the exchange ratio of one pair,

y=the exchange ratio of the other,

snd. xa yy.
Outside the square (above and to the left) are the exchange ratios, by
tenths, from 0 to 1; within are the cross-over ratios. To find the cross-

If the appir p quadrant:
If x or y or both are below thè values given in the sui the cross-
over pe is below the value giyen i in the tahe. Thus if x low .2

10; hss eross-over anti is below —
If x or y or both are above t values Es in the table, the cross-
over value is above that given in o table
In the lower right-hand qua adrant:
it æ or y or both are above the values in the table, the cross-over value
is below that given in the table.

254 .THE AMERICAN NATURALIST [ Von. LIT

r y or both are below those in the table, the cross-over value is
above that of the table

In the other two quadrants (upper right and lower left):
If x is smaller and on larger than in the table, the eross-over ratio is
above that o tab
If x is larger and y cs than in the table, the cross-over value is
below that of the table.
Example: x and y given; limits of C required:
x==.16,. y==.29; C<.38 and >.26
x= 09; y=:08; 0<:18-and >.0
AA TO CO <A and 0.0%
eS mir y hs: 0 <.00 and >.48
C given; limits of x and y required:
C=.434, x and y both =.434, or both =.566
C=.021, x and y both =.021,'or both = .979
OA y She
C06; wf 04:33 96
C==.50; x=.w50 : y= any. value from 0 to 1.

The formula for examination is:
C=x>+y-—2xy
(in which x and y are proper fractions).

1. Two exchange ratios, x and y, give the same cross-
over ratio (C) as do their complements, 1 — x and 1 — y.

or

x+y — 2xy = (1—x) + (1— y) -2(1—x) (1 — y),
as will be seen by performing the operations indicated in
the second member of the equation. But this second
member is the value of C for exchange ratios 1—x and
1—y.

For example, the two exchange ratios .2 and .3 give the
same cross-over ratio as do the two exchange ratios .8
and .7; for both cases C=.38. This relation is seen in
the symmetrical constitution of the table; the cross-over
ratio resulting from .1 and .2 is the same as that from
9 and .8; the cross-over ratio resulting from exchange
ratios .4 and .7 is the same as that resulting from .6 and
.3, ete. The rule holds equally for values not found in

the table; thus the cross-over ratio resulting from .011

No. 615] THEORIES OF CROSSING OVER 255

and .031 is the same as that resulting from .989 and .969.

2. If one of the two exchange ratios is changed to its
complement, the cross-over ratio is changed to its com-
plement.

That is, if the cross-over ratio resulting from x and y
is C, the cross-over ratio resulting from x and 1—y, or
yodi xia 1 C;

For:

x + (1— y) — 2x(1 —y) =1— (x+y — 2xy)

But the first member of this equation is the cross-bver
ratio from x and 1—y, while the second member is 1
minus the cross-over ratio from x and y. The same
result is reached if we take y and 1 — x.

Thus, as the table shows, the cross-over ratio resulting
from .2 and .3 is .38, so that the cross-over ratio from .2
and .7 is .62, as is likewise the cross-over ratio from .8 and
3 (.38+.62=1). Similarly, the cross-over ratio of .011
and .031 is .0413; hence the cross-over ratio from .011 and
.969 is .9587.

3. When the cross-over ratio is less than %, the ex-
change ratios x and y are either both greater than % or
both less than 14; one can not be less than Y, the other
. greater. That is:

If C< 1 then either x < % and y<% or x>¥% and
y>. For let us suppose that x = 12 — a and y = 12 +
b, in which a and b are any positive quantities. Then
‚C =x + y — 2xy = 2 + 2ab. Therefore x can not be less
than % and y more than 4.

On the other hand, if x = 12 — a and y= % — b, or if
x= +a, y= 42 + b; in either case C=x + y — 2xy =
Y) —2ab. So that in these cases the cross over-ratio C is
less than %.

4. Conversely to 3, when the exchange ratios x and y
are both less than 1, or when they are both more than Y,
the cross-over ratio is less than Y.

That is, when x<% and y< Y, or when x>% and

y > 16; in either case C < Y. This was proved under 3.
5. When the cross-over ratio is greater than Yo, one ex-

256 THE AMERICAN NATURALIST [ Von. LII

change ratio is less than 1⁄2, the other greater than 4.
That is: If C>%, then x< Y, y >%. This also was
proved under 3.

6. Conversely to 5, when one exchange ratio is less than
lo, the other greater than Y, the cross-over ratio is
greater than Y. That is: If x< Yo, y > Yo, then C > Y.
This also was proved under 3.

All these relations are evident in the table.

7. When the cross-over ratio is less than Y, the two
exchange ratios are either both equal to or less than the
eross-over ratio; or both equal to or more than the com-
plement of the cross-over ratio. They can not have any
value lying between the cross-over ratio and its comple-
ment. That is: When C < Y, either x and y each = C or
x and y each = 1 — ;

This is an EE TEAS important principle, on (hich the
final test of the theory depends. It is proved as follows:

In 3 we saw that if the cross-over ratio is less than 12,
either x and y are both less than 14; or both of them are
greater than 1⁄2.

(a) Let us take first the case where x and y are each
less than 4%. In this case, in the formula C =x +y —
2xy, the quantity 2xy is smaller than x, and smaller than
y. For since x is less than Y, 2x is less than 1, whence
it follows that 2xy is less than y; and the same reasoning
shows that 2xy is likewise smaller than x. Hence the
formula for C subtracts from the sum of x and y a quan-
tity smaller than y; it therefore leaves a quantity larger
than x; and the same reasoning shows that it leaves a
quantity larger than y. Only in the limiting case that
x=0 does y=

(b) Take next the other possible case, in which x and
y are both greater than 4%. In this case 1—x and 1— y
are both less than 1⁄2. Thence it follows (by the reason-
ing just employed) that

(1—x) + (1— y) — 2(1—x)(1— y)
is greater than 1—x and greater than 1—y. But, as
was seen in (1),

No. 615] THEORIES OF CROSSING OVER 257

(1—x)+(1— y) — 2(1—x)(1—y) =x +y- 2xy=C
So that in this case C>1—x and C >1—y. .
Thus the fraction C is nearer to 1 than the fraction
1—x. If therefore we subtract the fraction C from 1,
it will leave a smaller number than if we subtract the
smaller fraction 1 — x from 1. Thatis:1—C <x.

And in the same way it can be shown that 1— C <y.
Only in the limiting case that y=1 does x = 1 — C.

The general principle in this section can be expressed
as follows:

When the cross-over ratio is less than o, the two ex-
change ratios, x and y, either both differ from 0 by less
than the cross-over ratio, or both differ from 1 by less
than the cross-over ratio.

This relation is well seen in the table. For example,
for the cross-over ratio .38 the two exchange ratios are
either .3 and .2 (both less than .38), or they are .8 and .7
(both greater than .62, the complement of .38).

8. Conversely to 7:

If both exchange ratios, x and y, are less than %, both
are equal to or less than the cross-over ratio.

If both exchange ratios, x and y, are greater than Y,
both are equal to or greater than the complement (1 — C)
of the cross-over ratio.

9. When the cross-over ratio C is above 1%, one of the
exchange ratios (x) is equal to or less than the comple-
ment of the cross-over ratio (1— C), while the other (y)
is equal to or more than the cross-over ratio (C).

Or otherwise expressed:

When the cross-over ratio is above 12, one of the ex-
change ratios (x) differs from 1 by an amount equal to
or more than the cross-over ratio, while the other (y)
differs from 0 by an amount equal to or more than the
cross-over ratio. Thatis, when C>%,1—x=C,y Z0,
orx=1-—C,y=C.

This can be proved by methods similar to those em-
ployed in 7.

10. Conversely to 9:

260 - THE AMERICAN NATURALIST [ Vou. LII

high or very low), they can not together give cross-over
ratios of the more intermediate values. For example (as
our table shows), if two factor pairs each give, with any
other, cross-over ratios below .10, they can not give
together a cross-over ratio lying anywhere between .18
and .84. If the two pairs each yield any cross-over ratios
lying below .20, they can not give together a cross-over
ratio lying between .32 and .68. These and many similar
relations, illustrated in the table, are inherent in the
theory we are considering, but are completely opposed to
what is found in nature.
IV

These facts completely refute any theory which holds
that the observed constant cross-over ratios between pairs
of factors are the result of constant exchange ratios be-
tween the two members of a given pair—exchange ratios
that are characteristically diverse for the different pairs
(such theories as that outlined by Goldschmidt, 1917).
The refutation is independent of the question of the
nature of the forces involved; whatever the forces, if they
give constant average exchange ratios for each pair, the
results are bound to be inconsistent with the observed
cross-over ratios. No theory will hold that does not
provide for diverse relations between the different factors
in the same chromosome, such that some tend to cling
together more frequently than others.

Possibly some elements of the theory that diverse ex-
change ratios are characteristic for different pairs might
be retained, if there be added provision for modification
of the exchange ratio in a given pair, depending on
whether or not exchange occurs in some other pair. It
might be held, for example, that A and a are more likely
to exchange if in the same cell B and b have exchanged;
or the reverse. This would give a theory of mixed type,
which added to the forces regulating the exchange be-
tween two members of a pair, other forces causing two
given pairs to tend to do the same thing, or the opposite

No. 615] THEORIES OF CROSSING OVER 261

thing. The “variable force” theory would therein ap-
proach the chiasmatype theory, in which the diverse rela-
tions between the factors belonging to different pairs are
the primary, if not the exclusive, elements considered.

As theories of other type become successively modified
so as to take into account the known facts: the fact that
the chromosome actually is a linear aggregate; the fact
that the two chromosomes while in this linear condition
pair and intertwine; the fact that cross-overs occur only
at the period when this occurs; the fact that two reces-
sive allelomorphs when mated do not produce normals,
while two recessives not allelomorphs do; the fact that
after two factors, A and B, are found to hold together in
one generation, if we mate their cross-overs A-b and
a-B, we now find that it is A and b, not A and B that
tend to hold together (Bridges, 1917); the fact that when
a given factor is lost from a chromosome, others that
have low cross-over ratios with that factor are also lost
(Bridges, 1917a);—when the modifications required for
bringing these facts into relation with each other and
with others are introduced, it appears that the resulting
theory will come more and more to resemble the chiasma-
type theory. No theory is adequate that does not include
and bring into relation the facts just mentioned. for a
correct theory is nothing but a presentation of the facts
in their correct (verifiable) relations.

PAPERS CITED

Bridges, C. B.

1917. An Intrinsic Difficulty for the Variable Force Hypothesis of

Crossing Over. AMERICAN NATURALIST, 51: 370-37

Bridges, C. B.

1917a. Deficiency. Genetics, 2: 445-465.
Goldschmidt, R.

1917. Crossing Over ohne Chiasmatypie? Genetics, 2: 82-95.

Sde T. H., and Bridges, C. B.

1916. j linked Inheritance in Drosophila. Publication No. 237,

Carnegie Institution of Washington. 87 pp.

SHORTER ARTICLES AND DISCUSSION

NOTE ON THE COLORATION OF PLANES MINUTUS!

Ir is well known that the coloration of the grapsoid crab Planes
minutus, a constant member of the Sargassum fauna, is ‘‘homo-
ehromie”” to a high degree, not only as to tint and mottling, but
also in the frequent oceurrence of a bloteh of pale yellowish or
blank white upon the carapace; this has generally been supposed
to be a mimicking of the white patches of encrusting bryozoa and
Spirorbis tubes, which commonly infest the Sargassum.?. Ex-
periments made to discover the extent of possible color changes
in the adult Planes when it is placed over variously pigmented
artificial bottoms have led to no result, other than to show—con-
formably with what is known for some other crustacea possessing
a dense body pigmentation, as contrasted with a relatively scanty
supply of well-scattered chromatophores—that the power of
color adaptation is decidedly limited. It is, therefore, of interest
to make record of an instance in which pronounced color adapta-
tion of Planes had occurred in nature.

In January, 1916, after a rather severe gale, there was found
stranded upon one of the reef ‘‘heads’’ at Bermuda a large
‘‘ Spanish cedar” tree. It is certain that the tree had been in
the sea for some time, as the surface layer was thickly populated
by Teredo and boring amphipods. The trunk, the stumps of the
roots and the submerged branches of the tree were covered with
a forest of barnacles, Lepas anatifera, from among whose smoky-
brown erectile peduncles were obtained a vast number of adult
Planes minutus that were adhering to the more or less honey-
combed parts of the exposed bark and wood. Without excep-
tion the crabs were deep brownish-red, save for the frequently
occurring dorsal white patch. This pigmentation harmonized
precisely in general tint with the mahogany-colored surface of
the cedar tree.

The interest of thin: case lies in its demonstration that these
erabs—prominent members of that specialized gulf-weed fauna
which has been urged as part of an argument for the antiquity

1 Contributions from the Bermuda Biological Station for Research, No.
84. po

2.Cf. Verrill, 1908, Pl. XII; ns and Hjort, 1912, p. 671, Pl. VI.

No. 615] SHORTER ARTICLES AND DISCUSSION 263

of the floating beds of Sargassum (Collins, 1917), probably
for generations experiencing no other habitat than the gulf-
weed—having yet retained a considerable capacity of color
adaptation. Among Sargassum the hues of Planes vary consid-
erably,* but the color of the present specimens was very much
darker and redder than that of any I have seen described. The
color agreement could hardly have resulted from a general stain-
ing of the crabs following ingestion of pigment derived from the
tree, as the characteristic white blotch upon the dorsum was
fully as well, if not somewhat more, developed in many of these
specimens, than in the common ones living upon gulf-weed.
Spectroscopic examination of alcoholic extracts of these crabs
showed that the pigment was not detectably different from that
of Planes taken on Sargassum. Whether the white patch repre-
sents in this instance an inherited tendency to lack of pigment
on that area, or is rather to be regarded as (in addition) a
mimicking of the white valves of the accompanying Lepads, is a
question; the conspicuous development of the white shield, its
large size and precise outline in more than 50 per cent. of the
individuals, suggests the possibility of the latter alternative.
Presumably the floating cedar tree was invaded by Planes
larvæ, which developed upon this dark reddish-brown substratum,
and, like Hippolyte in the experiments of Gamble and Keeble
(1900), produced there a pigmentation of corresponding appear-
ance. In this way a coloration might be acquired which the
crabs probably could not, at least quickly, have accomplished by
adaptive color change in the adult state. No color changes were
detected when these dark crabs were kept for six days upon
Sargassum, in bright light. i
REFERENCES
Collins, F. 8.
1917. The Sargasso Sea. Rhodora, Vol. 19, pp. 77-84.
Gamble, F. W., and Keeble, F. W.
1900. ERWA varians: A Study in Color-Change. Quart. Jour.
Micros. Sci, Vol. 43, pp. 589-698.
Murray, J., and Hjort, J.
1912. io Depths of the Ocean. London, 8vo., xx + 821 pp.
Verrill, A.
1908. oi Crustacea of pe aa I: Brachyura and Anomura.
Trans. Conn. Acad. Arts and Sci., Vol. 13, pp. 299-474.
AGar’s ISLAND,
BERMUDA MW... CROZIER.
3 Cf. Murray and Hjort, 1912, Pl. VI. a

264 THE AMERICAN NATURALIST -° [VoL. LII

THREE MUTATIONS IN PREVIOUSLY KNOWN LOCI

THREE mutations is the sex-chromosome have occurred in my
cultures of Drosophila melanogaster (ampelophila). Two were
reappearances of genes already known, namely, white and rudi-
mentary ; the third was the appearance of a new gene at the white
locus and has been named coral, symbol w“. In each case it is
clear that the changes occurred in the wild type gene of a mater-
nal chromosome. The evidence also indicates that the new gene
arose relatively late in the history of the egg in each case, whereas
if the mutation had occurred in the early oogonial stages several
individuals with the new gene should have appeared. In eases
where such information is known, it seems worth recording since
it will make possible a later consideration of the relative stabil-
ity of genes by a summing up of the frequen, of mutations in
the different loci.

As has been pointed out by Muller! recessive genes might ex-
ist for a long time before making an appearance, in case they
were closely linked to a lethal. The character produced by the
gene would ultimately be allowed to appear as the result of a
cross over which would separate the gene and the lethal from
the same chromosome. Previous to the time of crossing-over the
character produced by the gene would never be seen, since all
individuals pure for it would also be pure for the lethal and not
survive. The gene could be indefinitely transmitted: along with
the lethal through heterozygous individuals. I mention this
point because it is necessary in establishing the time of origin
of a mutation to consider whether its appearance may be due
merely to its recent | separation from a lethal, which had obscured
it. The three mutations dealt with here could not have been
masked by a lethal because they were in the X-chromosome, and
the presence of a lethal would have been apparent, as it would
have produced a lethal sex ratio. No such lethal ratio has been
found in connection with any of the three mutants either before
or since their appearance. In these cases, then, it is safe to as-
sume that the appearance of the first mutant marks the time of
the mutation. If the mutation had occurred in earlier genera-
tions, several individuals bearing the character would have ap-
peared instead of one.

In the case of the reappearance of a character, careful consid- -
eration must be given to the possibility of contamination, as has

_1Proc. Nat. Acad. Sci., Vol. 3.

No. 615] SHORTER ARTICLES AND DISCUSSION 265

been pointed out by Morgan and Plough This possibility has
been taken into account and is discussed with reference to the
appearance of each gene in that particular section.

ORIGIN AND DESCRIPTION OF CORAL

Coral arose in a mating of an eosin miniature bar-eyed male
to a forked female with normal eyes. This female was a ‘‘sec-
ondary exception”” from an XXY mother which had had no sex-
linked eye color in her pedigree. Among 279 offspring that
_were of the expected classes and showed no lethal sex ratio, there
was found about the middle of the count one heterozygous bar
female which seemed to be ‘‘exceptionally’’ dark eosin. An eosin
eye color in a female would be impossible to account for, since
to be a female she must have obtained one of the mother’s chromo-
somes, both of which carried normal red factors, as well as
receiving the eosin bar chromosome of the father. A mating
to one of her red-eyed brothers showed at once that the supposed
eosin female was actually heterozygous for eosin and for a new
allelomorph (coral) as she gave two kinds of sons, eosin and
coral, while the daughters were eosin and the compound eosin-
coral. The eosin-coral females are darker than pure eosin fe-
males and the original female was of this nature.

Coral is the seventh mutant allelomorph to be found in the
white locus and counting the wild type gene forms with them a
system of eight allelomorphs. In the order of their discovery
these are: red, white, eosin, cherry, blood, tinged, buff and coral.
Coral does not show bi-colorism, but is the same for males and
females. It is similar to the color of very dark coral. It is
darker than all the other members of this series with the possible
exception of blood which according to the description of Hyde
in his discussion of blood? shows a considerable variation of
color according to cultural conditions. The color of coral is very
close to the darker shades of blood, but is much darker than the
lighter shades and does not show any such variations in range of
color. Coral is distinctly darker than cherry and the other
lighter members of this series. Coral is a dull color and does
have the brightness of color of the wild stock, neither does it
show the fleck in the eye.

The original coral. female was re-mated to a white male from
stock and behaved genetically, as would be expected on the as-

2 AMER. NAT., Vol

3 Genetics, 1, Kariah, 1916.

266 ‘THE AMERICAN NATURALIST [ Vou, LII

sumption that coral was a member of the white allelomorphie
series. The heterozygous white-coral compound in the female is
intermediate in color between the two pure stocks. Coral is re-
cessive to red. A coral male crossed to a yellow-white female
gave all yellow-white sons and the intermediate (compound)
white-coral daughters. Evidently the mutation took place in the
wild-type gene of the mother, since it is that gene which did not
oceur in the daughter while the eosin gene of the father is re-
tained. It also occurred near the maturation divisions as only
one individual of the kind appeared. If the change had oe-
curred in the early stages of the egg, it would probably have re-
sulted in several of the offspring showing the new character.

REAPPEARANCE OF WHITE

In a cross of a bar male to a red-eyed female, which produced
251 offspring without a lethal sex ratio, one male was obtained
which was white, although there was no white in the pedigree of
either parent. This fly was found in one of the last counts of the
bottle and had the appearance of being a young fly. Counts
were made from the bottle every two days. Since I had no cul-
tures going at that time which contained white and had had no
white flies in my etherizing bottle previously, the fly can not be
accounted for by assuming that it had remained in the etherizing
bottle from a previous count of another bottle. It is highly
probable, though not absolutely certain from these considera-
tions that this white male was not due to contamination, but
rather to a mutation in the wild type gene of a maternal chromo-
some. We may be sure that this change took place in the ma-
ternal chromosome rather than in that of the father, since males
always receive their one X-chromosome from the mother except
in relatively rare cases of non-disjunction, and in this case the
male would have been bar.

In appearance the new white is not distinguishable from ‘the
white of the original stock and is quite without color in both
males and females. Dr. A. H. Sturtevant has been testing the
effect of various concentrations of aleohol in extracting color
from the eyes of flies which are members of this multiple allelo-
morphic series and kindly added this new white to the material
which he tested. He reported that the new white is acted upon
exactly as is the original white. Genetic results showed the new
white to behave as an allelomorph of the old. The new white
male was crossed to a red sister and the offspring were all red.

a

No.615] SHORTER ARTICLES AND DISCUSSION 267

The F, generation gave females all red and the males in. equal
classes of red and white, which is the genetic behavior expected
for a sex-linked gene. To test whether this white was in the
same locus as the old white, a white male of this stock was
crossed to a yellow-white female from the original stock. The
sons were yellow white and the daughters were white, not yellow,
since yellow is recessive and was not carried by the father. No
difference could be observed in eye color between either sex of
the new white, or the daughters compounded from the two
whites, and the males and females of the original white. It
seems reasonable to conclude then that the white gene has reap-
peared by a second mutation from the red gene.

SECOND ORIGIN OF GENE FOR RUDIMENTARY WING

There appeared in a eross of an eosin miniature male to a
broad, vermilion, forked female (both from stock cultures, all
characters mentioned being sex linked) one son which was ver-
milion, forked like the mother, but which also had shortened
wings. This wing character was later shown to be rudimentary.
Crossovers in later generations showed that the maternal gene
for broad was also present, but its effect was obscured by the
rudimentary in all cases where both occurred together. This
male so obtained and bearing genes for broad, vermilion, rudi-
mentary and forked was outerossed to a virgin wild type female
to test whether the new character was of a genetic nature. The
F, flies were normal in all respects. One pair of these produced
117 sons which were classified with respect to the characters ver-
milion, rudimentary and forked, while no attention was paid to
broad, which did show in certain crossovers where it was sepa-
rated from rudimentary. Out of 117 males, 3 were crossovers
between rudimentary and forked, which gave a percentage of
crossing-over of 2.6, whereas the value given by Morgan and
Bridges* is 1.4 on the basis of a much larger number of flies.
There were 27 crossovers between rudimentary and vermilion,
which is a percentage of 22.2, while the above authors put it at
24.1. The nature of the crossovers obtained showed that the
gene for the wing character was between vermilion and forked,
which agrees with the assumption that it is a new appearance of
rudimentary. The crossover values obtained are sufficiently
near to expectation to justify this assumption in view of the
small number of flies. Crosses were made to the stock rudimen-

4 Carnegie Pub. No. 237, 1916.

268 THE AMERICAN NATURALIST [VoL. LII

tary to make sure that the new gene was at the rudimentary
locus. Since homozygous rudimentary females show a high de-
eree of sterility, the rudimentary stock is kept by crossing it to
forked and using normal-winged females that are heterozygous
for both rudimentary and for forked. One of the new rudimen-
tary males was crossed to such a heterozygous female and the
new rudimentary was shown to be an allelomorph of the old, as
both rudimentary sons and daughters were obtained in practi-
eally equal numbers. The new rudimentary stock resembled the
old as regards the sterility of the homozygous females. Miss C.
J. Lynch in this laboratory tested several and reported that they
showed the same high degree of sterility. Since the new char-
acter has the same appearance as old rudimentary, this seems to
be merely the reappearance of that gene.

In this case it is clear that the change occurred in one of the
maternal sex-chromosomes which already carried three sex-
linked genes. The linkage relations of the new gene to these
maternal genes make its origin in the maternal chromosome cer-
tain. Moreover, the male could have received his sex-chromo-
some only from his mother, as otherwise he would have been an
XO male and would haye been sterile.® The fly could not be ac-
counted for on the assumption of contamination, as there are no
flies of that particular constitution in the laboratory. The muta-
tion was from the normal gene at the rudimentary locus. The
appearance of only one individual indicates that the change oc-
eurred late in the history of the egg.

SUMMARY
1. Two mutations have occurred at the white locus in the nor-
mal red gene, giving rise to a reappearance of white and to a new
gene which produces an eye color called coral.
2. Coral is the eighth member of the multiple allelomorph
series at the white locus.
3. Rudimentary reappeared as a change from the normal gene
at Miot locus in a maternal chromosome.
; LITERATURE CITED
Bridges, C. B.
1916. Non-disjunction as Proof of the Chromosome Theory of
Heredity. Genetics, I, 1-52, 107-163.
Hyde, Roscoe R
1916. e New Members of a Sex-linked Eco Allelomorph Sys-
tem. Genetics, I, 535-580.

_ 5 Bridges, Genetics, II, 1916.

No.615] SHORTER ARTICLES AND DISCUSSION 269

Morgan, T. H., and Bridges, C. B.
1916. Sex- linked Inheritance in Drosophila. Carnegie Publication No.
237.

Morgan and Plough.
1915. Appearance of Known Mutations in other Mutant Stocks.
ER. NAT., Vol. XLIX,
Muller, H. J.
1917. An Qnothera-like case in Drosophila. Proc. Nat. Acad. Bei.,
Vol. 3 D. E. LANCEFIELD
COLUMBIA UNIVERSITY

EVIDENCE FROM INSULAR FLORAS AS TO THE
METHOD OF EVOLUTION

EVIDENCE as to the rôle which hybridization plays in evolu-
tionary change may be obtained from various insular floras by
a comparative study of the history of those plant types in
them which are prevailingly self-fertilized and those which are
prevailingly cross-fertilized, both as to the rapidity with which
new local species are produced and as to the frequency with
which old species disappear. With these points in view, analy-
ses have been made of the vascular plants in the floras of eight
islands or island groups: Ceylon, Mauritius, Socotra, New Zea-
land, Hawaii, Galapagos, Juan Fernandez and St. Helena.* In
all these there is a conspicuous, often predominant, element
in the flora which is strictly local or endemic, indicating that
each island has been the theater of considerable evolutionary
change.

Information is necessarily lacking as to the method of fertili-
zation of most of the species, but our general knowledge of the
reproduction of the higher plants allows us to divide them into
three main types. The dicotyledons and petaliferous monocoty-
ledons, possessing floral organs which in the great majority of
cases are attractive to insects, are doubtless prevailingly cross-
pollinated. In the glumaceous monocotyledons, on the other
hand (chiefly Graminee, Cyperacex and Juncacee), the floral
organs are not so constructed as to favor cross-pollination, and
it will probably be agreed that crossing is much less common

* These analyses are based on the following authorities: Trimen, Hand-
book of the Flora of Ceylon; Baker, Flora of Mauritius and the Seychel-
les; Balfour, Botany of Socotra; Cheeseman, Manual of the New Zealand |
Flora; Hillebrand, Flora of the Hawaiian Islands; Stewart, Botany of the
Guipa Islands; Johow, Flora de las Islas de Juan Fernandez; Melliss,
St. Helena; and Hemsley, Report on the lhe saa cui Expedition:
Botany.

270 THE AMERICAN NATURALIST [ Von. LIT

among them than in the petaliferous types. Finally, in the
vascular eryptogams, the very frequent occurrence of bisexual
gametophytes seems to insure a still greater prevalence of self-
fertilization.

The vascular flora of each island was divided into thea three
groups which were studied comparatively. Determination was
first made as to the percentage of local or endemic species in each
group. This degree of endemism provides us with a rough meas-
ure of the extent to which new forms have been developed on the
island, and thus allows us to compare the rapidity of evolution
in one floral group with that in the others. In the following
table are set forth the percentage of endemic species in each of
the three main groups which we have mentioned, and for each
of the islands:

TABLE I
PERCENTAGE OF ENDEMIC SPECIES IN VARIOUS FLORAL ELEMENTS
3 p 2 ha g
£ É El g< | 3 oe | £8 E
31318 1881 s [393 939| 5
ie ce ie E Š i
pags trae and Petaliferous Mono- mi
cido o ey A g 0 0
Pie. 65%
Glumaceous Monocotyledons ......... 11 ¡14 S. 56 160 130 137 87
Average, 38%
Vascular Cryptogame. i.o oimn le $120 510 180.151 DS 44
rage, 23%

It is evident that the proportion of endemic species is much
higher among those types which we have reason to believe are
prevailingly crossed than among those which are prevailingly
selfed, being highest among dicotyledons, lower among gluma-
ceous monocotyledons and lowest among vascular cryptogams.
The same fact appears among genera, for 95 per cent. of the
endemic genera of these islands belong to petaliferous types and
only 5 per cent. to the glumaceous monocotyledons and vascular
eryptogams. These facts all point to the importance of hybridi-
zation as a factor in the production of new species.

The other aspect of evolutionary change, namely the disap-

10f course not all the endemic forms can be regarded as of local origin,
since certain of them may be isolated relicts of types formerly more widely
spread. The proportion of these, however, which have not subsequently

undergone specific change, and thus developed true local types, is probably

No. 615] SHORTER ARTICLES AND DISCUSSION 271

pearance of species, seems also to be influenced by the method
of fertilization. Many of the genera which are themselves not
endemic on any island are nevertheless represented there now
only by endemic species. In such cases it seems clear that the
first representative of the genus to invade the island has since
disappeared there entirely and been replaced by local species.
Table II gives the percentage of such genera (not endemic but
represented only by endemic species) for each of the three plant
types which we have discussed and for all the islands.

TA
PERCENTAGE OF THE NON-ENDEMIC GENERA WHICH ARE REPRESENTED ONLY
Y ENDEMIC SPECIES

pay vind a Got ee
a 5 2 = 5 g
RIA IE
19173 e Sai ans
| = Z e] E A
Dicotyledons and petaliferous mono-|
DO o A aso 9% 128% 29% (44% 57% | 16% 52% |100%
Average, 42 %
°
Glumaceous porro ee A FONE I2 IO 40 411.788 T 83
age, 26 |
TER Sense A oa cre « 3 0 ¡0 ae 18 |.0 19 25
, 10.5 % | i

It is evident that genera in nibh A ihe ‘original species’’ has
become extinct are proportionally commonest among dicoty-
ledons, less common among glumaceous monocotyledons and rare
among vascular eryptogams, thus suggesting that hybridization
has resulted in the ‘‘swamping out'” of the early forms. If local
adaptation and natural selection alone were at work, it is hard
to see why extinction should not be equally common in all these
groups. The facts point to the importance of hybridization in
- completely altering specific type when a group of individuals
have been isolated from the main body of the species.

Against the soundness of these conclusions several points may
be urged. Vascular eryptogams are perhaps inherently less
variable and quick to produce new species than flowering plants.
It may be, too, that eross-fertilization is much more common
among them than is generally believed. Whether the recognized
“species”? among these plants is the equivalent of the “species”
among angiosperms, or is a much more inclusive group, is also
a matter of doubt. These points can not well be brought against
the glumaceous monocotyledons, however, as contrasted with the
petaliferous types. Whatever its interpretation, the fact seems

212 THE AMERICAN NATURALIST [VoL. LE

clear that among dicotyledons and petaliferous monocotyledons
new types are produced and old types lost much more quickly
than anywhere else in vascular plants, a fact which in the light
of our knowledge of methods of reproduction certainly supports
the view that hybridization has been a powerful factor in evo-
lutionary change.
SUMMARY

Evidence from a comparative study of endemism in various
elements of certain insular floras tends to show that among cross-
fertilized types new species are developed more rapidly and old
ones lost more frequently than among self-fertilized types, thus
emphasizing the importance of hybridization as a factor in evo-
lutionary change. Epmunp W. SINNOTT

CONNECTICUT AGRICULTURAL COLLEGE

A LAND PLANARIAN FOUND AT BERMUDA!

In 1902 Professor Verrill recorded (‘‘The Bermuda Islands,””
p. 436, Fig. 237), that there had been reported to him the find-
ing at Bermuda of a ‘‘worm’’ which appeared to be a land
planarian. With the possible exception of this worm, which may
have been a Bipaliwm, no land planarians have been seen at
Bermuda. While collecting earthworms, in September, 1917, I
obtained among moist decaying leaves in a ‘‘fertilizer pit’’ at
Point Shares, Pembroke Parish, a single specimen of a flatworm
which seems to be a species of Geoplana. The ‘‘pit’’ was in use
as a dumping ground for garden refuse, and since no land plana-
rians appear to be native to Bermuda, the worm may have been
introduced in company with plants. It was 50 mm. long and

mm. wide, pale greenish blue on the ventral surface,—which
bore a rather small oral sucker in the usual position,—the ground
color of the dorsal surface being a deeper shade of the same
greenish blue, but marked with two deep blue or black longi-
tudinal stripes running the whole length of the animal. Two
well-developed pigment spots were present, one on either lateral
margin of the anterior end. It is not impossible that this species
might become permanently colonized at Bermuda (although no
other specimens have been found), and this note may therefore
be of use in fixing the date of its earliest observed appearance.

. J. CROZIER
AGAR’S ISLAND, BERMUDA
1 Contributions from the Bermuda Biological Station for Research, No.

THE
AMERICAN NATURALIST

Vout. LIT. -~ June-July, 1918 Nos. 618-619

THE RÔLE OF REPRODUCTION IN EVOLUTION!
PROFESSOR E. M. EAST
BUSSEY INSTITUTION, HARVARD UNIVERSITY

Tue establishment of methods of reproduction which
maintain variation and inheritance mechanisms on a high
plane of efficiency is naturally a fundamental requirement
in organic evolution. Since, however, inheritance mech-
anisms presumably equivalent are common to every method
of reproduction, one should be able to interpret the evolu-
tionary tendencies in the matter by comparing their
effectiveness in offering selective agencies their raw ma-
terial. Some will hold this statement to be a self-evident
truth; others may maintain as strongly either that the
premises are wrong or that the conclusion is not justified
even if the premises be granted. Perhaps it is safer to
ply the middle course; if the case is not so obvious as a
Euclidian axiom, as a compensation rigorous proof may
be less difficult.

As a basis for argument, let us sketch the general trend
of reproductive evolution in plants and animals.

Ordinarily, one speaks of two types of reproduction
among organisms, asexual and sexual. This is a conven-
tion that has taken on the dignity of a ‘‘ folkway ’’ among
- biologists. Its employment should imply assent to the
proposition that the varied forms in which each of these
classes presents itself are inherently equivalent, and that

1 Read by title at the Symposium of the American pod of Naturalists
on the subject ‘‘ Factors of Organic Evolution,’’ Jan. 5,

274 THE AMERICAN NATURALIST [ Vou. LIT

the groups considered as units are fundamentally distinct,
but it is doubtful whether any such implication would be
admitted by the majority of its users. In fact one could
hardly maintain that simple division, sporification, the
production of gemmules, true budding, fragmentation
with regeneration of parts, and the various kinds of
apogamy and parthenogenesis on the one hand, and all
nuclear fusions on the other, can be grouped together as
if they are of the same evolutionary value, if this term be
used in any narrow or special sense; but from a broader
viewpoint, the conventional classification has a real and
deep meaning which perhaps the biologist has grasped
instinctively. |

There are both asexual and sexual methods of repro-
duction in nearly all groups of animals and plants; among
animals the second has almost supplanted the first, among
plants the two have continued side by side. In neither
kingdom was sex developed as a more rapid means of
multiplication, since, as Maupas showed, a single infuso-
rian may become the progenitor of some 50,000 individuals
during the time necessary for one pair to conjugate.
Some other requirement was fulfilled; and fulfilled ade-
quately if we may judge by the number of times sexual
differentiation arose and the tenacity with which it was
retained.

Just when sexual reproduction first originated in the
vegetable kingdom is still a question. Among the lower
forms only the schizophytes, flagellates and myxomycetes
have passed it by. Perhaps it is for this reason that
these forms have remained the submerged tenth of the
plant world. It is tempting, as Coulter (1914) says, to
see sex origin in the Green Alge. There, in certain
species, of which Ulothria is a good example, spores of
different sizes are produced. Those largest in size germi-
nate immediately under favorable conditions and produce
new individuals. Those smaller in size also germinate
and produce new individuals, but these are small and

their growth slow. Only the smallest are incapable of

Nos. 618-619] THE ROLE OF REPRODUCTION 275

carrying on their vegetative functions. These come to-
gether in pairs. Two individuals become one as a pre-
requisite to renewed vigor. Vegetative spores become
gametes. Something valuable—speed of multiplication
—is given up for a time that something more valuable in
the general scheme of evolution may be attained.

This is indeed an alluring genesis of sex. Let us use
the indefinite article, however; no doubt it is a genesis of
sex, but it can havdly be the genesis of sex. Various mani-
festations of sex are present in other widely separated
groups of unicellular plants, the Peridinex, the Conjugate
and the Diotome*—+the Conjugate being indeed the only
great group of plants in which there is no asexual repro-
duction. In these forms one can not make out such a good
case of actual gametic origin, but the circumstantial evi-
dence of sex development in parallel lines is witness of its
paramount importance.

After the origin of sex, many changes in reproductive —
mechanisms occurred in plants, but almost all of them
resulted merely in greater protection of the gametes, in
increased assurance of fertilization, or in provision for
better distribution. First there was a visible morpho-
logical differentiation of gametes, the one becoming a
large inactive cell stored with food, the other becoming
small and mobile. Then came the evolution of various sex
organs, and finally the alternation of generations. In the
higher plants a long line of changes have occurred con-
nected with the alternation of generations; the spore-pro-
ducing type has developed from a form of little impor-
tance to that which dominates the vegetable world, the
gamete-producing type has degenerated until it consists
of but two or three cell divisions. In these variations
there is reproductive insurance, something which also may
be said of those manifold adaptations which provide
zygotic protection either in the seed or the adult plant,
but they are no more direct changes in reproductive
Mechanism than are the diverse means which arose to

secure dispersal. In fact in all of these changes no new

276 THE AMERICAN NATURALIST [ Vou. LII

process of fundamental evolutionary significance oc-
curred, unless it be the various mechanisms devised to
promote or to insure cross-fertilization, and which may be
interpreted as variations tending to perfect sexuality.

Coincident with the general trend of plant evolution
just mentioned, two important changes in the nature of
retrogressions occurred, which have persisted in many
species. A new type of asexual propagation arose,
apogamy, which though it appeared under several guises,
apogamy in the narrow sense, partl genesis and poly-
embryony, is none the less asexual reproduction returned
under another name and apparently with no particular
advantages over the older types. Further, hermaphrodit-
ism was developed and has persisted in numerous lines.
We may be wrong in calling hermaphroditism a retro-
gression, for it has the great advantage of a certain
economy of effort in the production of gametes, but never-
theless it is certainly a change which per se is in the
opposite direction from that established when sex was
first evolved. A moment of consideration not only makes
this clear, but gives us a pretty satisfactory proof that
the gain made when continuous multiplication was halted
for a time by the intervention of a fusion at the genesis
of sexual reproduction was in some way connected with
the mixture of dissimilar germplasms. This conclusion
is hardly avoidable from the fact that although herma-
phroditism retained the cell fusion mechanism of gono-
chorism it was still necessary for Nature to evolve means
for eross-fertilization. And the multitude of ways in
which she solved this problem must mean that an im-
mense advantage was secured.

In spite of the great morphological differences between
animals and plants, the essential evolutionary changes
affecting reproduction in the two kingdoms have been
so similar as to be almost uncanny. ‘Accepting the divi-
sion of animals into twelve phyla as recognized by many
modern zoologists (Parker and Haswell), one finds the
following facts regarding reproduction. Asexual repro-

Nos. 618-619] THE ROLE OF REPRODUCTION 277

duction in the narrow sense is common in Protozoa, Porif-
era, Coelenterata and Platyhelminthes, and is sporadic in
Molluscoida, Annulata, Arthropoda and Chordata. If
fragmentation and regeneration be included, Echinoder-
mata and possibly Nemathelminthes are added. If
parthenogenesis is included, Trochelminthes is admitted.
Thus only the Mollusca have no form of asexual reproduc-
tion, and zoologists would hardly feel safe in maintaining
its absence there since the life history of so many forms is
unknown. This being the case, one must admit that
asexual reproduction has been found satisfactory for
most of the great groups of animals as far as actual
multiplication is concerned. For other reasons, however,
it evidently did not fulfill all requirements, since sexual
reproduction is established in every phylum. Further;
omitting the Protozoa in which it is difficult to decide such
sexual differences, gonochorism is present everywhere
except in the Porifera, and hermaphroditism everywhere
except in the Trochelminthes, although in Nemathel-
minthes, Echinodermata and Arthropoda it is rare.
Now if our conclusions regarding the true róle played
by sex in evolution are correct, hermaphroditism is a
secondary and not a primitive phenomenon. In this we
follow Delage, Montgomery and Caullery rather than the
majority of zoologists. We believe it to be the only
logical view in spite of the fact that the Porifera, usually
considered so unspecialized, are all hermaphroditic.
Perhaps the Porifera are farther along in specialization
than is admitted, for to find the substance nearest chemi-
cally to the so-called skeleton of the sponges one must
tùrn to the arthropods (the product of the spinning glands
of certain insects). Hermaphroditism, therefore, as in
plants, is from this viewpoint a regression. And as in
plants it was not found adequate. In giving up diecism
for monecism, something was lost, and this something had
to be regained by further specialization. Hence, even as
in the vegetable kingdom one finds the essential feature of
bisexuality, mechanisms providing for mixtures of dif-

278 THE AMERICAN NATURALIST [ Vou. LIT

ferent germplasms, restored by means of protandry,
protogyny or self-sterility.

In even such a brief consideration of the more im-
portant changes which have occurred in the reproductive
mechanisms of animals and plants, one thing stands out
impressively. Both animals and plants have adopted as
the most acceptable and satisfactory modes of reproduc-
tion, methods which are identical in what we deem to be
the essential features, something that can be said of no
other life process. These significant features are the
preparation of cells which in general contain but half of
the nuclear material possessed by the cells from which
they arise, which are differentiated into two general
classes that show attraction toward each other, and which
will fuse together in pairs to form the starting point of a
new organism. This parallel evolution is of itself valid
evidence of the importance of the process. Let us return
to our original proposition for its interpretation.

First, is there any evidence that sexual reproduction
differs from asexual reproduction in what may be called
the heredity coefficient? In other words, does one method
hold any advantage over the other as an actual means for
the transmission of characters? I have answered this
question in the negative, but it must be confessed that the
basis for this answer is a long and intimate experience in
handling pedigree cultures of plants rather than the study
of a large amount of quantitative data bearing directly
on the problem. Quantitative data are to be found, of
course, and plants furnish the best material because of the
ease in handling large numbers of both clons and seedlings
side by side; but even with the best of plant material,
several undesired variables are present. Practically the
inquiry must take the form of a comparison between the
variability of a homozygous race when propagated by
seeds and when propagated by some asexual method.
The first difficulty is that of obtaining a homozygous race
and thus eliminating Mendelian recombination. The
traditionally greater variability of seed-propagated

Nos. 618-619] THE ROLE OF REPRODUCTION 279

strains is due wholly to this difficulty, I believe. It may
be impossible to obtain a race homozygous in all factors.
There may be a physiological limit to homozygosis even
in hermaphroditic plants. The best one can do is to use
a species which is naturally self-fertilized, relying on con-
tinued self-fertilization for the elimination of all the
heterozygous characters possible. I have examined many
populations of this character in the genus Nicotiana and
have been astounded at the extremely narrow variability
they exhibit. Even though one can not grow each mem-
ber of such a population under identical conditions as
to nutrition, the plants impress one as if each had been
cut out with the same die. Qualitative characters such as
color show no greater variation, as far as human vision
may determine, than descendants of the same mother
plant propagated by cuttings. Further, in certain char-
acters affected but slightly by external conditions, such
as flower size, the sexually produced population not only
shows no greater variability than the asexually produced
population, but it shows no more than is displayed by a
single plant. ‘Yet one must remember that in such a test
the seeds necessarily contain but a small quantity of
nutrients, and for this reason the individual plants are
produced under somewhat more varied conditions than
those resulting from cuttings, hence it would not have
been unreasonable to have predicted a slightly greater
variability for the sexually produced population even
though the coefficient of heredity of both were the same.

I have made similar though less systematic observa-
tions on wheat—an autogamous plant almost as satis-
factory for such a test as Nicotiana—with practically iden-
tical results. I do not know of any published data on the
subject, however, taken either from these or any other
plants. In fact, there are few other plants from which
data could be obtained with so little likelihood of experi-
mental error.

On the other hand, zoology has furnished a consider-
able amount of such evidence (cf. Casteel and Phillips,

280 THE AMERICAN NATURALIST [ Vou. LIT

1903; Kellogg, 1906; Wright, Lee and Pearson, 1907).
One need only mention Kellogg’s work on bees as a type.
Kellogg assumed that if amphimixis were the principal
cause of the continuous variations postulated by Darwin
and Weismann as the most important source of material
for the use of the natural selection,? parthenogenetically
produced individuals should be less variable than those
produced sexually. /A' statistical investigation showed,
however, that the characters of drones probably are more
variable than those of worker bees of the same race.
Since Kellogg believes Darwin’s judgment that “males
vary more than females”? to have been disapproved, he
concludes that ‘‘amphimixis is not only not necessary in
order to insure Darwinian variation, but there is no evi-
dence (that I am aware of) to show that it increases
variation.”

It is hardly necessary to point out here the numerous
mathematical and biological pitfalls which should be con-
sidered before one could accept as valid the statistical
differences that appear to exist when coefficients of varia-
tion based on such data are examined. It should suffice to
note that the researches of Wright, Lee and Pearson
(1907) on wasps of the species Vespa vulgaris showed
just as great a difference in variability between workers
and drones in favor of the former. Apparently, the sta-
tistics in these two nearly related groups lead to opposite
conclusions; in reality probably neither statistical differ-
ence is significant as far as the question we are discussing
is concerned. The only conclusion justified by such data
would seem to be that the coefficient of herédity is as high
in the production of asexual as it is in the production of
sexual forms. |

Moreover, one can not expect anything more definite
from this method of attack. Biologists may differ as to

21t should be noted here that all parthenogenetic eggs are not mere
spores. Some preparation often occurs through the emission of one polar
body. This may be merely a kind of recapitulation, a vestigial process no
longer having any significance whatever, but since we are not certain it
seems to the writer that the evidence from plants at present must be re-
garded as stronger.

Nos. 618-619] THE ROLE OF REPRODUCTION 281

the definition of fluctuation, mutation, etc., but they are
generally agreed that germinal variations, be they great
or small,are in most species so rare they can not be gauged
by the use of ordinary statistical methods. For this rea-
son, a comparison between the variability of the drones
and of the workers of a pure race of bees is not likely to
show any difference between these two modes of repro-
duction in the matter of the frequency or the type of the
germinal variation produced, and can not answer the ques-
tion as to whether sexual reproduction contributes more
material for the use of natural selection than asexual re-
production. A study of variability in crossed races,
where the effect of Mendelian recombination can be con-
sidered, would be a more logical attack upon the second
problem, but is hardly necessary in view of the other
evidence available.
One is then justified in claiming there is no experimental
evidence to show that sexual reproduction in itself is not
an exact equivalent of asexual reproduction in the matter
of a heredity coefficient, but is this also true for germinal
= variation? ¡We believe it is. Variations there are in
both asexual and sexual reproduction, but it can not be
maintained that they occur more frequently in the latter.
There are insects in Oligocene amber apparently identical
with those of to-day, proving that constancy. of type is
possible through long periods of time under sexual repro-
duction; yet germinal variations occur to-day in some-
what noteworthy numbers, as Morgan’s work on Dro-
sophila shows, although the proportion of these varia-
tions which show possibilities of having an evolutionary
value, as evidenced by persistence in natural types, is
probably small. On the other hand, the number of varia-
tions produced under the dominance of asexual repro-
duction can not be said to be less numerous, even among
organisms of a relatively high specialization. If there
are those who doubt this statement, let them refer to the
immense list of bud-variations in the higher plants com-
piled by Cramer (1907).
There would be little reason in ogashing the claims

282 THE AMERICAN NATURALIST [ Von. LIL

further, since even though there does not seem to be a
sufficient difference between sexual and asexual reproduc-
tion in the matter of variation frequency to make it a
subject of experimental proof, certain theoretical points
raise the suspicion that there is such a difference. All we
would maintain is that to account for the general persist-
ence of sexual reproduction by such a cause, the differ-
ence in its favor should be so great that it could easily be
determined experimentally. Since this is not true, we
believe the hypothesis should be discarded.

The points of theory referred to are these. It will be
allowed by all that there is some considerable evidence of
the chromosomes being the most important conservators
of hereditary factors—the physical bases of heredity in
whatever form they may be. If it is assumed then that
changes in constitution in these cell organoids are fol-
lowed by changes in type, and that such changes in con-
stitution are equally probable in all chromosomes, it
follows that parthenogenetic individuals having the hap-
loid number of chromosomes should show a larger propor- '
tion of germinal variations than members of the same
species having the diploid number of chromosomes, be-
cause variations of all kinds should be recognizable in the
former case, while in the latter, recessive variations could
not be detected until the first or second filial generation,
and then only when the proper mating was made. ‘There
is some evidence that this reasoning is not wholly improb-
able. But variations occur much more frequently in
heterozygotes than in homozygotes. To me this simply
means that bud-variations are detected more frequently
in heterozygotes than in homozygotes: and an interpreta-
tion is not hard to find. Retrogressive variations are
much more frequent than progressive variations, and a
retrogressive variation in a particular character shows
only when the organism is heterozygous for that character.
If a retrogressive bud-variation arises in a homozygote
and gametes are afterwards developed from the sporting
branch it is not at all unlikely that the variation may show

in the next generation, but it will be attributed then to

Nos. 618-619] THE ROLE OF REPRODUCTION 283

gametic mutation. If one compares asexual and sexual
reproduction from the standpoint of frequency of varia-
tion only, then sexual reproduction may seem to hold the
advantage over asexual reproduction in the usual sense;
but parthenogenesis, which is certainly a form of asexual
reproduction, is in theory better adapted than sexual re-
production for giving large numbers of variations.

If, therefore, one is constrained to agree that the bulk

of the evidence points to a practically identical coefficient
of heredity for both forms of reproduction, and that varia-
tion in the sense of actual changes in germinal constitu-
tion may occur with greater frequency in asexual repro-
duction, if there is any difference at all between the two
` forms, he is driven either to the ‘conclusion of Maupas
that continued asexual reproduction is impossible through
some protoplasmic limitation or to the conclusion of Weis-
mann that a mixture of germplasms offers sufficient ad-
vantages to account for everything. This is the dilemma®
unless one wishes to maintain that efficient mechanisms
-for nutrition, adaptation, protection and distribution
could not be evolved or maintained under asexual re-
production.

The contention of Maupas can not be dealt with experi-
mentally any more successfully than the question as to
the inheritance of acquired characters since experimental
time and evolutionary time are not of the same order of
magnitude. The long-continued experiments of Wood-
ruff in which vigorous strains of paramecium have been
kept dividing asexually for several thousand generations,
however, as well as the botanical evidence that numerous
species having no sexual means of multiplication have
continued to exist during long periods of time, weight the
balance against him. One need not hesitate to concede
that all of these organisms are rather low unspecialized
types; the modern development of genetics has built up
such a solid structure in favor of Weismann’s view that
there is little need of argument along the older line. ~

3 Naturally another hypothesis wholly new to biology may be submitted
at any time.

284 THE AMERICAN NATURALIST [ Vou. LII

The main argument in favor of Weismann’s viewpoint
does not take long to state. It is this: Mendelian heredity
is a manifestation of sexual reproduction. Wherever
sexual reproduction occurs, there Mendelian heredity will
be found. The very fact that it describes the sexual
heredity of both animals and plants is sufficient proof of
its generality in this regard. Now if N variations occur
in the germplasm of an asexually reproducing organism,
only N types can be formed to offer raw material to selec-
tive agencies. But if N variations occur in the germ-
plasm of a sexually reproducing organism 2” types can be
formed. The advantage is almost incalculable. Ten
variations in an asexual species mean simply 10 types, 10
variations in a sexual species mean the possibility of 1,024
types. Twenty variations in the one case is again only 20
types to survive or perish in the struggle for existence; 20
variations, in the other case, may present 1,032,576 types
to compete in the struggle. It is necessary to hedge the
argument by pointing out that these figures are the maxi-
mum possibilities in favor of sexual reproduction. It is
improbable that they ever actually occur in nature, for 22
types really to be found in the wild competing for place
after only 20 germinal variations would mean an enor-
mous number of individuals even if the 20 changes had
taken place in different chromosomes, and if the varia-
tions were linked at all closely in inheritance the number
required would be staggering. But there are breaks in
linked inheritance, and the possibility is as stated.

These advantages remain even though it should be shown
later that the more fundamental and generalized char-
acters of an organism are not distributed by Mendelian
heredity. Loeb (1916) believes that the cytoplasm of the
egg is roughly the potential embryo and that the chromo-
somes, distributed as required by the breeding facts of
Mendelian heredity, are the machinery for impressing the
finer details. There is something to be said for this point
of view, though at present it is but a working hypothesis.
But granting its truth it does not detract from the ad-
vantages gained by sexual reproduction. Even the most

Nos. 618-619] THE ROLE OF REPRODUCTION 285

strict mutationist would hardly maintain that evolution in
general has come about through tremendous changes in-
volving sterility between the mutant and the parent types.
It seems unnecessary to deny such possibilities; but the
weight of evidence is in favor of the majority of varia-
tions being comparatively small, changes in detail, the
very kind which are known to be Mendelian in their in-
heritance.

Yet sexual reproduction in itself does not assure these
advantages, though they are based upon it. There must
be means for the mixture of germplasms. This oppor-
tunity was furnished originally by bisexuality. Then
came hermaphroditism, manifestly an economic gain, yet
on the whole unsuccessful except as functional bisexuality
was restored by self-sterility, protandry, protogyny or
mechanical devices which promoted cross-fertilization.

The prime reason for the success of sexual reproduc-
tion then, as Weismann maintained, is the opportunity it
gives for mingling germplasms of different constitution
and thereby furnishing many times the raw material to
selective agencies that could possibly be produced through
asexual reproduction. Further, there are three minor ad-
vantages which rest upon the same mechanism. They are
minor advantages only when compared to the major, and
should not be passed by.

Let us first consider heterosis, the vigor which accom-
panies hybridization. This phenomenon has long been
known. It is characteristic of first generation hybrids
both in the animal and vegetable kingdoms. It affects the
characters of organisms in much the same manner as do
the best environmental conditions. in other words, the
majority of characters seem to reach the highest de-
velopment in the first hybrid generation. ‘The hybrid in-
dividual therefore holds some considerable superiority
over the individuals of the pure races which entered into
it, and is thereby the better enabled to survive and to
produce the multiplicity of forms which its heterozygous
factors make possible. The frequence of this phe-
nomenon, for it is almost universal, together with the fact

286 THE AMERICAN NATURALIST [ Vou. LII

that it seems impossible to fix the condition, led Shull and
the writer independently to the conclusion that certain
factors in addition to their functions as transmitters of
hereditary characters also had the faculty of carrying
some sort of a developmental stimulus when in the hetero-
zygous condition. The recent work of Morgan on linked
characters, however, makes it possible to give another
interpretation, as Jones (1917) has demonstrated. If it
be assumed that several variations have occurred in each
of one or more chromosomes, then it can be shown that the
first-generation hybrid between such a variant and the
race from which it arose will bring together all dominant
or partially dominant characters. In the second hybrid
generation, on the other hand, Mendelian recombination
steps in and makes it improbable that many individuals
shall have such a zygotic composition. And only in the
rare cases where the proper breaks in linkage have oc-
curred can a homozygous individual of this type be
produced.

The latter hypothesis holds the advantage that it
furnishes hope for a homozygous combination as valuable
as that of the first hybrid generation no matter how
rarely it may be assumed to occur, but whether it holds
for the majority of organisms or not may depend on a
future decision as to the frequency of side-by-side
synapsis as compared to end-to-end synapsis. Our knowl-
edge of linkage rests almost entirely on Morgan’s work
on Drosophila where side-by-side synapsis occurs at the
maturation of the germ cells. If the break in linkage be-
tween groups of characters apparently carried by a single
chromosome, which Morgan finds to be so exact in Dro-
sophila, should actually depend on Jannsen’s theory of
chromosome, twisting at synapsis, then some other type
of inheritance may be found in species having end-to-end
synapsis. Perhaps this is the reason why the (Enotheras
have such a peculiar heredity, for in them Davis (1909)
thinks end-to-end synapsis prevails. But, be this as it
may, the vigor of first generation hybrids is a fact and

not a theory, and the advantage it brings to the hetero-

Nos. 618-619] THE ROLE OF REPRODUCTION 287

zygotic individual in competition with its fellows can not
be gainsaid.

The investigations of Shull and of the writer on the
effects of cross- and self-fertilization have brought to
light another series of facts with a bearing on the problem
under discussion. It has been shown that the apparent
deterioration of cross-bred species when self-fertilized is
in large measure and perhaps wholly due to the loss of
hybrid vigor* through the formation of homozygotie
Mendelian recombinations and not an effect of inbreeding
per se because of the union of like germplasms. This is
a plausible argument against Darwin’s idea that con-
tinued inbreeding is abhorrent to Nature. It may even
be said to be a valid reason for declining to accept
-` Maupas’s belief in the impossibility of continued asexual
reproduction, for there is no very good reason for dis-
tinguishing between continued asexual propagation and
continued self-fertilization. Inbreeding simply brings
about the opposite effect from crossing, and we can see
no reason for the comparative failure of naturally inbred
types in the wild other than the lack of chances for
progress. The one is the conservative manufacturer who
continues the original type of his article, the other is the
progressive who makes changes here and there without
discouragement until the acceptable improvement is
found. In fact, if this argument be overlooked, the in-
bred types which have persisted hold some advantages
over the cross-bredtypes. The self-fertilized species are
inherently strong and vigorous, witness tobacco and
wheat. They stand or fall on their own merits. They
are unable, as are cross-bred species, to cover up in-
herent weakness by the vigor of heterozygosis. Cross-
fertilized maize has become the king of cultivated plants
because of its variability, but many of our best varieties
carry recessive characters very disadvantageous to the
species.

The next secondary advantage of sexual reproduction is

4 Accepting the view that the vigor of the first hybrid generation is due
to dominant characters meeting makes this argument even more forcible.

288 THE AMERICAN NATURALIST [VoL. LIT

the division of labor made possible by secondary sexual
characters, using the term very, generally and including
even such differences as those which separate the egg
and the sperm. Itis not known just how these differences
arose or by what mechanism they are transmitted. The
greatest hope of reading the riddle lies in an investiga-
tion of hermaphroditic plants, for there are technical
difficulties which seem to preclude their solution in ani-
mals. For example, breaks in the linkage between sex-
linked characters occur only in the female in Drosophila,
and as the sex chromosome is double in the female, it
can not be determined whether the differentiation be-
tween male and female is due to the whole chromosome or
not. But this ignorance does not give reason for a denial
of the great advantage which sexes bearing different
characters hold over sexes alike in all characters except
the primary sex organs.

The only glimpse of the truth we have on these matters
comes from recent work on the effect of secretions of the
sex organs on secondary sexual characters. The effect of
removing the sex organs and the result of transplanting
them to abnormal positions in the body have shown that
in vertebrates the secretions of these organs themselves
activate the production of the secondary sexual char-
acters. This does not seem to be the case in arthropods,
however, so one can not say that primary sexual differ-
entiation and secondary sexual differentiation is one and
the same thing. !

Finally there is a presumable advantage in gonocho-
ristic reproduction in having sex-linked characters. We
say presumable advantages, for all of the relationships
_ between sex and sex-linked characters are not clear. The
facts are these: One sex is always heterozygous for the
sex determiner and the factors linked with it. Now it
may very well be that there is an actual advantage in the
heterozygous condition, as we have seen above. But
should the so-called vigor of heterozygosis prove to be
only an expression of the meeting of dominant characters,
still a possible advantage accrues to this phenomenon be-

Nos. 618-619] THE ROLE OF REPRODUCTION 289

cause the mechanism contributes toward mixing of germ-
plasms. As an example, let us take the Drosophila type
of sex determination. There the sperm is of two kinds:
the one containing the sex chromosome and its sex-linked
factors, the other lacking it. The eggs are all alike, each
bearing the sex chromosome. It follows then that the
male always receives this chromosome from his mother,
who may have received it from either her father or
mother. Moreover, further variability may be derived
from the linkage breaks which occur always in the female.
This last phenomenon is hardly worthy of special men-
tion, however, until it is shown to be typical of such
reproduction.

This short reconnaissance presents the pertinent facts
in the situation as they appear to the writer. A very
great number of interesting things connected with repro-
duction during the course of evolution have not been men-
tioned. This is because it is felt that the essential feature
of the róle of reproduction in evolution is the persistence
of mechanisms in both the animal and plant kingdoms
which offer selective agencies the greatest amount of raw
material. Other phenomena are wholly secondary.

LITERATURE CITED
Casteel, D. B., and Phillips, E. F. 1903. rs amine the Variability of
Drones daa Workers of the Honey Bee. Biol. Bull., 6: 18-37.
Coulter, J. M. 1914. The Evolution of Sex in "Plant Chicago, Uni-
1-

versity of Chicago Press. Pp.
Cramer, P. J. S. 1907. Kritische Ubersicht der bekannten Fille von
Knospenvariation. Natuurkundige a eae van de Hollandsche

Maatschappij der Wetenschappen. Derde eto Deel VI, Derde
Stuk. Haarlem, De Erven o pp. iii—xviii + 4
me hie M. 1909, Cytological Studies on (Enothera. T Ann, Bot, 23:

Thy i. r 1917. Dominance of Linked Factors as a Means of Account-
ing for Heterosis. Genetics, 2: 466—479.

Kellogg, L A 1906. Variation in Parthenmogenetie Insects. Science,
N: S; 695-699.

Loeb, J. on The Organism as a Whole. N. Y., Putnam. Pp. v-x+
379.

Wright, A., Lee, A., and Pearson, K. 1907. A Cooperative Study of
Queens, Drones nd Workers in Vespa vulgaris. Biometrika, 5: 407-
422,

CONTINUOUS AND DISCONTINUOUS VARIA-
TIONS AND THEIR INHERITANCE IN
PEROMYSCUS. II

DR. F. B. SUMNER

SCRIPPS INSTITUTION, LA JOLLA, CALIF.

IV. Herepiry oF THE RACIAL DIFFERENCES

In view of the long-recognized correlation between cer-
tain of these subspecific characters—namely, those relat-
ing to pigmentation—and certain factors of the physical
environment, the possibility has suggested itself that the
characters in question might be purely *““ontogenetic,””
1. e., produced anew in each generation by the action of
external physical factors. The simple experiment of
transplanting mice from one habitat to another has dis-
posed of this suggestion.

As I have more than once reported elsewhere (1915a,
1917, 19176) entirely negative results have been reached,
so far as climatic influences are concerned. Neither the
transference of the desert race to Berkeley, nor the trans-
ference of both the desert and the redwood races to La
Jolla have resulted in any demonstrable change, at least
up to the third cage-born (‘‘C,’’) generation.

The La Jolla test is the more satisfactory of the two,
since the number of animals employed is very much
greater. Thus far, however, only the C, animals (38
rubidus and 96 sonoriensis) have been killed, measured
and (in part) skinned. The C, generation is still kept
alive for breeding purposes, but the characteristic racial
differences are obvious. On comparing the skins of the
palest and the darkest rubidus, or the palest and darkest
sonoriensis of the C, generation, with the extremes among
_ the wild grandparents, it will be seen that the range of
color variation has not appreciably changed.

Nos. 618-619] INHERITANCE IN PEROMYSCUS 291

Not only are the larger color differences which distin-
guish these main races heritable, but certain lesser differ-
ences which distinguish narrowly localized sub-races have
been shown to be genetic characters. In a recent paper
(1917) I discussed an aberrant colony of ‘‘rubidus’’ in-
habiting an isolated sand-spit fronting on the ocean.”
The evidence for the inheritance of these peculiarities of
color may now be stated somewhat more strongly than
was done in that paper. Upon preparing the skins of the
three C, members of this sub-race, born and reared at La
Jolla, it was found that all three were of the aberrant
hue.!*

As regards differences relating to the measurable parts,
certain preliminary explanations are necessary. It was
early found that the cage-born mice depart from the
wild type in certain rather striking respects. They are,
on the average, considerably smaller than the latter, and
have tails, feet and ears which are shorter not only abso-
lutely but relatively. In extreme cases these malforma-
tions may fitly be termed deformities. Not rarely, too,
the dorsal tail-stripe becomes so diffuse that definite out-
lines can no longer be distinguished. Measurements re-
veal the fact that this stripe becomes narrower, on the
average, in the cage-born animals. Furthermore, the fer-
tility of the captive generations is greatly reduced.

ese abnormal characteristics resulting from captiv-

ity are manifested much more strongly by the Eureka
race than by the desert one, or by the race which is native
to this locality (La Jolla). In fact, my original stock of
rubidus, consisting of over a hundred animals of the wild
generation, has dwindled down to one male and six fe-
males in the C, generation. In contrast to this, no diffi-
A there neglected to point out the analogy between this race and that

bred this form in the laboratory and found its peculiarities to be ‘‘ really
genetic.’’ Morgan’s findings (1911) in respect to this question appear to
have been complicated by the appearance of artificially induced abnor-
malities.

16 In my earlier statement, based upon the appearance of the living
animals, it was said that two of the three were exceptionally pale.

292 THE AMERICAN NATURALIST [Vou. LII

culty has been found in maintaining approximately the
original numbers of the other two races, despite the steril-
ity of a large proportion of the individuals. Here, then,
we may note in passing, is another interesting racial dif-
ference of a physiological nature.

This deterioration of the stock, it must be pointed out,
is progressive. Each generation probably presents more
abnormalities than the preceding one. The causes of this
condition are at present entirely unknown to me. Mal-
nutrition or intoxication, resulting from pathogenic bac-
teria or protozoa in the alimentary canal, may be men-
tioned as possibilities. Many of the animals are now
being reared in small open pens, where they are allowed
to burrow in the ground. A preliminary test of this plan
encourages us to hope that the troubles referred to may
thus be avoided.

The following table presents mean values for certain :
characters in the C, generation, for the three races which
are now being reared at La Jolla.

TABLE V

No. of | Body Foot Ear Tail-Stripe

| Sex| Cases | (Mm.) pk Ciel: ) | (Mm.) | (Mm.) | (Per Cent.)
Bn 22° Beis g 21 85.5 93.1 20.26 | 16.39 37.3
Q 13 87.2 91.5 20.30 | 16.90 39.5
La Jolla iio ds g 61 87.4 79.6 19.62 | 17.66 30.7
Q 89.4 79.1 19.56 | 17.74 abt
Victorville, ao SF 45 85.3 75.6 18.65 | 16.43 25.9
2 49 87.0 74.8 18.81 16.90 26.2

It is plain from this table that, in respect to the four
characters other than body length (tail length, foot
length, ear length and tail-stripe), the three races have
maintained the same relative positions in the series as
formerly. When arranged with reference to tail, foot
and tail-stripe, the series, as before,is: Eureka > La Jolla
> Victorville. Asregards ear length, the earlier arrange-
ment likewise holds, viz.:

cel cian eee forville.

Another significant fact does not appear from the fore-
going table, however. The modifications of the three

Nos. 618-619] INHERITANCE IN PEROMYSCUS 293

races, indicated by the consistent reduction of all these
values, has not affected them to an equal degree. In re-
spect to all four of the characters (body length not here
considered) the local (La Jolla) race has been least modi-
fied, while in respect to three of them (tail length, tail-
stripe and ear) the Eureka race has been most modified.
Thus there has been a mean reduction of 11 per cent. in/
the tail length of rubidus, 74 per cent. in that of sonorien-
sis, and only 3 per cent. in that of gambeli. There is,
therefore, a convergence between the Eureka and the La
Jolla races, and if I had only these two under comparison,
I might have been disposed to conclude that local condi-
tions had brought about a modification of rubidus in the
direction of gambeli. But the case of sonoriensis, which
actually diverges farther from the local race in the C,
than in the parent generation, shows that this explanation
is not the correct one. These differences in the degree
of modification are probably indices of the susceptibility
of these three races to the malign influences of captivity,
which have already been discussed. In harmony with
this view is the fact that the Eureka mice are likewise far
less fertile, under local conditions, than either of the
other races.

V. HEREDITY OF INDIVIDUAL DIFFERENCES WITHIN EACH
Race

It has been shown that a wide range of individual varia-
bility occurs within each race in respect to just those
characters by which one of these groups is distinguished
from another. These major differences which distin-
guish one race from another have been shown to be
hereditary. Is this likewise true of those minor differ-
‘ences which distinguish one individual from another of
the same race?

This question can be answered by the well-known
method of computing coefficients of parental-filial corre-
lation—the ‘‘coefficients of heredity’’ of Pearson. I re-
alize that the validity of this measure of the force of
heredity has been called in question,*” on the ground that

17 By Pearl (1911) and Johannsen (1913), among others.

294 THE AMERICAN NATURALIST [ Vou. LIT

it does nothing more than reveal the presence of geneti-
cally different strains in a mixed population. This, how-
ever, is exactly what I wish to do in the present case.
The fact that within a ‘‘pure line,” where the phenomena
of heredity should be least obscured, this coefficient is
said to be zero is entirely irrelevant to the present situa-
tion. What we wish to ascertain is the degree to which,
for example, long-tailed parents tend to have long-tailed
offspring. Whether these differences among the parents
are due to ‘‘mutations’’ or ‘‘fluctuating variations,””
whether they are due to single ‘‘unit factors”? or ‘‘multi-
ple factors”” or no factors at all, are admittedly matters
upon which these coefficients throw no light. Such ques-
tions must be decided upon other grounds.'®

In computing these correlations between parents and
offspring, we are restricted to characters which are inde-
pendent of the absolute size of the individual? Char-
acters which fulfil these requirements fairly well are the
relative tail length (ratio to body) and-the relative width
of the tail-stripe (ratio to cireumference). My data show
that the former is largely, and the latter almost wholly,
independent of the size of the mouse.

-he coefficients are given in Table VI. I have not
thought it worth while to include their probable errors,
since the significance of the set as a whole is indicated by
the magnitude of most of the figures and by the fact that
all but two out of the 24 are positive. The weighted
means of these coefficients, combining the four races and
two sexes, are: relative tail length, + 0,297 ; tail-stripe,
- 18 One objection, valid in certain cases, has been raised against the use of
this coefficient. I refer to the contention that it may reveal resemblances
which are due not to genetic relationship, but to environmental influences

were trapped in the same restricted area, while all the offspring were
reared in captivity under conditions which were practically identical for all.

Nos. 618-619] INHERITANCE IN PEROMYSCUS 295

TABLE VI
rubidus gambeli (La Jolla) sonoriensis
Tail Tail-Stripe Tail Tail-Stripe Tail Tail-Stripe
HALDOL da es +0.36 +0.08 +0.15 +0.36 +0.23 +0.26
Far separa Ae Fi gers +0.24 +0.27 +0.11 +0.53 +0.42 +0.17
Mother-son .......... +0.64 +0.57 +0.51 +0.36 +0.28 +0.26
diia AS +0.84 +0.47 —0.02 +0.51 +0.15 —0.02

+0.302. The average of four figures given by Pearson”
(father-son, father-daughter, ete.) for the heredity of
stature in man is + 0.335. There is thus found, in these
mice, approximately the same degree of resemblance be-
tween parents and offspring, in respect to these two char-
acters, as is found to occur in man in respect to stature.”

Since the heritability of these individual differences
has been proven by means of correlation coefficients, the
practicability of selection experiments with such charac-
ters is evident, provided that sufficient numbers of normal
animals can be reared. Experiments of this nature have
already been commenced by Mr. H. H. Collins and myself.
The characters chosen for these tests are coat color, tail
length and width of tail stripe.

VI. HYBRIDIZATION

Successful hybrid matings have been made (1) between
the Berkeley and the Victorville mice, (2) between the
Eureka and the La Jolla mice, and (3) between the Eu-
reka and the Victorville mice.

In the first case, moderate numbers of F, and F, ani-
mals were reared. In respect to coat color, the 25 F, in-
dividuals, when adult, ranged from a condition similar to
that of an average Berkeley gambeli to a condition as pale
and as yellow as that of the average sonoriensis. More-
over, there was here no marked tendency for the mean or
intermediate condition to preponderate numerically over
the extremes.

20 1900, p. 458.

ai It is likely that the abnormal condition of the cage-born mice, de-
seribed above, has resulted in minimizing this correlation. Some of the dif-
ferences among them are doubtless due, not to heredity, but to differences in
the incidence of the disturbing factor (infection?). The parents, having
grown mab or quite to maturity in the wild state, were not subjected to
this influen:

296 THE AMERICAN NATURALIST [ Vou. LIT

Now it is of interest to note that the F, generation, con-
sisting of 40 specimens, presented very nearly the same
range, in respect to coat color, as did the F,. The two
dark extremes were of an almost identical shade, as were
likewise the pale extremes.?? No argument for ‘‘segre-
gation’’ could be based upon this series which would not
apply with equal force to the F, series.

Twenty F, mice, resulting from random matings of the
F, animals, presented a range of variation which was
actually not as great as that observed in the F, genera-
tion. The smaller number may perhaps be responsible
for this difference.

It must be added that both of the parent races present
a rather wide range of variability as regards coat color,
and that series of the two overlap rather broadly in this
respect. This circumstance complicates our interpreta-
tions much more in the present case than in that of the
crosses between rubidus and sonoriensis, which will be
considered later.

The Berkeley and Victorville races have been found to
differ in only two of the characters which were subjected
to careful measurement. The former race has a broader
dorsal tail-stripe and slightly shorter ears. The second
of these differences is a trifling one, however, and is not
always evident when small series are compared. More-
over, the absolute length of the ear is largely dependent
upon the size of the body.

Thus the width of the tail-stripe is the only accurately
measured subspecific character which is available in con-
sidering hybrids between these two races. Table VII
gives the mean value and the variability of this character
in the two parent races** and in the F, and F, genera-
ations of hybrids.

- Despite the small numbers of animals here concerned,
two facts are of some interest. (1) The mean width of
_ the tail-stripe in both generations of hybrids is very
22 These remarks are based upon a comparison of skins, after death.
28 The parent mice here used were cage-born animals, which, as before
stated, were somewhat modified by captivity. They represent the limited
stock which was reared in Berkeley during the earlier stages of the ex-

Nos. 618-619] INHERITANCE IN PEROMYSCUS — 297

nearly midway between the mean widths found in the
parent races. (2) There is no increase in the variability

TABLE VII
sag hada paca Mean Standard Deviation
gambeli (Berkeley) ........ 23 33.5 4.09
ABC O A 38 26.5 4.57
al ce os eee s 24 30.2 3.96
He Ry binds 6) as ois 38 29.3 3.54

of this character in the F, generation. Indeed, the stand-
ard deviation for the latter animals chances to be the
lowest of the four values given.

It seems worth while to indicate the actual distribution
frequencies of the tail-stripe measurements for these four
groups of mice (Table VIII).

TABLE VIII
aver
BOS Suse
Berkeley |
gambeli ... 1 ETA 31213 2|2|4|1/1/1/33.5
sonoriensis ..l21212 1321421221413 |212|1|1|2/2 | | 26.5
F; hybrids...| | 1 1} lilalilalal1lal1j2/2 1 30.2
F> hybrids. .. 1l1l113l3l4lelelalsia/3l41111 29.3

In contrast to this case of the tail-stripe, it is interest-
ing to note that, in respect to relative tail length, the F,
generation shows a somewhat higher variability than the
F, generation or than either of the parent races. But it
so happens that this is a character in which the parent
races do not appreciably differ.

In reality, it is probable that none of these differences
in variability is significant, in view of the small numbers
of individuals concerned. The explanation offered for
certain differences between other hybrids to be discussed
later—namely, that the F, and F, generations differed in
the relative degree of abnormality—does not seem to ap-
ply here. The mean body length is about equal in the
two generations, as well as the mean length of tail and
foot.

The crosses between the Eureka ma the La Jolla races
have not been carried far enough to render any report
upon them a at present. |

298 ` THE AMERICAN NATURALIST [Vou. LIT

Hybridization of the Eureka with the desert mice was
first accomplished nearly two years ago, and thus far 30
F, and 20 F, animals have been reared to maturity and
measured. A very serious drawback has been the great
infertility, under local conditions, of rubidus and any-
thing having rubidus ‘‘blood.’’ Still more serious is the
abnormal state of a large proportion of the cage-born ani-
mals, which affects some of the very parts that we are
chiefly concerned with in these crosses. Fortunately, the
coat color remains nearly, or quite unaltered.

` My series of 30 F; skins, taken as a whole, presents a
condition about intermediate between that of the parent
races. They exhibit, however, a wide range of variation,
the darkest individuals being nearly as dark as some of
the palest wild specimens of the Eureka race, while the
palest individuals differ little in shade from a medium
mouse of the desert race (Fig. 13).

In the F, generation we meet with a range which is
little, if any, greater. The darkest skin** is somewhat
darker than the darkest in the F, generation. On the
other hand, the palest skin is scarcely as pale as the palest
in the F,. The preponderating effect is that the hybrids
of the second generation, like those of the first, are inter-
mediates. If these differences of coat color are condi-
tioned at all by Mendelian ‘‘unit factors,’’ there must be
more than one pair of allelomorphs concerned. The
monohybrid ratio is obviously lacking, and there is no
segregation into distinguishable classes. Furthermore,
it must be borne in mind that no indirect evidence for
segregation can be pointed out in the F, arcada which
is not equally manifested ‘in the F,.

TABLE [X24
Body Tail Tail-Stri
Noot | Mm.) | Pecat) |, Pot (Mm) (Per Cent)
Animals |
i Mean | Mean | S. D. | Mean 8.D, | Mean | 8. D.

rubidus (cage-born) .-| 61 | 87.80 | 98.5 _ | 20.82 38.3
sonoriensis (cage-born)| 121 | 86.48| 78.4 19.22 24.7 ;
Pi hybrids........... 21 | 86.43| 88.0 | 4.3 | 20.09! 0.85 | 313 | 4.0
Fs hybrids oie 20 | 84.70 | 84.3 6.9 | 19.45 | 1.12 | 33.6! 6.5

~ %4This happens to be a badly shrunken skin, which perhaps does not
reveal the original color. The next darkest (shown in the photograph) is
of almost exactly the same shade as the darkest F, skin.

a

INHERITANCE IN PEROMYSCUS

Nos. 618-619]

us

, and Fa hybrids ERARA

of rel sap ate and dark di of P. m.
s (second line), and of
J).

Fie, 13. Ski
(first line), of P.
these races (third otis pat nes

300 THE AMERICAN NATURALIST [Vou.LIl

As regards the measurable characters which differen-
tiate the parent races, some light upon their behavior in
hybridization may be derived from the data in Table IX.

Here, too, one fact seems plain, despite the small num-
ber of individuals, and the modifications which these mice
share with cage-born mice in general. This is the inter-
mediate condition of the hybrids, both F, and F,, as re-
gards tail length, foot length and tail-stripe. The ear
measurements have not been introduced into these com-
parisons, since the parent races do not differ in this re-
spect.

As regards variability, it is seen that the standard de-
viations of the F, generation are all three larger than
those of the F,, and the differences seem great enough to
be of possible statistical significance. I have, however,
tabulated the frequencies of the various values for these
three characters (not reproduced here), and I find that
the probable explanation of this increase of variability in
the F, generation is an increase in the amount of abnor-
mality in the latter. This is to be inferred from the fact
that the extension in range is chiefly in the direction of
the lower values, while in two of the three cases the up-
permost figures actually fall below those of the F,. Now
the abnormal influences of captivity operate by decreas-
ing (oftentimes considerably) the values for these very
characters. As further evidence for this interpretation
is the fact that the average body size of the F, mice is
here less than that of the preceding generation.

A considerable number of back crosses have been ob-
tained between each of the hybrid combinations above
discussed and one or both of the parent races. The num-
ber of individuals in any one series is, however, small, so
that it is hardly worth while to deal with the results of
these crosses here. They afford as little evidence of
complete segregation of the subspecific color types as do
the hybrids previously considered.

(To be concluded)

25 Standard hoe for the parent races have been computed only for

the two sexes separately. They are, therefore, not included in this table.

Nine of the 30 F, animals are not here included, since the rubidus parents
of these were caught at some distance from Eureka.

INTERNAL FACTORS INFLUENCING EGG PRO-
DUCTION IN THE RHODE ISLAND RED
BREED OF DOMESTIC FOWL. III

DR. H. D. GOODALE

MASSACHUSETTS AGRICULTURAL EXPERIMENT STATION, AMHERST, Mass.

Winter Egg Production and the Genetic Constitution
of Rhode Island Reds:—Pearl, *12, found his Barred
Plymouth Rocks fell into three well-defined classes in re-
gard to winter egg production, viz., those that did not lay
at all before March 1, the zero class; those that laid less
than thirty eggs with a mean at about 16 eggs, the medi-
ocre producers; and finally those that laid over thirty
eggs, the high producers. Pearl has stated, however,
that the existence of the two classes of birds, mediocre
and high, is the important point rather than the number
of eggs at which the dividing line falls, which may be
below 30 eggs in some flocks and above 30 in others.
Since the record of No. 5080 is that of a typical true medi-
ocre producer; it is clear why the division point need not
fall at a particular number of eggs. If, however, the
numerical record of a pullet is made the only basis for a
division point, it may be pointed out that the point could
be shifted by differences in environment. If, on the
other hand, under the same general environment one
flock was found to have a different point of division from
the second flock, the latter could not be considered genet-
ically like the first. Thus far, no satisfactory division
point has been found for our Rhode Island Reds, due
probably to the few records of the type shown by No.
5080 (that is, to an absence of true mediocre producers)
and to the great variability in age at first egg, associated
with a comparatively uniform rate of production after

301

302 THE AMERICAN NATURALIST [ Vou. LIT

the appearance of the first egg. While a true mediocre
producer such as No. 5080 can easily be distinguished
from a high producer in Barred Plymouth Rocks, it is
impossible to draw a division line on the basis of the
number of eggs produced where the egg production is of
the sort observed in our Rhode Island Reds. (See Figs.
3 and 4.) Except in the case of a few individuals, the
only evidence for the existence of two genetically distinct
groups, such as Pearl found for Barred Plymouth Rocks,
is in the shape of the left-hand portion of the curves of
winter egg production. If the zero producers be omitted
as a wholly artificial group (cf. Figs. 8 and 17) this evi-
dence becomes less satisfactory, especially as the excess
of numbers on this side of the curve may well be due to
environmental factors, since the effect of such factors is
almost always in the direction of decreased production.
Moreover, in a later paragraph it is shown that the shape
of these curves depends upon certain clearly recognized
factors.

In order to eliminate any miscomprehension in regard
to the characteristics of a mediocre producer, I have gone
over the matter personally with Pearl. It appears that
his Barred Plymouth Rocks, with the exception, of course,
of the zero producers, begin to lay at about the same age,
but lay at widely differing rates. My Rhode Island Reds,
on the contrary, begin to lay at widely different ages, but
lay at a fairly uniform rate. In the Barred Plymouth
Rocks, therefore, there are three distinct types of winter
records, zero producers, mediocre producers (Fig. 12,
Nos. 274 and 284) and high producers (Fig. 3). The
same types of records have appeared in a flock of Brown
Leghorns which we have trap-nested. ,

In our Rhode Island Reds, records of the mediocre
type are so rare that they must be referred either to some
non-genetic origin, or a chance union of two hetero-
zygotes, such as results in the appearance of an occasional
recessive in a flock that characteristically has some domi-

Nos. 618-619] EGG PRODUCTION 303

nant character such as rose comb. It is quite possible
that while our flock of Rhode Island Reds as a whole is
homozygous for high fecundity, it is not entirely so. It
is possible, of course, that many of our Rhode Island
Reds are genetically true mediocre producers, but that
the addition of some other genetic factor has so altered
the rate of production that the records can no longer be
recognized as true mediocre records.

It seems clear, then, that our Rhode Island Reds fall
into the class of high producers observed in Barred
Plymouth Rocks, but that the great variability in maturity
results in a portion of the flock giving numerical results
like those of the Barred Plymouth Rocks, but which in
reality are not at all equivalent biologically. It is also
clear why Castle’s (*15, 16) recent criticism of Pearl’s
theory of egg production can not be considered to be es-
tablished, in so far as it concerns winter egg production.

Moreover, it will be desirable in the future to distin-
guish clearly between the two sorts of numerical results.
A discussion of the bearing of a division point at 30 eggs
upon the problem is given below.

It has been maintained by Pearl that winter egg pro-
duction is a satisfactory basis for selection, both for itself
and as an index of the annual egg production. This

would be true only when there is a definite relation be-
tween winter and annual production such as he has found
for Barred Plymouth Rocks. Since winter egg produc-
tion forms a part of the annual egg production with which
it is to be compared, it is evident that the coefficient of
correlation may be expected to have a positive value of
some magnitude, unless there is a definite tendency for
birds that lay well in the winter to be poor layers in the
summer. On this account it has seemed best to us to
compare winter production with that of the same hens
for the remainder of the year. The number of birds (77)
available for the comparison was not large; but the cal-
culated coefficient (.365 -+.067), while statistically sig-
nificant, being six times its probable error, is too small

304 THE AMERICAN NATURALIST [ Von. LII

for use in breeding operations. While there is an evi-
dent tendency for a high winter producer to be also a high
producer for the rest of the year, it seems also true, as
far as the 1913-14 records are concerned—those of 1915
being at this writing incompletely worked out, though of
the same general order—that there is no special tendency
for birds of late start to stop early rather than late. Of
course it is essential that a bird lay well in the winter if
she is to make a good yearly record, and in this sense the
winter egg production may be of value as a measure of
fecundity, but good winter production does not insure a
good annual production, nor does a low winter production
necessarily mean a poor annual production. It is true,
however, that birds that make the very highest records
must lay throughout the entire year. From the data in
hand it seems probable that winter egg production of
Rhode Island Reds is not as valuable a measure of the
innate fecundity capacity of a bird as it is for the Barred
Plymouth Rocks.

On the average, the flocks, if grouped according to the
month hatched, have an annual record that differs by an
amount equal to an average winter month’s production,
viz., 10 eggs, as shown by Table VIII. Or, to put the
matter a little differently, the average egg production of
pullets hatched in March, April or May is approximately
the same from February 1 to November 1. We have no
evidence that the early hatehed birds, on the average, stop
laying earlier in the fall than those hatched later.

The influence of time of hatching on a division point at
thirty eggs is very marked in the Rhode Island Reds. In
Table VI the egg production of pullets laying at all dur-
ing the winter is divided into an over-thirty and an under-
thirty class. In the former are 68.8 per cent. of the
March-hatched pullets, 58.7 per cent. of the April-hatched
pullets, while of the May-hatched birds there are only
26.6 per cent. The means of the over-thirty class are
63 eggs for the February-hatehed pullets, 54.9 for the
March, 46.3 for the April birds, while that for the May

Nos. 618-619] EGG PRODUCTION 305

TABLE VIII
Eee PRODUCTION OF PULLETS HATCHED IN APRIL, MAY AND JUNE, 1915;
SHOWING NUMBER OF EGGS LAID PRIOR TO AND AFTER MARCH 1, ALSO
PRIOR TO AND AFTER FEBRUARY 1

Months Hatched

|
|
|

March | April | May
Number of pulle o Ee eet 139 140 116
Total eggs before pragi A Nig WL cok E ee 5,780 4,470 2,253
Total eggs after March 1 to end of year........ 12,951 12,974 10,982
Av. number of eggs before March 1............ | ; 19
Av. number of eggs after March 1............. | 93.2 92.8 94.6
Grand total 20. diia EA | 18,731 | 17,444 | 13,235
Av, yearly production. .....-...+++-+++ss+++:| 134.8 124.7 114.0
Av. number of eggs for Febru pro Ar UA AER SE a 10.3 12.0 11.9
Av. number of eggs before Feb E aa Hu ae Se aos 31.3 19.9 7.5

Av. number of eggs after F February 1 1 to end of
year. 103.5 104.8 106.5

Note.—The data above the double bar is for birds that completed the
year. A part of the data below the double bar was compiled at the end of
the winter and thus includes some birds that died during the summer.

birds is 40.5. The means of the under-thirty class are
18.8 for February and for the other three months ap-
proximately 16 eggs. Now, the mean for the abstract
numbers 1 to 30 is 15.5, a value which is not far from
the observed mean, since the value of 18.8 eggs obtained
for the February birds is probably due to the small num-
ber of individuals involved. As the value of the mean
for the under-thirty birds is practically alike in all cases,
and as the value of the means for the over-thirty class
for the various months decreases with decreasing age,
it is evident that the value, 16.1, comes mainly from the
: relation of the abstract numbers and has little or no sig-
nificanceinitself. This point will be returned to shortly.
It will be noted in Table VI that the percentage of over-
thirty pullets dropped very suddenly in passing from
April to May. It means, 1 take it, that many of the May-
hatched birds did not reach a winter egg production of 30
eggs or over because of the time they were hatched. In-
deed, something of this sort is to be expected when a
definite number of eggs is taken for a dividing line at a
particular point on the calendar. If the production curve

306 THE AMERICAN NATURALIST (Vou. LII

for May-hatched pullets be examined it will be noticed
that there is a large percentage of pullets in the 16-20,
21-25, and 26-30 groups. If the average monthly pro-
duction of about 10 eggs be added to these groups, most
of them become over-thirty birds and the percentage of
over-thirty birds becomes nearly the same as that for the
April-hatched pullets. This is important, since it indi-
eates that May-hatched Rhode Island Red pullets mature
too late to furnish data comparable to Pearl’s Barred
Plymouth Rocks, for which it is stated (Pearl ’15) that
April- and May-hatched pullets alone give normal records.

Birds that begin to lay by January 1 and lay at the
rate of 15 eggs per month would lay 30 eggs before
March 1. This rate is about the lowest continuous pro-
duction that has been noted, for birds that lay at a less
rate usually lay intermittently. Pullets that lay at the
rate specified and which begin sufficiently early in the
season make very good winter records. In a recent
paper Pearl, '15b, has remarked:

Any bird laying 18 or more eggs per month in the months of Novem-
ber, December, January and February may certainly be regarded as a
high winter producer.

The statement as written might imply that this rate is
maintained throughout all four months, but, as this means
72 eggs and as the context implies, we take Pearl’s state-
ment to mean 18 eggs for any one month. At this rate a
bird beginning to lay January 1 would lay 36 eggs before
March 1. Since, however, the mode of the frequency
curve of age at first egg falls at the 251-260 days groyp,
and since the median also falls near this group, it means
that nearly one half of a flock of pullets hatched May 1
will not begin to lay till after January 15, and therefore
will lay less than 30 eggs. The slope of the curve indi-
cates, moreover, that the upper limit of the range falls at
about 311-320 days, which means that a few individuals
begin to lay so late in life that they can not be expected
to make normal records. Ten months from hatching time
brings birds hatched the last of April into laying some

Nos. 618-619] EGG PRODUCTION eo

time in February. Three hundred and ten days seems,
moreover, to mark the approximate boundary of a group
of stragglers which perhaps corresponds to Pearl’s after
March 1 group of producers, i. e., his zero producers.
This group is represented in the otito by the shoulder
which begins at this point, there being of course an over-
lap with the larger group, beginning somewhere about
280-290 days.

The relation of the abstract numbers involved in the
series of data relating to winter egg production is a mat-
ter of some importance. The mean of the abstract num-
bers from 1-30 is 15.5, or from 0-30, 15, a value that cor-
responds closely to the mean value of the number of eggs
laid by the under-thirty class. The mean value of the
numbers above 30 beginning at 31 and proceeding to some
other higher number such as 50 or 80, will depend in part
on the value chosen for the higher number. If 80 be
chosen as the higher limiting value, then the mean of the
numbers 31-80 is found to have a value of 56.2. The
mean values just given hold only when the abstract num-
bers are taken one at a time or when they are arranged
in a symmetrical fashion about the mean, as, for example,
in the ratio of 1:2:3:4, etc., and back to 4:3:2:1. If
the 1-30 winter egg production group represented a defi-
nite genotype one would expect a symmetrical or nearly
symmetrical distribution of the concrete numbers about
their mean. If, however, the numbers (abstract or con-
crete, as the case may be) had some other sort of arrange-
ment—as, for example, if they formed part of a normal
curve of variation—a different set of mean values would
be obtained, depending upon the steepness of the slope
of the curve. If, for example, the classes 1-5, ete., to 25-
30 inclusive, are formed and the central values appear in
the ratio of 1:1.3:1.7:2:2.3:2.7, the mean will be 18.2.
On the other hand, the mean of the abstract numbers
above 30 that form the remainder of the normal curve,
under some circumstances may shift downwards, as, for
example, when the mode of the curve is at 50 with the

308 THE AMERICAN NATURALIST [ Vou. LIT

upper limit of the range at 90. If the observed distribu-
tion of winter egg production is mono-modal, and if its
curve, base, mode and range correspond approximately
to those of the curve assumed in the preceding para-
graph, the mean of the numbers 1-30 will be larger than
the mean (15.5) obtained from the abstract numbers in the
manner chosen at the beginning of the paragraph. That
is, there is a tendency for the means of the two parts of
the mono-modal curve, i. e., the ones derived from 1-30
and 30-90, respectively, to approach each other in value.
If, however, the curve of winter egg production is a com-
posite of two curves, i. e., mediocre and high producers,
then the means tend to approach those of the abstract
numbers involved when distributed symmetrically.

Some doubt exists as to how our data are to be inter-
preted in the light of the statements in the preceding para-
graph. The mean of the under-thirty group—viz., 16—
while slightly higher than that of the mean for the ab-
stract numbers—viz., 15.5—probably is not significant,
though it may perhaps be taken as an indication of the
existence of two genotypes. Since, however, the mean
of the over-thirty group increases with increasing age
and since the percentage of birds falling in the over-
thirty classes increases with increasing age, it seems
probable that the shape of the observed egg production
curves is due to the artificial division point at March 1,
while the irregularities in the curves are due to too few
numbers. Since these irregularities largely disappear
if the class intervals are doubled and since, if the time
limits were to be extended so as to permit all birds to
begin laying, it appears probable that the curve would
become a symmetrical unimodal curve. Moreover, this
conclusion is strengthened by the fact that the observed
mean, 46.3, of the over-thirty class of the April-hatched
pullets with its upper range at 77 is lower than the value
(56.2) noted for the abstract numbers 31-80. For Barred
Plymouth Rocks, however, the means do not rest upon
the relation between the abstract numbers, but are an
expression of the average production of the two types.

Nos. 618-619] EGG PRODUCTION 309

A comparison of the winter egg production curves ob-
tained with the Rhode Island Reds with the similar curves
for various seasons for the Barred Plymouth Rocks

us
30

25

8

à

3

y 20

€

w

=

Sus

E

5

a

10

si

pm EE oe
el ms aee a AS SAS RE Ss 1045
w us m5 RI OS 408 BE AT NS 855
WINTER EGG PAODVCTION

Fig, 14. omparison of a flock of Barred Plymouth ta pullets (Maine,
1907-08, solid line) Pub wo flocks of Rhode Island Red pullets (Mass. i
broken line; Mass., 1913-14, dotted mee): to show the difference in the sh pe of the
winter egg prod sa curves. or the present purpose it seemed unnecessary
to regroup our sapi to correspond exactly to the Maine Station grouping.

brings out some interesting points. These curves are

shown in Figs. 14, 15, 16 and 17. In Fig. 14 the con-

tinuous line is the frequency curve for the winter egg pro-
Da

ho a do

PERCENT

de

0
0-4 59 1014
ESOS

Fig. 15. Comparison of a flock of Barred Plymouth ma pullets (Maine,
1912-13, dotted line) with a flock of Rhode Island Red pullets (Mass., 1913-14,
solid line) to show the similarity in the distribution of prt egg production.

duction at the Maine Station for 1907-08, the flock reared
by Pearl from the birds previously bred by Gowell for
increased egg production. The line composed of short

310 THE AMERICAN NATURALIST [ Vou. LIT

dashes is the curve of egg production for the winter of
1913-14, made by our Rhode Island Reds hatched in
March and April. This flock is the first flock about which
we have full data. They were the daughters of the orig-
inal flock, 1912-13, as described above, and for all essen-

e oR

PERCENT

ne

1-3 6-1011 61-88
eee?
. 16. Winter egg production of the Rhode Island Red pullets cma
in donk 1913. The records of these pullets have been separated from the
records shown in Fig. 14 (solid line).

tial purposes were a random selection of individuals.
This flock, one generation removed from its standard-bred
ancestry, clearly differs from the 1907-08 flock of Pearl’s
in containing many more fairly high birds. The data for
the 1913-14 flock, grouped in classes of 1-5 instead of 1-10,

PERCENT

OHhGAGAvHeSORERSRTESS

-9 30-19 20-29. 3JO-IY 40-49 50-59 60-69 70-79 80-89 90-99

Fic. 17. Winter egg production of the ‘flock of pullets reared in pare from

eggs purchased of the breeder furnishing our foundation stock. The solid line
represents the March hatched spect dell the broken line, the March Slds plus
April hatched pe from the same source,

is shown by the continuous line in Fig. 15, where it is to be
compared with the winter record of 199 Barred Plymouth
4 Rocks made at the Maine Station for 1912-13.4 The es-
sential similarity of the two curves is self-evident. As
ê The data for the Barred Rocks was taken from Pearl (15c).

Nos. 618-619] EGG PRODUCTION Sit

_ Pearl’s pullets were (presumably) hatched in April and
May, while ours were hatched in March and April, the
curve for the April, 1913, hatched pullets alone is given
in Fig. 16, which shows that the separation of the March
pullets from the April pullets has not essentially altered
the shape of the curve. The zero producers were omitted
in order to study the shape of the curve when treating
the zeros as an artificial group. This omission, however,
would not alter the essential shape of the curve.

As already stated, new blood was added to our breed-
ing pens in 1915. The egg production curve made by the
pullets hatched in March and April for 1915-16 is shown
by the broken line in Fig. 14. Compare also the separate
curves for March- and April-hatched pullets shown in
Fig. 8. Although some individual records better than
any previously obtained were secured, on the whole egg
production was not as good as for 1913-14, although still
distinctly better than the production of the 1907-08 flock
of Barred Plymouth Rocks.

The figures when grouped according to the month in
which the birds were hatched show clearly the part played
by this factor in determining the sort of records made.
The curve for the March-hatched pullets is quite similar
- in appearance to the curve made by Pearl’s Barred
Rocks in 1913-14, except that the mode falls at a lower
number of eggs. That is, the Barred Rocks contained a
higher percentage of birds laying above 50 eggs than the
March-hatched Rhode Island Reds. The curve, however,
for the May-hatched Rhode Island Reds is much more
like that of the 1907-08 curve of Barred Plymouth Rocks,
while the curve for the April-hatched pullets is inter-
mediate between the two.

It seemed a matter of some interest to determine
whether our 1913-14 record was a chance record or
whether it was a fair sample of the original stock. Con-
sequently eggs were purchased from the original source
and chicks reared in 1915. The first lot was hatched
March 22 and all the pullets (39) that passed the test for

312 THE AMERICAN NATURALIST [VoL. LIT

vigor were put into a laying house in October. The egg- .
production curve is shown by the continuous line in Fig.
17. It bears a strong resemblance to that of the Maine
Station for 1912-13 and that of our flock for 1913-14.
There was in addition to the 39 pullets a smaller lot
hatched April 27. The curve for both lots eombined is
shown by the dotted line in Fig. 17. Again the curve is
essentially like the 1913-14 curve for Rhode Island Reds.
The writer would be glad to study the original stock in
more adequate fashion, but the amount of time, labor
and equipment required are so great that the end in view
did not seem to warrant the expense.

Since Pearl has not discussed the question of age at
first egg from the present standpoint, the possibility must
not be overlooked that the factor of maturity may play an
important part in his results. In a recent paper (Pearl,
16) he states that in a good laying strain the pullets
mature at five or six months on the average. Quite a dif-
ferent state of affairs exists in our Rhode Island Reds,
where the average age is much higher, about 83 months
for the entire flock. It is clear, however, from the discus-
sion that the factor of maturity does not influence his
results to any considerable extent.

Since the two principal internal factors responsible in
most cases for the number of eggs produced by a pullet
during the winter are, first, the date of the first egg, and
second, rate, similar sets of data may result from vari-
ability in either of the two factors. Now the date of the
first egg is dependent in part on the time of year the bird
is hatched, and in part on differences in maturity. The
former factor is under control and can be used to elimi-
nate differences in maturity. Rate has already been dis-
cussed in another place, but it should be noted that it
may be modified by such definite factors as broodiness,
or the presence or absence of cycles, as well as other
_ physiological factors that appear to be innate but which
can not be named at present. The combined effect of
rate and date of first egg, or better the length of time

Nos. 618-619] EGG PRODUCTION gis

elapsing between the first egg and March 1, is such that
the same number of eggs may result from a variation in
either one of these factors. If a flock of birds all begin
to lay about the same time of year, but vary greatly in the
rate at which they lay, a body of data superficially the
same as that produced by a flock of birds laying at a
fairly uniform rate but varying greatly in the time at
which they begin to lay might readily result, especially if
the variability of each set of factors (rate or maturity)
was very much alike. Radically different sets of data
would result only if the variability belonged to quite dif-
ferent types. The birds that make unusually high winter
records are those that start early and lay at a good rate
throughout the winter. A poor record, on the other hand,
results either from a low rate combined with a long period
of production (the date of first egg coming early in the
fall) or from high rate and a short period of production,
since the date of first egg comes late in the winter.

A late date of beginning egg production may be the
result of late hatching or of delayed maturity. That is,
an early-maturing bird if hatched too late may lay no
more eggs than a late-maturing bird hatched early in the
spring. Thus, the actual record of a bird is the result of
the influence of several factors, themselves very variable,
and gives much the same result as though a much larger
number of factors were involved.

Pearl has spoken of two production factors, L, and Lo,
but has not, so far as I know, assigned either to the two
chief factors concerned in winter egg production, viz.,
rate and age at first egg, nor does the context indicate any
such sense. At first sight it might seem possible that one
of these factors is a factor for early maturity, the other
for rate; for according to Pearl’s theory both factors
must be present to secure high production, since the L,
factor in the homozygous condition does not make a high
producer. So far as can be seen, there is no objection
to assigning one of Pearl’s two factors to rate. Some diffi-
culties are encountered, however, in assigning maturity

314 THE AMERICAN NATURALIST [ Vou. LII

to either of Pearl’s factors, for early-maturing birds
almost invariably lay more than the required number of
eggs even when their production is interrupted in some
way. Since, however, the theory demands that both
genetic factors be present for high production, the assign-
ment of one factor to rate and the other to maturity can
not be made.

THE SHAPE or WinteR Ece Propuction Curves

The curves of winter egg production are clearly com-
pound curves, probably belonging to the ‘‘S’’ type de-
scribed by Pearl and Surface for monthly egg produc-
tion in Barred Plymouth Rocks.

In the case of the winter egg production curves, it can
be shown that this type of curve is due primarily to the
variability in age at first egg, plus the date at which the
census of the flock is taken. The curve, however, is
modified somewhat by the variability in rate and by the
fact that the birds were not all hatched at the same time.
If a flock of birds were all hatched on the same day and
all laid at some uniform rate, say an egg per day, it is
clear that, on any given date between the date the first
pullet commenced to lay and the date the last one began,
the frequency polygon showing the number of eggs pro-
duced would consist of a zero class plus the remaining
portion of the polygon, beginning at class 1 and proceed-
ing through classes 2, 3, 4, ete., to the upper end of the
range. Now the number of birds laying one egg apiece
would depend upon the date chosen for the census and
would be the number of birds that reached a given age on
that date as shown by the curve of age at first egg. Thus,
if Fig. 1 be taken as our standard, and some date early
in the season be chosen for the census, say the date on
which the flock becomes 186 days of age, the zero com-
ponent of the egg curve would have a value of 99 per
cent. and the 1-egg class a value of .6 per cent. At mid-
season (256 days) the zero component would have an
ordinal value of 55.3 per cent. and the l-egg class a value

Nos. 618-619] EGG PRODUCTION 315

of 14.3 per cent., the 11-egg class a value of 9 per cent.,
the 21-egg class a value of 11.4 and so on. Towards the
end of the season, say at 306 days, the value of the zero
class would be 13.3 per cent., and the 1-egg class would
have a value of 3.1 per cent., the 11-egg*class a value of
5.1 per cent., the 51-egg class a value of 14.3 and so on.
At the close of the season, 366 days, the zero class disap-
pears, while each egg class has the percentage values
given in Fig. 1, beginning at the extreme right and pro-
ceeding to the left, 7. e., the egg curve is a mirror image
of the age at first egg curve.

The first modifying factor, i. e., rate, tends to flatten
the theoretical egg curve by shifting individuals from
one side toward the other. Thus, an individual fall-
ing in the 56 class, on the 100 per cent. rate, would fali
into the 36 class at a 66% rate. If a 50 per cent. rate
were selected as the theoretical rate, the 663 rate hen
- would be shoved over into the 76 class. If the date of the
first egg of a sufficiently large number of pullets all
hatched on the same day were plotted on a suitable cal-
endar, a duplicate of the age at first egg curve would be
obtained; but if several hatches are grouped together,
e. g., the four hatches occurring in any one month, it is
evident that the curve plotted on the calendar would be
considerably flattened and in turn would flatten out the
theoretical curve of egg production as based on age at
first egg. .

Selection.—Pearl's success in securing increased egg
production by breeding might be due to his methods of
selecting the breeders, regardless of all theoretical con-
siderations. Families that contained all high producers
were selected generation after generation to propagate
the high fecundity lines. Families in which true mediocre
producers appeared, i. e., where segregation took place,
were not used in breeding for increased egg production.
This type of selection could hardly fail to yield results,
provided that egg production is inherited. Nevertheless,
it is clear that fecundity is inherited in Mendelian fashion

316 THE AMERICAN NATURALIST [ Von. LII

in Pearl’s Barred Plymouth Rocks. However, the re-
sults obtained by Dryden at the Oregon Station show
that individual selection in pedigreed lines as opposed to
mass selection may result in improved egg production
quite as well as by the application of Pearl’s theory.

Egg production in the domestic fowl may seem at first
sight to be a highly desirable character on which to
study the influence of selection. It may be regarded as a
unit character if one so desires, and if, by selection, this
character is changed, it is clear that selection has been
effective. But it is also clear that the effectiveness of
such selection in this instance rests, in large measure, at
least, upon the influence exerted by various modifying
factors, such as broodiness or age at first egg, discussed
in this paper. It is possible to study these factors indi-
vidually both by themselves and also in their relation to
egg production. Broodiness is known to behave like a
Mendelian dominant, while Pearl has shown that the rate
of production during the winter cycle is dependent on two
genes, one sex-linked, the other a simple Mendelian char-
acter. We have found some evidence that the presence
or absence of a winter cycle in fowls that lay at all during
the winter? follows the Mendelian scheme. Since the
influence of the various modifying factors is so clear cut,
it is evident that egg production is a character wholly
unsuited for studying the possibility of the modification
of the germinal representatives of a character by selec-
tion. On the other hand, it is a good example of a char-
acter that varies continuously, but the continuity of whose
variability can be shown to depend upon several modify-
ing factors.

There is a point of some general interest regarding the
genetic composition of any given flock. Hardy (’08)
showed that the proportions in which a Mendelian char-
acter occurs tend to remain constant provided no selec-
tion is practised. Fanciers often practise a certain but

7 The absence of a winter cycle in this instance means continuous produc-

tion throughout the winter and spring, and not absence of egg production
as in Pearl’s Barred Plymouth Rocks.

Nos. 618-619] EGG PRODUCTION 317

indefinite amount of inbreeding. Under such circum-
stances there would be a tendency for the fecundity
factors to remain in about the proportions in which they
started. We may, therefore, expect to find ready-made
flocks of high producers, true mediocre producers, or
even zero producers as well as those containing the sev-
eral types. Thus, the original Barred Plymouth Rocks
of the Maine Station contained all three types, while the
Cornish contain only true mediocre and zero producers.

In spite of the fact that we have as yet been unable to
apply Pearl’s theory of egg production bodily to our
Rhode Island Reds (although it may yet be possible to
use it after making some modifications) there is no ques-
tion but that the ability to lay is inherited, as shown by a
better egg production in some families than in others. It
is clear also that some males produce offspring that on
the whole make much better records than those from
other males when the two groups of females with which
they are mated are very similar in their winter egg pro-
duction. In one instance, the difference between two sets
of offspring by two males was clearly due to a difference
in maturity. It seems clear, moreover, that some of the
internal factors, such as broodiness and maturity, segre-
= gate independently.

SUMMARY

1. The object of the present paper is to present a sur
vey of the problem of egg production based on the re-
sults of four years” study of egg production in Rhode
Island Reds. The presentation of the data is incidental
to this object.

2. On account of disturbing factors, data from two
years only is presented.

3. There are two main conclusions: First: egg pro-
duction in our strain of Rhode Island Reds differs in
several important respects from Pearl’s strain of Barred
Plymouth Rocks and also from Leghorns. Second: the
egg record of a hen by itself is an unsafe basis on which

318 THE AMERICAN NATURALIST [ Vou. LII

to breed for definite degrees of production, for it can be
shown that egg production depends on several more or
less independent internal factors and that the same num-
` ber of eggs may result from the action of different sets
of factors. It follows, therefore, that each factor must
be studied separately, both from the physiological and
genetic standpoints.

4. The factors reviewed are: date of first egg, age at
first egg, growth, rhythm and rate of production, includ-
ing here Pearl’s genetic factors L, and L,, broodiness,
moult, cycles, persistence of production in the autumn,
and stamina.

5. Date of first egg is shown to depend on the time of
hatching, the rate of growth of the young birds and some
elements, at present unknown, that determine the attain-
ment of sexual maturity. On the average it has been
found that those individuals that lay early in the fall
(October) lay more winter eggs than those that begin to
lay later.

6. On the average, pullets that lay relatively early (6
to 7 months) in life lay more eggs than those that lay at
a later period (8, 9, or more months) in life, other things
being equal. The variability in age at first egg appears
to be much greater for our Rhode Island Reds than for
Pearl’s Barred Plymouth Rocks.

7. Birds that lay rapidly, other things being equal, lay
more eggs than those that lay more slowly.

8. Some birds (true mediocre producers) lay very
slowly and irregularly, producing only a few eggs (1-10
or thereabouts) per month. Others (true high producers)
lay much better (15-28 eggs per month).

9. The effect of the age at which the first egg is pro-
duced on winter egg production is such that numerical
results, similar to those given by true mediocre pro-
ducers, may result.

10. Some pullets lay continuously or nearly so, for
long periods of time. Others lay relatively rapidly, but
lay in eyeles with a period of rest in between. These rest

Nos. 618-619] EGG PRODUCTION 319

periods may or may not be associated with broodiness.
In a large percentage of Rhode Island Reds, a winter
cycle comparable to that found in Barred Plymouth
Rocks, is absent.

11. Broodiness operates to reduce egg production very
materially, for the average production is about 40 per
cent. higher for the period prior to the time each indi-
vidual goes broody compared with average production
after that time. The apparent paradox that hens with
the greatest number of broody periods lay more eggs
than those with fewer broody periods is due to the fact
that increased production affords opportunity for more
broody periods. The presence of a large amount of
broodiness in our Rhode Island Reds differentiates them
from the Leghorns, which lack this characteristic.

12. The appearance of a moult often stops production.
A partial summer moult was noted.

13. Other things being equal, birds that lay late in the
fall lay more than those that stop early.

14. Small birds mature earlier, on the average, than
large ones and hence lay more winter eggs.

15. It was observed that while birds of poor stamina
might make exceptionally good records, that lack of
stamina tended to delay the appearance of the first egg
and hence lowered the winter records.

16. It is pointed out that while the results obtained in
Rhode Island Reds differ from those obtained by Pearl
in Barred Plymouth Rocks in egg production, this dif-
ference must be looked upon as a real difference just as
the two races differ in color.

17. Egg production is an unsatisfactory character on
which to study the possible effects of selection in modify-
ing the germ plasm, because in reality it is complex and
not a simple unit character.

18. The genetic constitution of our Rhode Island Reds
in respect to Pearl’s L, and L, factors has not certainly
been made out, but it seems probable that as a strain they
fall into Pearl’s class of high producers. True mediocre
producers are very uncommon in this strain.

320 THE AMERICAN NATURALIST [ Von. LIT

19. The relation between the means of the abstract
numbers, in the series 1-30 and 30 to some higher num-
ber, and its bearing on the use of the means of egg pro-
duction of two groups falling within the same limits is
discussed.

20. The curves of winter egg production are shown to
be compound curves.

21. A knowledge of the factors described is of im-
portance both from the commercial and biological stand-
points. As Pearl and Surface (’08) pointed out a number
of years ago, the income received from each bird will
depend not only on the number of eggs produced, but also
on the season at which those eggs are laid. A bird that
produces 100 eggs at suitable seasons may yield as much
income as a bird that produces 200 eggs at less profitable
seasons.

From the biological standpoint, a knowledge of the
separate factors is important because what might seem at
first sight to be a simple character is really extremely
complex. Obviously, then, it is necessary to attack the
problem from this standpoint.

LITERATURE CITED
Castle, W. E.
1915. Some Experiments in Mass Selection. Am. Nar., Vol. XLIX.
1916. Can erin Cause Genetic Change? Am. Nar., Vol. XLX.
Curtis, M, R.
1914. A Piomar Study, ete. IV. Factors Influencing the Size,
Shape, and Physical Constitution of Eggs. Arch. Ent. Org.,
Bd. XXXIX.

Dryden, J. A
1916. Poultry Breeding and Management. Springfield.
Goodale, H.
1918. Winter Cycle of Egg Production in the Rhode Island ao
of Domestic Fowl. Journ. Agri. Research, Vol.
Gowell, G. M.
1902. reading for Egg Production. Maine Agri. rae Sta. Bul. 79.
1903. Breeding for Egg Production. The same
1905. Poultry Experiments. The same, Bul, 117.
1906 oO Experiments. The same, Bul. 130,

1908. Iodoa Proportions in a Mixed Population. Science, Vol. 28.
Pearl, R.

Nos. 618-619] EGG PRODUCTION 321
+

1912. The ng of Inheritance of Fecundity in the Domestic Fowl.

Jour. Exp. Zool., Vol. 13; also Maine Agr. Exp. Sta. Bul. 205.

1915a. Seventeen Years Selection of a oe T Sex-linked
ende

lian Inheritance. AM. AT., Vol.
1915b. Mendelian at ot ARA in ine Dada Fowl and
Average Flock Product AM Vol. 49.
1915c. pa of the Winter Cycle in de egt Production of
estic Fowl. Jour. Agri. Research, Vol

1916. The Effect of Feeding Pituitary Substance ait Corps Luteum

Substance on Egg Production and Growth. XIV
Pearl and Surface.

1908. Poultry Notes. Maine Agri, Exp. Sta. Bul.

1909. A Biometrical Study of Egg Production in pa ‘Domestic Fowl.
I. Variation in Annual snag Production. U. S. Dept. Agri.
Bureau Am. Ind. Bul. 1

TOTE Ek dador Distribution p? Egg Production. The same,
Part II. y

Rice, J. E.
1913. Some Practical poe in the Management of Poultry for Egg
roduction in ultry Culture. Bull. 1. Issued by State
Board of siise Boston.
2

AN EXAMINATION OF THE POLICY OF RE-
STOCKING THE INLAND WATERS WITH FISH!

PROFESSOR W. M. SMALLWOOD

SYRACUSE UNIVERSITY

Tue large sums of money annually expended by both
the National Government and the several states in fish
propagation fall into two general fields of activity,. the
marine and the freshwater. The freshwater activity in
turn may for convenience be divided into the production
of food and game fish.

It is always proper to examine the conditions which
influence restocking; and just at this time it is especially
fitting to enquire into the efficacy of the methods. The
technique involved in securing the eggs and their care
during hatching have been well worked out. It was a
marked step in advance when these modern methods were
first put into practice. The money used in carrying*out
modern methods in the many fish hatcheries is efficiently
expended so far as the writer has been able to determine.
The fundamental scientific problems involved have been
solved so that the regular fish foreman can successfully
direct and supervise all of the steps in the process.

After the eggs have been hatched and the young fed for
a certain length of time, they are distributed to the ponds
and streams. The last act in the series is the one con-
cerning which we know the least. In order to gain an
insight into the actual conditions, a typical Adirondack
pond was selected for study.

_ The whitefish is the only species that has become at all
abundant as a result of the policy of the state. The fault
does not seem to be connected with the number of finger-

- lings placed in this lake, for the state has, indeed, been

1 Contributions from ‘the Zoological Laboratory, Liberal Arts College,

Syracuse University, C. W. Hargitt, director.

322

Nos. 618-619] RESTOCKING INLAND WATERS 323

generous. The problem that confronted the writer was
to discover the cause or causes for the obvious failure of
this lake to support an abundance of fish after these thirty
years of restocking.

In the study of the life of any given body of water or
area of land, a number of fundamental relationships have

T Spawning-Bed Point
& Tamarack Swamp Brook
3 Outlet

la Outlet Flow
er, ”

Seale L 2450" Lnlargrd From USGS. Sheet by HBWaha
Fie. 1.

to be established before an effective detailed study can
be made. After examining the conditions in several
Adirondack ponds and lakes, the writer felt the necessity
of examining the broader aspects of the problem with
the hope that such a study might reveal some of the

324 ` THE AMERICAN NATURALIST [ Vou. LII

causes that are influencing the present general food
supply for fishes, the future food supply, the plant growth
and other general problems related to the successful re-
stocking of waters in the Adirondacks.

Such a study implies that one understand the soil and
its origin. The Adirondack ponds are noteworthy for
their abundance of sand. ‘This sand is in the final anal-
ysis responsible for much of the modern life of these same
ponds and lakes to-day. The geological history of the
Adirondacks, especially its glaciology, is just becoming
well understood by the experts. For the purposes of this
paper it is simply necessary to keep in mind the fact that
as the glacier receded, the Adirondacks were surrounded
by a ring of ice. Within this ring of ice first the higher
peaks and later the lower areas were exposed. The gen-
eral result was that there were formed in succession a
series of temporary glacial lakes, the remnants of which
constitute the present Adirondack lakes and ponds.

Lake Clear, formerly known as Big Clear Pond, is lo-
cated near Lake Clear R. R. Junction. It is at the head-
water of the Saranac Lake system, so is free from the
usual migration that takes place when one pond receives
an outlet from another. The lake contains nearly 1,000
acres of water, which comes entirely from springs and
mountain brooks. The water is clear, cool and pure—an
ideal freshwater pond for restocking, one would say.
What has thirty years of restocking by the state accom-
plished? The following table indicates that this pond has
received 17,535,850 food and game fish.

In some regions, no less than eight successive lakes
have been revealed by the recent critical studies of glaci-
ologists. The net result is the accumulation of vast quan-
tities of sand from which most of the organic food has
been removed.

Lake Clear is one of the remnants of a much larger
glacial lake, the shores of which are easily made out.
This lake it has been proposed to call Lake St. Germain.?

2 From the unpublished account of Lake Clear by Mr. Harold Alling.

Nos. 618-619] RESTOCKING INLAND WATERS 325

FISH PLANTED IN LAKE CLEAR

Brook Trout Lake Trout Rainbow Trout | Brown Trout | White Fish
1887 20,000 30,000
yeaa tii 20,000 50,000
ic eel 4 A 150,000 5,000
1890 .....| 25,000 100,000 1,000
1891 .| 25,000
1892 20,000 100,000
1893 20,000 100,000
1894 15,000 15,000
1895 35,000 ,000
1 35,000 50,000
1897 11,000 10,000 100,000
1898 1,600 10,000
1899 5,000 10,000
1 36,000 35,000 3,000 600,000
1901 25,000 | 250,000
1902 750 25,000 5,000 800,000
,500 10,000 = 3,000 1,000,000
1904 18,000 8,000 1,000,
1905. 15,000 18,000 1,000,000
50,000 ,000,000
1907 5,000 50,000 760,000
1 138,000 800,000
=e 14,000 14,000 1,000,000
1910". 358 20,000 178,000
1011 ic: 35,000 1,000,000
1912 10,000 1,500,000
1012. 25,000 15,000 700,000
mu. 55,000 | 25,000 | 1,000,000
1915. .....). 3000 30,000 1,500,
ine 30,000 | 65,000 | 1,500,000
| 683,850 - | 1,067,000 | 91,000 6,000 15,688,000
UPPER SARANAC, N. Y., (Signed) Mio F. Otis

September 30, 1916

Its total area was possibly twenty times the present size
of Lake Clear and included the present St. Regis lakes as
well as several others. Preceding this fossil lake there
was a still much larger lake more than twenty-five miles
long. The station at Lake Clear and all of the level area
extending north to Gabriel’s Station is a small part of
the floor of this large lake. |
These conditions as outlined for Lake Clear in a gen-
eral way apply to nearly all of the Adirondack lakes and
ponds. Using Lake Clear as a center, there are in a
circle, the radius of which is fifteen miles, more than
seventy-five similar ponds and lakes, many of which are
restocked by the state, so that the following study may `

326 THE AMERICAN NATURALIST [ Vou. LII

be taken as describing typical conditions in the large area
of the Adirondacks.

Inasmuch as the past as expressed in the present
physiography has played such a large part in influencing
the present life of the lake, a brief description of the pres-
ent conditions is necessary as well as the special stations
at which collections were made. The names of these sta-

Fic. 2a, The shore beige stations I and II, The beach is modified
every spring by the ice. Note how free from vegetation this strip of the cee
is clear up to the boathouse.

tions are found in Fig. 1. To these should be added the
names of the two small bays, one centering at station 5,
which we will name Big St. Germain Bay, and the one
south of this, which we call Little St. Germain Bay.

The plant life, the ultimate source of fish food, is lim-
ited to the area between the shore and the 15-foot con-
tour line except for the floating algal forms. This is the
part of the lake, then, that is important for our study.

Nos. 618-619] RESTOCKING INLAND WATERS 327

From just west of station 2 to half-way between sta-
tions 4 and 5, the shoal is composed of rocks and sand.
The rocks are from the glacial till and similar to the soil
conditions in the ““fossil”? shore of Lake St. Germain as
exposed by the road east of the hotel. In front of Lake
Clear Inn at station 3, the glacial till and sand have been
washed away, leaving a small exposure of anorthosite
rock. Around stations 6 and 11 large rocks and glacial
till are common. The character of the soil in these three
areas determines the spawning habits of at least threg

A See

Fie. 20.

The shore opposite station XII.

species of fish*in the lake. The remaining part of the
shoal around the lake is wholly sand derived from the
anorthosite rock. Station 7 is a sand point, designated
as Spawning-bed Point on the oldest maps, although the
frost fish are the only ones that are known to spawn there
now.

The high water of early spring and the strong south-
west winds cause the sand to be removed from under the
trees and shrubs on the north and east shores. As the
ice breaks up in May, a strong southwest wind frequently
forces a large amount of ice on to the northeast shore and
on the north side on to the road. On the east half of the

328 THE AMERICAN NATURALIST [Vou. LII

north shore these ice-push shores form nearly every
spring, to be later washed into the lake by the heavy rains
of summer. These constantly shifting shores prevent
permanent vegetation, thus tending to give a barren ap-
pearance to much of the shore lines, Fig. 2a.

The easily modified shore extending for the most part
around the six miles of shore line may be taken as a good
indication of the lake bottom adjacent to it. The wave
action is constantly forming sand ripples which are as
constantly being changed by a heavy rain or a different

Fic, 3. Station 1, Mouth of Trout Brook, During @most of the summer
except after a heavy shower, the water in this stream at this place is not over
six inches deep.

direction of the wind. This makes it difficult for plants
to gain a foothold even if the sand were good soil for
them to grow in. The general result is, then, that most
of the shallow water is free from any but a very limited
plant growth.

The largest brook, station 1, Fig. 3, receives several
tributaries from Big Clear mountain and flows into the
lake the year round. The remaining brooks, especially
Sucker, Meadow and Tamarack, frequently become en-

Nos. 618-619] RESTOCKING INLAND WATERS 329 .

tirely dry during August. But after a severe rainstorm
all of the brooks carry a large amount of water, often
more than double their normal flow, for from twelve
to eighteen hours, when they return to the normal again.
Thus again the large amount of sand in the soil in this
region plays an important part in determining how long
the water shall be retained before it runs off. After
these rains the water in the lake around the mouth of the
brooks is colored dark by the organic matter brought
down by the water of these streams. Two important re-
sults follow: First, as this organic matter settles to the
bottom, a richer soil for plant growth is furnished; and,
secondly, fish tend to come to these places for their food.
The amount of water flowing from Trout Brook and the
frequent strong winds constantly shifting the sands pre-
vent plants from becoming established at this place. But

- the water is so dark here and so much cooler that this is

by all means the best place to 2... for brook trout, espe-
cially by fly-casting.

The summer food of fishes has been studied by so many
investigators that the main facts for the several species
are fairly well understood. But the more difficult prob-
lem of determining what the available food is during the
winter and what fish eat during this period is still prae-
tically unknown. One naturally thinks that all aquatic
life, like the deciduous trees, perhaps, enters into a rest-
ing state for several months, but that this assumption
is far from correct can be shown by the following obser-
vations.

Through the courtesy of Milo Otis, superintendent of
the Saranac Inn Hatchery, I have had sent to the zoology
department a large number of the so-called red hydra,
each month beginning with November and ending with
April. These red hydra come into the hatchery tanks
through the intake pipes in Little Lake Clear. These
pipes are from 30 to 40 feet below the surface of the water.
The important fact is clearly established that this very
simple organism, sensitive to temperature changes, lives

380 THE AMERICAN NATURALIST [ Vow. LIT

throughout the year and actively forms buds in January,
February and March in the Adirondack waters. Micro-
scopic sections of these hydra taken in February reveal
the presence of minute Entomostracans in the enteron.
Some of these minute Crustaceans taken from a jar con-
taining the hydra were submitted to Dr. C. D. Marsh for
identification. He reported that they were Cyclops
americana. As cyclops is the common food of hydra we
may assume that this species, which is very abundant, is.
eaten by these red hydra. At any rate, hydra feed on
minute animals so that animal food is a prerequisite for
their active growth.

The food of the cyclops in turn is the minute floating
algal plants. These must be in relative abundance in
order to support so many of the minute Entomostracans.
Thus the conclusive proof of the active, reproductive
habit of these red hydra throughout the winter estab- `
lishes the winter active life of cyclops and unicellular alge
in Little Lake Clear.

These observations indicate a greater amount of vital
activity in such cold regions as the Adirondacks than we
had been led to believe existed. If these minute and
simple forms of life live throughout the winter in an
active state, we may safely predict that most of the other
forms of life except the larger plants are also active and
that the winter food of fishes is probably similar to that
of the summer in many particulars.

During the summer of 1916, I had opportunity to ob-
serve the habits of the red hydra which are common in
Lake Clear, particularly at station 12. A number of col-
lections were made early in the summer and I attempted
to bring some of the live animals back to the university,
but in each instance the hydra died before reaching the
city. I then tried to acclimate them to aquarium life,
placing them in regular aquaria jars in the boathouse, but
- in each case the hydra died in from twenty-four to forty-
eight hours after being taken from the lake. As they
were kept in the lake water, the only explanation that

Nos. 618-619] RESTOCKING INLAND WATERS 331

seems applicable is that the water became too warm for
them. During these same days, hydra were coming into
the hatchery troughs at Saranac Inn and reproducing in
such abundance that it was necessary to clean out the
troughs every three days.

In these hatchery troughs, they grow so thick that a
perfect mat is formed, covering the bottom and sides in
patches two or three feet long. In the center of these
patches, the hydras become brown, and, if left for about a
week, may become nearly white. They look as if they
were dead, but when touched, contract. Hydra taken in
late September in Lake Clear were all brown, with a few
that were nearly white. This wide range in color is ap-
parently due to metabolic changes taking place in the
chloroplastic corpuscles.

The red hydra are a source of food for fish, for the trout
in the hatchery troughs eat them after they become a few
weeks old. On first hatching, the small trout are killed
by these hydra. After trout fry have eaten freely of
red hydra, their droppings are colored red, indicating
that the chloroplasts are not broken up in digestion.
Doubtless the young fish in the lake and small minnows
that secure their food from the stems of plants eat many
of these hydra.

A further question arises in the relation of the fish to
the several physiographic conditions in the lake. This is
determined by classifying the fish habitats. These are
stream, barren sandy-shoal, barren stony-shoal, vegeta-
tive and deep-water.

There is no vegetation growing in the brooks except
inside the bridge of Trout Brook where Potamogeton
robbinsii and some filamentous alge are found. At the
` mouth of Meadow Brook a few bullrushes and pond lilies
are seen; while the Divide Brook meanders over organic
débris in which a few scattering plants of P. robbinsii,
succeed in growing. There is a narrow fringe of yellow
pond lilies about twenty feet from the mouth of this latter
brook, and between the shore and this fringe of lilies a

332 THE AMERICAN NATURALIST [ Vou. LIL

few scattered plants of the seven-angled pipewort may
be seen. Ordinarily the stream habitat furnishes the
most favorable ground for vegetation, yet we see in these
streams a dearth of species and a limited growth that
clearly indicates limited and restricted food for such ani-
mals as live upon aquatic plants.

The barren sandy-shoal, the barren stony-shoal and the
deep-water habitats are each almost entirely free from
plant life. This leaves the vegetative to supply the neces-
sary food. Aside from two patches of bullrushes and two
small groups of yellow lilies, the vegetative habitat is re-

#stricted to the plants that form on the slope leading to
the fifteen-foot contour. At station 12, for possibly an
eighth of a mile, there is a thick fringe of aquatic plants
composed of Potamogeton prealongus and robbins. In
the southeastern part of the lake and also near station 8
two other thick areas of plants occur. These consist of
Potamogetons, with the addition of a third species; P.
oakesianus. Of the possible four miles of this slope
around the lake not more than one sixth supports plants.

The study of these habitats then shows that the fish
are limited to the vegetative habitat in their search for
such food as lives in turn upon aquatic plants.

In order to determine what the fish were actually living
upon, a study of the cional: contents was made, of which
the following is a summary

Salvelinus fontinalis Mitchill. Brook Trout: The
black-striped minnow (Leuciscus carletoni), grasshop-
pers, crayfish, snail (Campeloma decisa), a few insect
larvæ and a pumpkin seed made up the diet of the twenty
stomachs examined.

Eupomotis gibbosus Linnæus. Pumpkin seed. Col-
lected at station 8, 20 to 25 mm. long.—Daphnia and

cyclops with an occasional insect larva. Fish of the same
size from station 5 were feeding entirely upon daphnia
and cyclops. Judging from the number found in some
of these fish, I would estimate that these small fish must

- eat more than 1,000 crustacea daily. One specimen had

Nos. 618-619] RESTOCKING INLAND WATERS 333

more than 100 rotifers which belonged to the species of
Hydatina. In all, about one hundred of these small fish
were studied from various stations in the lake and all
were found to be eating the same food, with but a small
amount of individual variation. The same is equally true
of the adults of this species.

Ameiurus nebulosus Le Sueur. Common Bullhead.—
The bullheads in Lake Clear vary their diet, as plant re-
mains, crayfish, clams, snails, plumatella and daphnia are
all found.

Catastomus commersonii Lacépède. Common Sucker.
—Plant remains, crustacean skeletons, sand, plumatella
and débris are all found.

Notropis cornutus Mitchill. Shiner.—Daphnia and in-
sects constitute their diet. A number were found with
honey bees in the stomach.

Leuciscus carletoni Kendall. Black-striped Minnow.—
Insect larvæ, rotifers, algæ, plumatella and daphnia were
all found. ;

In view of the importance of the whitefish as food fish
the details of this study are given.

Coregonus clupeiformia Mitchell. Common Whitefish;
Labrador Whitefish.— These whitefish are by all means
the most numerous fish in the lake, as from three to four
thousand are taken in the fall nets at once. More white-
fish are caught than all of the other species combined, so
far as my observations go. One would naturally expect,
then, that more of dead whitefish would be found along
the shore than of any other species. If one happened to
make his observations just when the whitefish are dying,
the above would be correct; but not more than two or
three times during a summer are any considerable num-
ber of whitefish to be found dead upon the shores. The
following notes illustrate this point.

Dead whitefish collected between September 19 and 24,
1916, just as they drifted onto the north shore between
the small brooks stations 1 and 2, and the East Flats, give
the daily record as follows: September 19, two males,

334 THE AMERICAN NATURALIST [VoL. LII

124 and 83 inches long; September 22, 4 whitefish from
shore; September 23, 5 whitefish from shore; 3 others
partly eaten by crows. All of these eight fish appear to
- be in a healthy condition and show no evidence of star-
vation; September 24, 5 more whitefish from shore. The
intestines of three others were taken as the body of the
whitefish was already mutilated by crows. During this
week a strong southwest wind blew.

On July 2 and 3 a similar series of dead whitefish was
found on this same stretch of shore. A dozen fish were
noted, all of which were between 12 and 15 inches in
length. These had all been partly eaten by crows when
first observed, so that it was impossible to learn any-
thing about their food. The crows begin their attack
upon the body in the gill region and drag out the viscera
through this opening. After the visceral delicacy is eaten,
the dorsal muscles are gradually removed. The crows
‘ake from two to three days to eat up a whitefish. None
of these twelve fish was poor or showed any sign of star-
vation. A strong southwest wind had been blowing for
- several days.

During the past ten summers I have noticed a similar
series of conditions. Two or three times each summer, a
large number of whitefish are found dead on the north- —
east shore. Occasionally, I have noticed the skeletal re-
mains of whitefish on the west and south shores. During
these irregular times when whitefish are drifting on shore,
there are more of them than suckers, bullheads, brook or
lake trout. Rheighard (’13, p. 224) says:

Great numbers of dead suckers are thrown up on the beach in South
Fishtail Bay in July and August. Many of these have the character-
istie form of starved fish. The back is thin and sharp instead of round,
and the head is disproportionately large compared to the body... +
The emaciated fish do not appear to be diseased and are not usually
parasitized heavily enough to account for their emaciation.

Colbert (’15, p. 35) names five causes of death in fishes,
as follows: (1) Mechanical injury; (2) injury through
attacks of other species; (3) the beaching of individuals

Nos. 618-619] RESTOCKING INLAND WATERS 335

while pursuing or swallowing prey; (4) the accidental
beaching while attempting to escape enemies; (5) disease
and parasites.

These are the two most important of the recent obser-
vations on death in fishes. The size and general healthy
condition of the whitefish collected eliminate all but the
first cause given by Colbert. It is well known that white-
fish are easily killed by handling and do not have the
tenacity of life so characteristic of suckers or bullheads,
for example. The whitefish, although occupying the deep
basins of the lake, frequently come to the surface to play.
On a calm summer evening one can hear them as they
spring out of the water. The splashes which they make
are more numerous in the deep water, while the brook
trout are seen in shallow water near the mouth of the
brooks. This is a common habit of whitefish in Lake
Clear, especially during July and August. That they
come to the surface is also shown by the fact that at times
many are caught with not more than six or eight feet of
line. Their habit of coming to the surface makes it pos-
sible to see how wave actions might cause their death.
None of the other causes cited by the observer quoted
explains the death of these whitefish. In the detailed
study it was found that all of the dead whitefish collected
this summer were males, which renders it all the more
difficult to understand how wave action may be the only
cause of death.

The following study of the stomach contents of these
dead whitefish throws a good deal of light upon their
food habits:

No. 1. Male. Length 113 inches. Intestine de of food. Stomach con-
tents badly macerated and most of it impossible to identify. Apparently
almost entirely Daphnia kahlbergensis. No apok The mesentery con-
tained a large amount of fat

No. 2. Male. Length 114 inches. Stomach and intestine contain many
minute crustacean skeletons. The remains of three honey bees covered with
saprolegnia. A small amount of fat in mesentery.

No. 3. Male. Length 124 inches. Stomach and intestime contain nu-
merous minute macerated erustacea. Mesentery loaded with fat.

336 THE AMERICAN NATURALIST [ Vou. LIT

No. 4. Male. Length 113 inches. Thousands of minute macerated
erustacea in stomach and anterior part of intestine. Remains of three
honey bees. Small amount of fat in mesentery

No. 5. Male. Length 12 inches. One honey bee. No fat in mesentery.

No. 6. Male. Length 11 inches. Numerous Daphnia kahlbergensis.

S;
O. Male. Length 12 inches. One honey bee. Large amount of fat
in mesentery and around stomach.
Male. Length 114 inches. lag empty. Intestine partly
tal he gesta food. Large amoun
ale. Length 11 inches. a contained numerous Daphnia
eto and Leptodora hyalina. No copepods. The intestine was 5
inches long and +; of an inch in diameter and was packed full of cladocera
skeletons

The food in numbers 1, 6, 9 were identified by Dr. C. D.
Marsh. It is probable that the minute crustaceans po
in 2, 3, 4, 8 are the same as those found in 1, 6, 9.
material was so badly macerated that it was + a Buaibtt
to be confident of the identification.

The large amount of food found in the stomach and
intestine, and the presence of a great deal of fat in most
instances, is convincing evidence that these fish did not
starve to death. It seems strange that they were all
males. The honey bees eaten had evidently been in the
water several days, as practically every one was covered
with saprolegnia.

The following whitefish were collected by Milo Otis,
superintendent of the Saranac Inn Hatchery. The fish
were taken during November, 1916, in nets used to secure
spawning fish.

No. 10. Male. Length 11 inches. Stomach empty. Duodenum con-
tained minute crustacea, too macerated to identify. In the intestine were
found numerous winter eggs of Daphnia. These winter eggs appear to be
uninjured by the digestive enzymes. Some were found in the rectum in a
perfect condition, so that I feel P that most of them pass through
the digestive canal and into the water ready to grow into Daphnia. This is
an important fact besa the problem of an adequate amount of food is
taken into considerati

No. 11. Female. Sia 113 inches. gage: empty. Macerated
eladocera in intestine. Many tapeworms pre

No. 12. Male. Length 113 inches, con pt Intestine contains

No. 13. Female. Length 11 inches. Stomach and intestine mostly
m Parasites present. Ovaries full of eggs.

Nos. 618-619] RESTOCKING INLAND WATERS 337

No. 14, Male. Length 11 inches. Stomach empty. Intestine with
cladocera skeletons. Parasites present.

No. 15. Female. Length 84 inches. Stomach and intestine empty.
Celome full of eggs.

The following viscera taken from whitefish collected by
Milo Otis were received November 24, 1916:

No. 16. Stomach empty. Several parasitic flatworms present.

No. 17. Stomach contained 3 pumpkin seeds (Eupomotis gibbosus) 1
inch long; 1 partly digested gibbosus; 1 snail (Amnicola Limosa Say
identified by F. G. Baker). Parasitic flatworms present.

No. 18. aya contained many Daphnia kahlbergensis and Lepto-
dora hyalina. No e ods,

No. 19. Stomach ail numerous Daphnia winter eggs; 8 whitefish
e,

. 20. Stomach full. One pumpkin seed (Eupomotis egibbosus) and
mee visten eggs. Era flatworms present.
. 21. Stomach empty, with many parasites.

ig 24. Stomach full of whitefish eg

No. 25. Stomach contained 1 partly digested fish, probably gibbosus.

Nos. 26, 27. Each contained many Dap

No. 28. Stomach cel Daphnia = ite eggs.

In the jar containing the viscera of the whitefish num-
bers 16 to 28, there were, in addition to the above records,
20 pumpkin seeds that had been in some of these stomachs
and 107 whitefish eggs. 33 tapeworms were also found
in this residue.

Small pumpkinseeds were selected in order that the
food habits of very young fish might be compared with the
adults. Small fish 20 to 25 millimeters (about an inch),
25 to 30 millimeters, and adults 100 millimeters long (3 to
4 inches) were examined.

These young fish live almost exclusively on small clado-
cerans, with cyclops and daphnia predominating. Occa-
sionally one was found having only rotifers. Insect
larve do not play an important part in their food so far
as my studies go. The adults include, in addition, plants
and Plumatella.

Baker (1916, pp. 184-188) gives a summary of the facts
of the food habits of this species in which insect larve
and mollusca are seen to be the most important. In
Walnut Lake, Wisconsin, insect larve are mainly eaten;

338 THE AMERICAN NATURALIST [ Vou. LII

while in Douglas Lake, Michigan, and Oneida Lake, New
York, molluses are the more important. In Lake Clear, .
cyclops and daphnia may be said to be their main food.
These two animals happen to be almost the only food of
the whitefish. The pumpkin seeds in Lake Clear thus be-
come a hindrance in the stocking of this lake. The num-
ber of pumpkin seeds eaten by other fish does not appear
to be large. Itis suggested that food fish which live har-
moniously with whitefish but feed upon pumpkin seeds
would make a valuable combination.

Forbes (pp. 108, 1883) made numerous experiments to

determine the natural food of young whitefish. He found
that cyclops were more important than all the other or-
ganisms combined. Hankinson (p. 239, 1914) also noted
the almost exclusive diet of cyclops and daphnia. Baker
(pp. 159-161) shows that not only crustacea, but molluscs
are important as food for adult whitefish. My observa-
tions emphasize the limited diet of large whitefish in Lake
Clear where cyclops and daphnia are all that are eaten
during the summer. In the fall some specimens were
taken with the snail Amnicola in the stomach, but this
snail is present in limited numbers only, so can not be
very important as a source of food in Lake Clear. Dur-
ing the fall, small pumpkin seeds and their own eggs are
added to the daphnia-cyclops diet. It is to be noted that
all of the dead whitefish were in good condition, most of
them being fat. These all fed upon the daphnia-cyclops
diet.
These cladocerans produce winter eggs in large num-
bers which are not destroyed by the digestive juices of the
whitefish nor the other fish that feed upon them. These
eggs pass through the entire digestive canal uninjured.
This is an important factor in keeping up the number of
these minute organisms. Were these winter eggs di-
gested and used as food, it is probable that this, the most
important source of food for whitefish, would soon become
exhausted,

The brook trout taken in Trout Brook or at its mouth

Nos. 618-619] RESTOCKING INLAND WATERS 339

feed for the most part on insects; while those taken in the
lake eat minnows almost exclusively. The two taken
early in the spring (the ice did not break up until after
May first) had been feeding upon pumpkin seeds, cray-
fish and insect larve. No molluscs were found.

The food of the three species of minnows examined as
well as the suckers indicates that these fish live largely
on the same animals as the whitefish and the pumpkin-
seeds. The bullhead is the most general in its diet of any
of the fish studied, taking clams, snails, crayfish, minnows
and plants.

A detailed study of the food habits of the fish in any
lake is necessary before one can tell just what the several
species of fish actually eat. It is to be regretted that in
this lake the number of organisms suitable for food for
fishes is so limited. The result is that each species comes
into competition with the other species for food. The
result of this competition for the one abundant food,
daphnia-cyclops, prevents this lake from permanently
having large numbers of food fish.

The consideration of the history of the lake, the specific
babitatiou of the fish, the noteworthy dearth of aquatic
plants, the actual food of the fish and the restocking that
has taken place during the past thirty years leads to the
conclusion that restocking has not been and cannot be a
success. In estimating how many fish any given body of
water will support, one must first consider the variety and
abundance of aquatic plants. There can not be any more
animal food for the small fish and fingerlings than can
find subsistence on the aquatic plants of any given body
of water.

In this connection a second question might be asked,
why stock any Adirondack pond with a distinctively food
fish? Whitefish when not ‘taken in nets are caught at
- baited buoys which are placed early in the spring by the
local fishermen. I have counted 25 buoys scattered
around the lake in early June. At these buoys a consid-
erable number of whitefish are taken by relatively very

*

340 THE AMERICAN NATURALIST [ Vou. LII

few people, probably not over fifteen separate families.
Sportsmen who desire to fish for whitefish employ some
one who has a buoy located advantageously as a guide.
This guide ties the boat to his buoy. In a half day, from
six to twenty whitefish may be thus taken. No one is ex-
pected to tie to one of these buoys without the ‘‘owner’s”’
consent. The result then of this extensive stocking by
the state has been to enable a few local families and
guides to catch for their own use and to sell a few white-
fish. The general vacation transients are not benefited
nor is this excellent food fish taken in such numbers as
to yield any considerable amount of food. If the state
is to continue to restock such ponds as Lake Clear with
whitefish, then some better method should be devised to
take the whitefish.

The fundamental reason why the extensive restocking
of Lake Clear has not been a success is due to the glacial
origin of the lake. The mere fact that the sand in which
the plants try to grow has been resorted by the glacial
waters until most of the organic material, plant food, has
been washed away, produces a very limited number of
aquatic plants. Such plants are indispensable as a source
of food for the numerous minute organisms upon which
fish normally feed. One can hardly appreciate what a
large number of different animals fish eat unless he has
given careful study to this problem. Baker (pp. 157-
199) gives a summary of the different kinds of animals
eaten by fish and the variety is in striking contrast to the
one abundant group of organisms (daphnia-cyclops) in
Lake Clear. In order to have a general growth of fish
in a lake, I believe that an abundance of several kinds of
fish food is indispensable. The conditions in Lake Clear
well illustrate how the species that can most successfully
utilize the form of food that is abundant survives and
greatly increases in numbers, while the others remain
few in numbers.

The shores of Lake Clear are remarkably free from

dead fish and one rarely finds any dead fingerlings. It

' Nos. 618-619] RESTOCKING INLAND WATERS 341

would seem as if there were but one conclusion, namely,
that the most of the fingerlings are eaten by larger fish.

The probabilities are that there would not, then, have
been any more trout in Lake Clear, even if the state had
placed many more than it has during the past thirty years.
This lake really illustrates an instance of overstocking
simply because there was not an adequate amount and
variety of food. At present there does not seem to be
any way to remedy this deficiency.

During the past summer the states have been urged to
increase greatly their production of food and game fish.
This very general recommendation fails to recognize cer-
tain fundamental facts. The next step in advance along
fish propagation is one that will have to be taken slowly,
as it necessitates a great deal of critical information. It
is becoming more and more apparent that we must not
only know the breeding habits of the small minnows,
pumpkin seeds, etc., the fry of which serve as admirable
food for the food-fish fingerlings, but also the natural his-
_ tory of all of the life of a given body of water. It is a
well-recognized biological axiom that no organism can
live unto itself alone. This applied to our problem means
that a clear and adequate supply of water is not the only
factor that must be considered in deciding to restock great
bodies of water with fish-fry. But rather the intricate
and more or less obscure conditions that determine the
sum total of life in each body of water must be taken into
consideration. Such studies alone furnish a correct basis
for determining the extent to which an animal may draw
upon a given source of food, upon the available body of
food, and many kindred problems. Before the state can
wisely undertake to place more fingerlings in the ponds,
it ought to know whether there is enough available food
to keep them at least from starving.

The extensive restocking of most of the Adirondack
ponds is done mostly for the benefit of the sportsmen.
The benefit that has accrued to these and many other
vacationists is very great, but one can not help wondering

342 THE AMERICAN NATURALIST [Von.LII *

if this policy should not be restricted and adjusted. One
or two trout hatcheries would probably be able to pro-
duce all of the trout needed, while the rest might be
turned over to the production of food fish which should
be placed in the larger bodies of water that support an
abundant aquatic plant life. The conditions in Oneida
Lake are in such striking contrast to those in the Adiron-
dacks that one can unhesitatingly suggest that this lake
could support an incredible number of fish.

Intimately associated with this general problem is the
question of disease in fishes. Under this heading are in-
cluded the several forms of parasitism. Reference will
be made to two or three only. This is the field in which
the greatest progress has been made in the cure and pre-
vention of the diseases affecting man. But before pre-
vention can be applied, the life history of the parasite
must be understood. In the main, the work of the na-
tional and state governments has been confined solely to
hatchery problems. Here is a field that they should en-
ter, as many of the problems are too large and involve
too much expense for the individual.

Salmincola edwardsii (Olsson).

This parasite belongs to the copepod group of crus-
taceans, many of which are familiarly known as ‘“‘fish-
lice.”? Wilson says: ‘‘This family (Lernwopodide) of
parasites is widely distributed amongst fishes in both salt
and fresh water. Some of our best food and game fish
are infested with them, and when they once obtain en-
trance to a stock pond, fish hatchery, or aquarium they
usually multiply so rapidly as to become a serious
nuisance, and may even kill the fish”” (p. 569). There
, Are some one hundred and thirty-six different species of
animals that belong to this family, all of which are para-
istic during their adult life. In the genus to which our
specimens belong, there are twenty species, eighteen of
which live exclusively upon the several kinds of trout.

Salmincola edwardsii is found exclusively infesting the

brook trout (Salvelinus fontinalis). All of the specimens

Nos. 618-619] RESTOCKING INLAND WATERS 343

taken in Lake Clear were attached to either the dorsal or
anal fin. ` The gills of these fish were examined but no
fish-lice noticed upon them. This seems strange as
Fasten (1911-12) reports them as especially abundant
upon the gills. They were found on trout ranging from
four inches in length to those fourteen inches long, which
were caught in Trout Brook, in the lake near the mouth
of this inlet, station 1, and in the bay near station 5.
They were taken upon both large and small fish in each
of these locations. My specimens were collected during
July, August and September. The last specimens were
on a nine-inch male caught in Big St. Germain Bay, Sep-
tember 28, 1916. The egg-sacs were full of embryos.

This common ‘‘fish-louse’’ is easily recognized. From
the main part of the body two large egg-saes are sus-
pended. With an ordinary hand lens, the numerous
small embryos can be noted, and the small beak which
attaches the parasite to its host. Fasten has recently
worked out the life history and the habits of this inter-
esting parasite, which undergoes an extensive degenera-
tion after becoming parasitic.

These parasites are widespread in the United States in
the native trout streams, and in Canada and Europe.
The first scientific record of this particular parasite is by
Linneus in 1761. It seems strange that an animal could:
be known for so long and its habits not be understood
until within the past five years. The fact that it is so
widespread and has been known for so long indicates that
it is not a serious pest except under very favorable
living conditions. These are best found in the hatcheries
or stock ponds, where many fish live in a small enclosure.
The numerous parasitic larve then have little trouble in
finding a host. In the streams and ponds of this state,
we need not fear that the trout will be killed by them.
The chief reason is that the trout are few in numbers in
any given place, so that when the embryo parasites make
their escape, there is small chance of their ever becoming

attached to a trout. The result is that each trout in a

344 THE AMERICAN NATURALIST [VoL. LII

wild state does not usually harbor more than a half dozen,
and this number does no serious harm to the trout, espe-
: cially where they are attached to the fins. i

When this pest gets into a hatchery, but little can be
done. The infested trout at the Wild Rose Hatchery in
Wisconsin were treated with ‘‘solutions of copper sul-
phate, potassium chlorate, sodium chloride, and mixtures
of sodium chloride and potassium chlorate, but these had
no effect upon the parasite” (Fasten, 1911-12, p. 17).
About the only remedy that is effective is to destroy all
of the parasitized trout, which can be done after the
spawning season; for there is no reason to believe that
these parasites can be carried from one place to another
except in the manner described by Fasten. We do not
know how these parasites pass the winter, but in view of
the fact that they are not known to eat during their free-
swimming period, it is probable that those which live
through the winter do so as parasites upon trout. Dur-
ing the cold weather the growth changes would not take
place as rapidly, so that a given parasite might remain
on a trout for five or even six months during the winter.
Clinostomum marginatum.

This is a small Trematode that lives for a part of its
life embedded in the muscles of several food and game
fish. The popular name of ““grubby”” is used to describe
this condition. So little is known about this parasite that
renders thousands of dollars worth of fish unsuitable for
food that any new facts are welcome. In this connection,
the following field observations are recorded upon this
very annoying parasite. In an earlier paper (Smallwood,
"14) attention was called to the fact that these worms may
voluntarily leave the body of their host after the host dies.
On May 28, 1914, about 50 perch were taken from MeCau-
ley Pond between Saranac Lake village and Lake Clear
Junction. After returning to camp, it was noticed that
these perch were ““grubby”” and they were all left in a
pile on the ground. The next morning, I examined the

pile of perch and was able to pick up more than 100 flat-

Nos. 618-619] RESTOCKING INLAND WATERS 345

worms that had crawled out of their cyst during the night.
If I had cared to do so I could have gathered several
hundred specimens during the day as they continued to
escape from the cysts.

This observation answers one question, namely, that
these parasitic worms remain in the body of the perch
all winter, as the worms were in the same stage of ma-
turity as those taken in July, August and September.
It further suggests, also, that Clinostomum marginatum
may return to an aquatic habit directly from the body of
the fish.

Guides and fishermen claim that perch are free from
““grubs”” at certain seasons of the year, a belief which I
am unable to confirm. But if more extended studies show
that their belief is true, the method by means of which
the ‘‘grubs’’ leave the fish has been found.

So far as we know, these parasites are mainly confined
to members of the bass and perch families, although a
few cases are on record of trout being infested with them.
Occasionally one finds a few bullheads that are ““grubby.””
In Lake Clear Lepomis gibbosus and the minnows are
largely infested by them.

I can also confirm my observations (p. 11, 14) that fish
eating parasitized minnóws are not themselves infected
through this avenue. One brook trout was found with
a partly digested minnow in its stomach. In the flesh
of the minnow a dead parasitic clinostomum was ob-
served. :

I would say that the number of fish harboring this
parasite in Lake Clear is on the increase.

The following observations upon some larval stages in
Trematodes may help to start some one upon this prob-
lem without losing a season experimenting.

During the past three years, I have frequently found
in the muscles of perch and pumpkin seeds minute cysts.
Each of these cysts contains a larval Trematode.

The wall of the cyst is tough and easily ‘‘shells out”” of
the muscles, but is not readily penetrated by the ordi-

346 THE AMERICAN NATURALIST [ Vou. LII

nary killing fluids. This has resulted in my not being
able to secure suitable sections for detailed study. It
will probably: be necessary to make most of the observa-
tions upon living material, which will require that one
have a microscope and microscopic reagents while doing
field work.

The following general facts may help to direct atten-
tion to this important stage in the development of the
life history of Trematodes. Possibly some fish contain-
ing them may be found near a laboratory and thus readily
studied.

y | @
F 4. Lepomis megalotis (Rafinesque), long-eared sunfish. Taken in
Lake Clear, 1915. A portion of the skin has been removed to show the position
that the larval stages of certain undetermined Distomes occupy in the flesh.

Fig. 4 shows the usual position of these larval stages
in the muscles of the long-eared pumpkin seed taken in
Lake Clear during July, 1915. Each cyst is accompanied
by a small amount of pigment in the more advanced
stages, although I have seen many of them with no pig-
ment and nearly the same appearance as the muscle (Fig.
7). One must often pick the muscle fibers apart in order
to find the cyst as most of them do not show on the sur-
face. I have not been able to discover that they are
found in any definite region of the body, although they
are more numerous near the dorsal fin.

A photomicrograph of a young cyst shows a number
of blood vessels entering one end of the cyst. The cyst

Nos. 618-619] RESTOCKING INLAND WATERS 347

itself contains numerous blood corpuscles, as if they had
been emptied into the cyst. The difficulties of fixation
have prevented me thus far from studying the character
of the blood in these cysts.

As the cyst grows, the wall of which is made up of con-
nective tissue, partitions are formed which grow in from
the main outside wall. The result is that in the older
stages three to five separate cavities are found in each
eyst (Fig. 6), although but a single larval Trematode is
present in each cyst. I have opened a large number of
these cysts and thus far have found no exceptions.

Fie. 5. The perch (Perca flavescens) showing minute distomes embedded
in the skin and fins. These parasites ‘can be recognized in the photograph as
minute, black spots. The black pigment surrounds the parasite.

Not much can be said at this time in regard to the struc-
ture and changes through which this larva is passing.
That it is growing there can be no doubt, as different
stages showing the presence or absence of some of the
mature organs were found. In most of the whole larve
dissected and mounted, the excretory ducts are com-
pletely formed, while the digestive tract is limited chiefly
to the anterior end. Two suckers can be recognized and
a part of the reproductive organs. The whole animal is
so very small that it can scarcely be seen with the un-
aided eye. It will require better fixation before the spe-
cies can be determined accurately. I am inclined to

348 THE AMERICAN NATURALIST [ Vou. LIT

believe that it is the larval stage of either some species
of Holostomum or Clinostomum. The latter is more nu-
merous in this lake; but the size and pigment strongly
suggest Holostomum, the adult appearance of which is
shown in the photomicrograph (Fig. 7

From the studies thus far made, I am inclined to think
that it takes one season for this larva to transform into
the adult worm. My reason for this conclusion is that

viens taie aa of the pin cyst of ey undetermined distome.

Fic. 6.
Each cmi contains but e young worm, although there are several parti-
tions. The cyst is surrounded by dica! layers of connective tissue and is

embedded in the muscles

there is so little difference between the several stages that
I have secured.

In the hope of throwing some light upon the relation of
Trematodes to fish, the writer urged several years ago
that some ‘‘grubby’’ perch be placed in a separate tank
in one of the hatcheries and fed for one year to see what
happened, but the suggestion was rejected as not prac-
tical, although a number of specific experiments were out-

Nos. 618-619] RESTOCKING INLAND WATERS 349

lined. It is unnecessary to say that many thousands of
dollars’ worth of fish would be rendered available through-
out the United States, if this one disease alone could be
prevented. Perch and bass which are so generally in-
fected with Trematodes are delicious pan fish and usually
easily caught in large numbers. It would seem as if the
mere calling attention to this large amount of food an-
nually wasted would stimulate some organization to
finance the necessary scientific investigation. Such an
investigation would need to continue at least two years
and possibly longer; but the expense involved would be
a small fraction of the returns, if the disease could be pre-
vented.

““Grubby”” perch are but one illustration of the diseases
that occur in our fresh-water fish. There are a number
of diseases due to bacteria and others caused by certain
Soporozoa, Fig. 7. These diseases are usually epidemic,

Fic. 7. Leuciscus carletoni (Kendall), black-stripped minnow. vera in Lake
Clear, 1916. Infected with Myxobolus, a sporozoan para

killing large numbers of fish in a few weeks. The main
fact that is known now is that the fish die and several
specific microorganisms are observed to be associated
with certain ulcers, cysts, ete. But as to what causes
lead up to the fatal termination of the diseases, little is
known.

An admirable monograph upon carcinoma of the
thyroid in the Salmonoid fishes was prepared, but left in-
completed because of lack of funds. This is about the
only serious study of fish diseases that has been made in
the United States.

At this time, when all are anxious to help in whatever
way that they can, it may be permissible to suggest that

350 THE AMERICAN NATURALIST [VoL. LIT

scientists may render real service along the general lines
indicated in this paper. To be more specific, they may be
summarized as follows: `

1. A critical study of the life of the fresh-water ponds and streams is

very pat ble.

a seer the life history and food habits of the small fish and in-

ttebrates present.
(b) Yo determin e the amount and conditions of plant life—the ultimate
of food for all animals.

(e) m determine if more plants can be made to grow in water naturally
ng sufficient plant li

(d) To 4 determine if it is somible to introduce enough natural food for

fry and fingerlings to keep them from starving, as many of them

probably do now in such ponds as many of those in the Adiron-

dacks,
(e) To determine to what extent the natural food of the fish is eaten by
invertebrates living in the same water

When these problems are solved, the next step in effi-
ciency in fish culture can be taken. It will require the
cooperation of many scientists. The result will be the
substitution of an intelligent method of restocking in
place of the present one, which is often unintelligent or
politically influenced.

2. A detailed study of the several diseases occurrin
(a) It is necessary to know the complete life history of dc parasites as
the Trematodes before any one can formulate la meas-
- There are at present at least four different species of Tre-
idos found in our fresh-water fish and the ARNO life his-
tory of each is unknown.
(b) The causes leading up to the usual epidemies in fish must be deter-
i these can be prevented. This is a problem for the
bacteriologist and the protozoologist.

When these problems are worked out, an enormous
amount of fish food will be conserved for human needs.
SYRACUSE UNIVERSITY.
October 29, 1917

BIBLIOGRAPHY

The an all list of references includes only those works cited in the
; pages. A more extended reference to the literature will be found
in E. c. Baker’s paper, which is the first one given in the list below.

Nos. 618-619] RESTOCKING INLAND WATERS 351

Baker, Frank Col
1916. The sade of Mollusks to Fish in Oneida Lake. Technical
Publication No. 4. The New York State College of Forestry
at Syracuse University, pp. 1-366.
Bean, Tarleton H.
1903. Catalogue of the Fishes of New York. Bull. No. 60. New
ork State Museum, pp. 1-784
Calkins, Gary N

1899, Baport upon the Recent Epidemic among Brook Trout (Salve-
linus fontinalis) on Long Island. Fourth Annual Report of
the Commissioners of the Fisheries, Game and Forests of the
State of New York. Report for the year 1898, pp. 175-189.

Colbert, Roy J.

1915. An Ecological Study of the as of the Douglas Lake Region,
Michigan, with special reference to the morta some of the spe-
sa Michigan Geological sie Biolo ge T Publica-

on No. 20, Biology Series No. 4, pp.

1911-12. Te Brook Trout Disease in Wisconsin Tun Report of

e Wisconsin Commission of Fisheries, pp. 11-21.
1913. is Behavior of a Parasitic Copepod. Jour. Animal Behavior,
i 6-60.

1914. Fertilization in the Kenio aes Lerneopoda ‘edwardsit.

Biol. Bull., Vol. 27, pp.
Fenneman, N. M.

1910. Lakes of epr Wisconsin. Published by the State of

i

ra ae
The zo Food of the Common Whitefish (Gorogonus -ni
formis Mitchill). Bull. Ill. State Lab. Nat. Hist., I, No.
pp. 95-110.
Gaylord, H. R., and Marsh, M. C.
1912. Carcinoma of the Thyroid in the Salmonoid Fishes. Bull. Bur.
Fisheries, Vol. XXXII, 1912. Document No. 790. Issued
April 22, 1914
Gurley, R. R.
1894. The Myosporidia or Psorosperms of Fishes, and the Epidemics
Produced by Them. U. S. Commission of Fish and Fisheries.
Co mer Seu a. Published 1894. Pp. 65-359.
Hankinson, Thomas L.
4. Youn pa > Whitedsh 3 in Lake Superior. Science, N. S., Vol. 40, No.

1024, p. 239.
1915. serrat on the Fishes of A County, Mi chigan.
Michigan Geological and Biological Survey. Publication No.
20, Biological Series No. 4, pp. 11-
Needham, J. C., and others.
1903. Aquatic Tiset i in New York State. Bull. 68. New York State

Museum.
Needham, J. C., and others.
1905. May Flies and Midges. Bull. 86. New York State Museum.

352 THE AMERICAN NATURALIST [Vor. LIT

nen Paul
1908. A Plan of Promoting the Whitefish Productions of the Great
es. Bull. U. S. Bureau of Fisheries, XXVII, pp. 643-84.

Smallwood, W. M.

1914. Preliminary Report on Diseases of Fish in the Adirondacks.
A Contribution to the Life History of Clinostomum margi-
natum. Technical Publication No. 1. The New hase State
College of Forestry at Syracuse University, pp.

Smith, Henry I
1874. The Crustacean pies of the Fresh-water Fishes of the
United States. U. S. Fish Commission. Report for 1872-73.

e. Charles B.

15, North American Parasitic Pl os Belonging to the Lernæo-
poda, with a revision of the entire family, No. po Pro-
ceedings U. S. National prada Vol. 47, pp. 565-72

SHORTER ARTICLES AND DISCUSSION
MODIFICATIONS OF THE 9:3:3:1 RATIO

A. CHEMICAL EXPERIMENTS PARALLELING THE SEVERAL POSSIBLE
MODIFICATIONS OF THE MENDELIAN F, DI-HYBRD PHENO-
YPIC NON-BLENDING Ratio 9:3:3:1

Tue foundation F, di-hybrid ratio 9:3:3:1 consists of 9 in-
dividuals having somatic traits both ‘‘A’’ and ‘‘B,’’ 3 individ-
uals having ‘‘A’’ only, 3 having ‘‘B”’ only, and 1 having neither
“A” nor ‘‘B.’’ Or, if each allelomorphie pair consists, not in a
gene and its absence, but in genic entities contrasted in quality
or quantity and showing clean-cut dominance and recessiveness,
in 9 ‘‘A’’ and “*B”” both dominant; 3 ‘‘A’’ dominant, ‘‘b’’ pe
cessive; 3 ‘‘a’’ recessive, ‘‘B’’ dominant; 1 ‘‘a’’ and ‘‘b’’ both
recessive. This di-hybrid ratio was one of de early discoveries
of Mendel! himself, but after the revival of genetical studies in
1900 experimental breeding had not continued long before modi-
fications of this ratio became apparent. Thus Bateson? men-
tions a number of cases in which the 9:3:4 F, ratio is found.
It is apparent in such cases that two unit trait-pairs are in-
volved, that the dominant phase of one of them standing without
that of the other in 3 individual F, somas is not to all patent as-
pects different from the 1 individual possessing the dominant phase
of neither of the two trait-pairs involved. In the gametes of in-
dividuals of families that produce the 9:3:4 ratio the segrega-
tion and recombination of genes are, however, just as clean-cut
and follow the same rule as in pedigrees giving the unmodified
foundation ratio 9:3:3:1; only the somatie working out of the
genes is different in the two ratios.

Barring blending, linkage, crossing-over, non-disjunction, sex-
limited inheritance and other special phenomena, which limita-
tions preserve intact the numerical entities 9, 3, 3, and 1, the

1 Mendel, G., ‘‘ Experiments in Plant Hybridization’’ (reprinted in Eng-
lish in Bateson ’g ‘*Mendel’s Principles of Heredity,’’ pp. 334-379), p
351, 1866.

2 Bateson, W., ‘‘Mendel’s Principles of Heredity,’’ p. 80, 1913,

354 THE AMERICAN NATURALIST [VoL. LIT

Bim omar Illustrative of Black Skin-Pigment ees wis
h of the 16 zygote tubes

Male Gam

HUMAN SKIN-COLOR
IN WHI AGRO Fa M eles
passing 16+ 162
+ L
CUMULATIVE ae PS
GENES.
AB At as ae

>
e 2 Black
ES
GE ©
¡Ms
as A fe
T q Y
El E) Cea
vB) y k
54 O y
on 6
®| © pS
Ra Ù P Octoroon
CARS
A (pass-for-while
3
J
a a deville: hake
2 Az4cc. on of as. and 17.5 clear water:
Gene Bo ace of cantina tice ri eb ea ce Lee a
@=40.c. of. c.Ifnóta Ink and 940.5 c.c. clear water.

following table presents in orderly fashion all of their possible
ratio-recombinations :

Series A
Case E... 93:85 L
Case IL 9:3:4.
Case III. 9:6:1.
Case IV. 9:7.
Case V- 10:3;:3,
Case VI. 10:6.
Case VII. 12:3:1.
Case VIII. 12:4,
Case IX. 13:3.
Ca o- X., 15:1,

Nos. 618-619] SHORTER ARTICLES AND DISCUSSION 355

FiFemale Gametes
e O LS ES
=z
xXXAB OIC AS Xan HINA
R
w
X
y AABB AABE- AUBB AQ BE
Y y
¿O
z X
gy AABE AAC E AQ BE- AQE E-
an
1
3 X
= AQBB AaBe @aBB aaBe
y ee
E H
$ . AQBE AQEE aaBe aace
Fz Zyqgotes: Phenotypic Ratio 9:3:3:4

/Hustralive
Suéstralum xx = ban nce less).
ETa E ta bear om of Fe Ch e H CI (Yeller).
Gene B= Agueoys solution of Rides € PISA

The chemical experiments described in the accompanying fig-
ures parallel chemically what must happen to the F, genes in
their development into traits in the F, somas in each case of a
modified somatic di-hybrid ratio. Each drawing represents a
2-inch wooden block 74 inches square, with twenty-four ¿-inch
holes, 14 inches deep, suitable for holding test-tubes. The four
holes at the top hold test-tubes containing chemicals representing °
the four types of F, female gametes possible when di-hybridism
is being considered ; the four to the left perform the same service
for the four male gametes. The sixteen holes blocked off in the
square hold test-tubes for the sixteen types of zygotes resulting,
by the checker-board method, from all possible unions of the
four male and the four female gametes. X X represent the sub-
stratum or the sum-total of hereditary qualities other than the
traits under consideration. The lettering under each test-tube
circle is the genetic formula for the gamete, or zygote as the

356 THE AMERICAN NATURALIST [ Vou. LIT

Fy Female Gametes

DO FO RO
ORO

O
g 3 AQ@BB AQBe
hs y
JO
1,8 x
Q AABE AQBE aaee
S$/38
JO =,
= x
Q AQBB AABE aaBB aaBe
$
©;
AGBE aaBe aaee
Fz Zygoles:Phenolypic Ratio 9:2:4
illustrative E: :
MAR r Colfer toca)
Gene A= Age f: fi

Gene B = Dilute EI CCaloriasa 5).

case may be. A word within a test-tube circle names the color
of the reaction-product, which color is the index or analog of the
somatic working out of its accompanying genetic formula of the
soma.

Instructions for performing the experiment illustrative of
each ratio-modification are found in detail in the drawing for
the selected case. The chemicals representing gene ‘‘A’’ and
gene '*B”” and the substratum are, as indicated, poured into the
test-tubes representing the F, female and the F, male gametes.
Each gamete-holding test-tube should contain approximately 20
c.c. of the appropriate genes and substratum, Of this mixture
4 c.c. represent 1 gamete; and 1 such gamete is to be poured into
each one of the 4 zygote-holding test-tubes, directly below or
directly to the right, as the case may be, of particular gamete-
holding test-tube, following the checker-board method. Thus
each of the 16 zygote test-tubes will contain chemicals appro-

Nos. 618-619] SHORTER ARTICLES AND DISCUSSION 357

Fy Female Gameles
Case
Wiis

XXLAE xXxxXaAB

xXXIAB

X
Ñ

Xx

=
Gy
X
N e
AABB AABE AQBB AQBe
¿ $
d
KOF
IO}
y AABE AAL E ACBe AQC e
S q
e
= A@QBB AaBe aa,» acaBae
>
u y
8
O eu vey
Aa BE AQ EE aaBeée aa eec
Zyqotes:Phenolypic Ratio 9:6-1
Saks coco mae d

A= Ag Sak, KI cj A a ob theatosrn t Le Er
“Gene B' (White). a y
E B- Ho Hur 7 £ PL (C>-Mx05). (White).

priately representing 1 F, male and 1 F, female gamete free,
like the constituents of a united sperm and egg, to interact in
the zygote and in subsequent ontogenesis.

Some of these ratios, such as exist in Cases Nos. V, VI and
IX, are much more difficult chemically to contrive than others,
such as Nos. I, II, III, IV and X. It is not surprising, then,
that nature provides more readily cases in inheritance repre-
senting the latter ratios, while the former are being found only
by the most diligent genetical study. Not all of these ratios
have yet been found in nature by experimental breeding, but,
from time to time, a geneticist reports the discovery of a new
di-hybrid F, ratio which proves to be a member of this series.
Doubtless all of them will be found in time, but without intro-
ducing the special phenomena earlier referred to, Series A, con-
sisting of 10 cases, exhausts the possibilities of F, phenotypic di-

hybrid ratios.

358 THE AMERICAN NATURALIST [ Vou. LIL

Fi Female Gametes
Case C)
TY
XXAB xxAs xxan xxa s
O;
:
y AABB AABE AQBB AaBe
y
O! (om)
y X
$ AABE AALS AGBEe AQLE
Ol © ©
8
$ x
| AQBB AaBE QQBB aaBe
y $
8
Oj
AQBE Aaee aque aqQee
ante” Phenolypic Ratio 9:7

ee of AE oh ba ia z pre ta te A as fone Ao Poa.
cae Benen a slightly clouded with phenol-

B. A CHEMICAL EXPERIMENT PARALLELING A MODIFICATION OF
THE MENDELIAN F, Di-HyBrRD PuenNorypric Ratio 1:2:1:-
2:4:2:1:2:1 INvotvine Somatic BLENDING AND GENIC
SEGREGATION IN THE F, GENERATION

The ten cases described in the Series A are non-blending, in
which the entities 9, 3, 3, and 1 are kept intact, that is, they are
simply recombined. The mono-hybrid ratio on which they are
based is the normal dominant-recessive 3: 1 relation. The typical
blending ratio in the F, mono-hybrid is 1:2:1. Carrying this
- latter ratio into the di-hybrid classification in the same manner
as the 3:1 ratio was carried into the 9:3:3:1 classification
gives us the following: 1:2:1:2:4:2:1:2:1, a total of sixteen
individuals divided into 9 classes.

In Davenport’s® study on the ““Inheritance of Skin-Color in

3 Davenport, C. B., **Skin-Color in Negro-White Crosses,’’ Pub. 188 Car.
Inst. Wash., 1913.

Nos. 618-619] SHORTER ARTICLES AND DISCUSSION 359

Fz Female Gametes

se OOO,

dese

e 3 AABB: AABE AQBB AQBE

loleses

©

TERRE

Y

NOS cocos
AQBE AQE QABE aaee

F2 Zygoles: Phenotypic Ratio 20:3:3

Mustrative Experime
Ena pes Mss ARE irre ay TOR

Negro-White Crosses” he found that the amount of black skin-
pigment in an individual is determined by two genes in each >
parental gamete, and in measuring the intensity of black skin-
pigment in the members of a great many highly hybridized
families he found 5 definite maxima in the curve plotting the
distribution of individuals according to the per cent. of black
skin-pigment carried. Theoretically (unless genes “A” and
“B” are exactly equally potent in developing into a definite
per cent. of blackness showing in the unweathered skin) there
should have been 9 maxima in the above-described distribution
curve. Arbitrarily giving the genes the following potentiality
for somatic expression, ‘‘A’’ 16 per cent., ‘‘B’’ 19 per cent.,
‘Sa’? 1 per cent., ‘b’ 2 per cent. (which is not far from the
che facts of the eas) the somatic frequencies and black skin-
color percentages would run as follows: One 70 per cent., two
54 per cent., one 36 per cent., two 55 per cent., four 38 per cent.,

360 THE AMERICAN NATURALIST [ Vou. LIL

F1 Female Gametes

XXAB XNAS XXaB Meta E
--
ae es)
X .
Y AABB AABE AQBB AaBe
y $
+
X N
3 AABE AACE AQBe AQee
y R
<=
CC
= AQBB AGBE ea BB aaBe
Q y
8
CA 6
AQBE aaee aaBe aaee
aa aa a A AA
Fz Zygotes: Phenotypic Rato 10:6
Illustrative bxperiment:—_ ý

xx
Gene A= Dilute H CI + fein lo match “Gene B” (White).
ERN E TO" IDEN

two 21 per cent., one 40 per cent., two 23 per cent., and one 6
per cent., in a total of 16 individuals.

But practically 5 instead of 9 somatic types were actually
found because gene ‘‘A’’ and gene ““B” are so nearly equal in
value that it is not possible in a given individual, simply by
measuring the skin-color pigment, to determine whether gene
““A”” or gene ‘‘B’’ is responsible for the quantity of melanin
possessed. Thus in the 5 classes the two 54 per cent. cases
(AABb) and the two 55 per cent. cases (AaBB) constitute the
three quarters blacks or ‘‘sambos,’’ the two 21 per cent. (Aabb)
cases and the two 23 per cent. (aaBb) cases constitute the one
quarter blacks or ‘‘quadroons,’’ while the one 36 per cent.
(AAbb) case and the one 40 per cent. ( aaBB) case are so nearly
like the four 38 per cent. (i. e., the ““mulatto?” AaBb) that they
constitute a single class of six individuals. Hence the contracted
ratio 1:4:6:4:1.

Nos. 618-619] SHORTER ARTICLES AND DISCUSSION 361

Fy Female Gametes
Case
Wz
XXAB LLAC xxaB XX ae
A
el 20
X a
7 AABB AABE AQBB AQBE
+
y - |
OÀ ~
RS AABE AAEE AQBL AQE£Ee
Y
JO] S
X
= AABB AQaBe aaBB aaBe
a y
T: +
a AQBe AQ e aaBe aaee
Fz Zygotes: Phenolypic Ratio 12:3:1

Mustrative Experiment:

= Water poe 18 rhexs}.

As Davenport clearly points out in the matter of black skin-
pigment, each gene finds its definite somatic expression regard-
less of the presence of other genes. Since the two genes work
out into the same somatic trait (i. e., are duplicate genes) which
somatic expressions differ only in quantity, they might well be

called cumulative genes. The accompanying diagram (Ex-
periment Illustrative of Black Skin-Pigment Inheritance) gives
specific directions for running this experiment in a manner very
closely paralleling what Davenport found in nature.
experiment illustrates only one case in a long possible
series of modifications of the foundation F, di-hybrid phenotypic
blending ratio 1:2:1:2:4:2:1:2:1.

C. Orner F, Di-HyBRID SERIES

In a case in which gene ‘‘A’’ does not blend with its allelomor-
phic mate ‘‘a’’ in the F, soma, but presents a typical case of

362 THE AMERICAN NATURALIST [Vor. LII

t
i

Fi Female Gametes
a) SAS
XXAB XXAG XXAB xxas
LA]
<-
or
X
n AABB AABE AQBB AQBE-
v
De 9
Y <
O}
0 AABE AACE AQBL£ Aaee
k]
3O}
G X
= x
AQBB AQBe aaBB aaBe
Ty
O
x
AQBe aace aaBe aage
E Zygotes: Phenolypic Ratio 12:4
IHustralive 2 Experiment :—

Water (Colorless).
Gene lz guezas solution of Fe Cl, and litmus (Red).
Gene B= Dilute H CI (Colorless).

Mendelian dominance, while gene ‘‘B’’ blends with its mate
““b”” in the F, soma, the di-hybrid F, ratio resulting from com-
bining two such trait-pairs is found by multiplying the F, mono-
hybrid dominant relation 3 + 1 by the F, mono-hybrid blending
relation 1+ 2 + 1, giving when expressed as a ratio the somatic
F, di-hybrid relation 3:6:3:1:2:1. Among other special cases
to be considered are those involving crossing-over, non-disjune-
tion and sex-limited inheritance.

When the genic and somatie specifications of a ratio are
known it is not difficult to contrive a chemical parallel for it.
The use of such experiments consists not only in clarifying the
conception of the particular situation at hand, but also in pro-
viding in the laboratory a nearer approach than is usually used
for demonstration to what is actually happening in heredity. The
chemical analogy between what happens in the zygote-holding test-
tube and what actually happens in the somatic working out of the

Nos. 618-619] SHORTER ARTICLES AND DISCUSSION 363

Fy Fernale Gametes

O
E
O

xxXAB XLAC xxaB xxasL£
A
ls EDI,
u X
Y AABB AABE AQBB AQBE
¡el
9 X
10] AABE AAGE AQBE AQE- e
y
=~
Ol be
= x
My ACBB AaBe aass aaBe
$
Or >
w
AQBE Aate aaBe aace

F2 Potes: Phenotypic Ratio 153:3

Hiustralive E

Subctralum xx ey PERES REA JO PAC AGAR pisa volume
Gene A= Golution of NAg OH a volume to 2 time
Ge

ze B= D hake Ba GY Times (Rea,

segregable genes in the members of the F, ratios is doubtless
much closer than any completely mechanical contrivance can
show. In the experiments the blending quite properly is shown
complete in the zygote and soma, but if the analogy were carried
still further one should be able to dip into the gamete-holding
test-tubes and there find that some of the original F, genes remain
unchanged—the unchanged germ-plasm which is left behind—and
to lift out well-defined gametes with ‘‘A’’ or ““a”” and “B” or
‘bh’? variously combined, according to the constituent genes of
the zygote, for in the living germ-cell cycle, not the blending of
the soma, but clean-cut segregation is the rule. Representing
the gametes by capsules containing genes in solid form, which
capsules and genes would be slowly soluble in the substratum of
the zygote, would drive the analogy a little closer to nature.

364 THE AMERICAN NATURALIST [ Vou. LIT

Fy Female Gametes

| Care ©)

23 AB XXA€ xxaB LLAC-

PRAS
OJO
©
® 1

AABB AABE AQBB AQBE

AABE AAEE AaBE AQC

AQBB AQBE- aaB aaBpe

Aa Be AGEs aaBe aate

Male Gametes

XXaQOB

Fz Zygotes:Phenolypic Ratio 15:1
Mustrative we Experiment: _

Sent Bo Wate Wee ada a A E
Gene B= Dilute Hz SOg (Colorless).

REFERENCES
Johannsen, W. Elements der Exakten Erblichkeitslebre, pp. 526-8, 1913.
Laughlin, H. H. The F, Blend Accompanied by Genie Purity. AMER.
Nar., pp. 741-751, 1915.
Shull, G. H. A Simple Chemical Device to Illustrate Mendelian Inherit-
ance. Plant World, Vol. 12, pp. 145-153, 1909.
CoLD SPRING HARBOR H. H. LAUGHLIN.

THE FACTORS FOR YELLOW IN MICE AND NOTCH IN
DROSOPHILA

IN a recent number of THE AMERICAN NATURALIST! appeared a
paper by Ibsen and Steigleder on ‘‘Evidence for the Death in
Utero of the Homozygous Yellow Mouse.’’ In summing up (p.
751), after stating that ‘‘our evidence tends to confirm the con-
elusion of Castle and Little that in mice homozygous yellow zy-
gotes are produced in the yellow X yellow mating, but that these

1 Vol. LI, No. 612, December, 1917, pp. 740-752.

Nos. 618-619] SHORTER ARTICLES AND DISCUSSION 365

zygotes fail to develop normally after implantation in the
wees ” they suggest “that in mice there may be a ‘lethal fac-
tor,’ similar to those so well known in Drosophila, which is so
closely linked to the factor for yellow that they are practically
at the same locus and there is consequently no crossing-over.””

If the postulated factors are indeed so closely linked that
crossing-over never occurs, would it not be simpler and even more
logical to assume that but a single allelomorphic pair of factors
is involved? Such an assumption necessitates a second, namely,
that one of the factors of this pair is dominant in one of its ex-
pressions (yellow) and recessive in a second (lethal).

Morgan? has recently made a similar assumption for the mu-
tant sex-linked factor in Drosophila which causes ‘‘notch’’ in
the wings. He states that it is ‘‘dominant in regard to the wing
character but recessive in its lethal effect. A female with notch
wings carries the gene in one of her X-chromosomes and the nor-
mal allelomorph in the other X-chromosome. Half of her sons
get the former and die, the other half get the latter X-chromo-
some and live. As there are no lethal bearing males, the females
must in every generation be bred to normal males.”’

If Morgan is right in his assumption, there seems to be no rea-
son, a priori, why factors having both dominant and recessive
expressions should be limited to sex-chromosomes.

Stating that crossing-over has not been observed in a certain
number of generations, even though the number of both individ-
uals and generations is large, is, however, a very different matter
from proving that crossing-over never occurs. Just where the
evidence ceases to be negative and becomes positive, perhaps de-
pends upon the ratio between the space occupied by a factor
and the total length of the chromosome bearing it, in its relation
to the law of probability, and to possible factors restricting
erossing-over.

The data necessary for such a calculation does not appear to
be available for either mice or Drosophila. Which interpretation
is assumed must therefore for the present be a matter of per-
sonal preference. The continued failure to find crossing-over
will in each case tend to make the one involving a single allelo-
morphic pair of factors more probable.
LIAM A. LIPPINCOTT
KANSAS STATE AGRICULTURAL COLLEGE,

January, 1918

2 Amer, NAT., Vol. LI, No. 609, September, 1917, pp. 513-544,

NOTES AND LITERATURE

Genetics in Relation to Agriculture. By E. B. Bascock and
R. E. Cuausen. New York: McGraw-Hill, 1918. Pp. ix-xx
+ 675, Fig. 239, Pl. 4.

Every geneticist who opens this volume in a spirit of hope
and expectation will close it with appreciation and satisfaction,
with a feeling that a tremendous task has been worthily accom-
plished. Unquestionably it will fill a real need as a text-book.
In addition it should be of great service to the general biologist
—using that term in its broadest sense—who wishes to have a
compendium of genetical facts for ready reference.

The work is divided into three parts; it is really three vol-
umes in one. In the first part the fundamental principles of
. genetics are outlined, in the second plant breeding is discussed,
in the third animal breeding is considered.

There are thirty-nine chapters and an excellent glossary. The
bibliography of twenty-four pages is not complete and does not
purport to be complete. It is a list of the actual titles consulted
in the preparation of the manuscript, truly no light undertaking.

The reputation of both authors as careful investigators is so
firmly established that one is not surprised to find the immense
volume of genetic literature of the past eighteeen years judi-
ciously weighed and sifted, but that they found the time during
the selection of material for so much constructive and illumi-
nating criticism is perhaps not to have been anticipated. More-
over, since an investigator is not necessarily a good text-book
writer, one suspects that the general excellence of the book from
the pedagogical standpoint is due largely to Dr. Babeock’s spe-
cial training and long experience as a teacher.

After an introductory chapter on the methods and scope of
genetics, the authors have chosen to open the volume by treat-
ing variation. Several modifications of older classifications of
variation are given which may or may not prove useful. They
were probably included rather tentatively as introductory to
the main business of the chapter, the action of external stimuli
in modifying development and causing germinal variation.
Some of the illustrations here are not all that might be desired,

Nos. 618-619] NOTES AND LITERATURE ` 367

for example, the work of T. H. White on making new varieties
of tomatoes by adding large quantities of fertilizers, which is
rather hesitatingly cited. But perhaps one should eriticize mat-
ter which does not appear rather than matter which does appear.
In other words, if the reviewer may express his own hope in the
connection, it is to see a more extended, well-rounded discussion
of the action of the factors of environment in a later addition.

The statistical study of variation has been given a very pleas-
ing treatment. The mathematician will undoubtedly complain
that it is amateurish because it is given with a minimum of
technical language. But just there lies its value in such a
text-book. The necessary statistical tools are described so
clearly that the elementary student can hardly err in their use.
The more advanced student is properly directed elsewhere.

The physical basis of heredity is also discussed with great
independence. The chromosome hypothesis is accepted without
reservation. Cytological details not bearing directly on the sub-
ject in hand are not even mentioned, while those described—
though accurate as far as they go—are somewhat diagrammatic.
-It will be interesting to see how this treatment works out in the
classroom. No doubt many will find it a welcome change from
the interminable details of unknown significance that often fill
the pages of books on genetics. At the same time it is perhaps
to be regretted that the difference between the higher plants and
the higher animals in gametogenesis, and the bearing on genetics
of the conflict between supporters of pea re and tele-
synapsis, are not given more space.

A hundred and fifty pages are devoted to Mendelian inter-
pretations of breeding facts. A wealth of illustration is used,
and almost all types of Mendelian ratios are represented,
although it is not easy to gather them together in a general
ensemble. The older treatments of Plate and of Goldschmidt,
where every theoretical modification of the Mendelian ratio was
mentioned and then a case in point cited, had their value from
a pedagogical standpoint, and it is rather to be deplored that
their use was not continued. The newer data, in particular the
work of Morgan and his students, is admirably presented, how-
ever. For the first time text-book treatment of crossing over,
interference, and the various other phenomena discovered in

osophila, is given in such a manner that the beginning student
should be able to grasp the essential points without difficulty.

Interesting chapters on species hybridization; pure lines and

368 THE AMERICAN NATURALIST [ Vow. LII

mutations comprise the remainder of part one. The discussion
of mutations will probably not please DeVries, but it certainly
is more in accord with all the facts than most other essays on
the subject.

The applied genetics, animal breeding and plant breeding,
filling twenty-four chapters, ought to be received with great
approbation. Its compilation must have been very difficult,
and its usefulness should be in direct proportion to the work
involved. There is a great deal of new material of a strictly
practical nature such as the origin of sweet pea varieties, breed-
ing disease-resistant plants, etc. On the other hand, there is
really no definite line to be drawn between part one and parts
two and three. Theoretical genetics runs all through the book.
For example one finds the treatment of heterosis, graft-hybrid
chimeras and bud mutations in part two; and in part three dis-
cussions of Mendelian inheritance in domestic animals, acquired
characters, non-Mendelian theories of sex-determination, the hor-
mone theory, ete. The work is so well done that it must be read
to be appreciated. The reviewer has but one suggestion to make
concerning it, and this not in the nature of a criticism. Par-
tially sterile hybrids between species, as is shown in Chapter 12,
supply particularly useful material for the improvement of do-
mestic animals and cultivated plants. If one studies carefully
the history of domestic species, fragmentary as it is, he is as-
tounded at the enormous number of cases in which the ancestry
involves two or more species. Would it not be well to empha-
size this point by extended illustrations in a book on plant and
animal breeding? :

The authors are to be congratulated on having brought to-
gether the material for a classical text-book on genetics. Re-
vision will doubtless soon be necessary as is suggested in the
preface. If it is made with the care that such an excellent
foundation deserves the work will unquestionably go through
many editions.

The McGraw-Hill Company deserve no little credit for the
generous way in which they have seconded the authors’ efforts.
The typography, illustrations, paper and binding are extremely
good. Weunderstand that this is their first contribution toward
a series of agricultural ¢ext-books under the general editorship
of C. V. Piper. A standard has been set that is a good augury
for the future.

. E. M. E.

THE
AMERIC ¿AN NATURALIST

VoL. LII. August-September, 1918 Nos. 620-621

THE RELATION BETWEEN COLOR AND OTHER
CHARACTERS IN CERTAIN AVENA CROSSES!

PROFESSOR H. H. LOVE anv W. T. CRAIG

IN COOPERATION WITH THE OFFICE OF CEREAL INVESTIGATIONS, U. $.
DEPARTMENT OF AGRICULTURE -

SPECIES crosses among oats have not been studied to
any great extent, yet they offer some very interesting
problems. Trabut? says:

Hybridization between the cultivated species of oats has not yet been
methodically attempted to my knowledge, and there is here a very
interesting open field. It is true that we have yet to determine in what
degree a true hybridization will be possible. If Avena fatwa sativa
may be crossed with A. sterilis culta, a progeny may be produced having
very useful mixed characters. A. abyssinica will gain by being crossed
with the really superior A. strigosa. But in the matter of hybridization
there is much more to be gained from experimentation than from the
mere discussion of theoretical views.

Since this paper was read (1911) some studies have
been reported on with different species crosses in oats.
Zade? discusses results obtained from a cross between
fatua and sativa. He found F, to be intermediate and
that the F, gave types resembling fatua, sativa, and the
F, intermediate type. These with respect to awns and
hairs gave a 1:2:1 ratio.

Surface* has described rather fully some results ob-

1 Paper No. 70, ette of Plant Breeding, Cornell University, Ithaca,
N. Y.

2 Journal of Heredity, Vol. 5, pp. 84-85, 1914. ee from the
original article published 1912.)
3 Der Flughofer, Deut. Landw. Gesell. Ab., side ces 1912.
4 Genetics, Vol. 1, No. 3, pp. 252-286, May, 1
369

370 THE AMERICAN NATURALIST [ Vou. LIT

tained from a cross between Avena fatua X Avena sativa
var. Kherson. The two parent forms differ in a number
of characters some of which may be given here. The
Avena fatua is brown or black in color, has both kernels
awned and pubescent and has the typical wild type of
base surrounded by a tuft of basal hairs. The Kherson
is yellow in color, has few or no awns and lacks pubes-
cence on the glumes and has the typical sativa base on
which is found an occasional hair.

The F, type is intermediate between the two parents.
The colic of grain is brown but somewhat lighter than
the wild parent, and the larger kernel in the spikelet is
usually awned,> while the smaller or upper kernel is
never awned. The lower grain exhibits a medium pubes-
cence while the upper grain is always smooth. The base
is intermediate between the two parent forms with tufts
of hair on each side but not all around the base.

Some of the results of the F, generation it be given
in Surface’s conclusion:

The data show that the wild parents carry genes for gray and prob-
ably for yellow color in addition to the black. These three colors seg-
gregate independently of each other. The observed ratio closely approx-
imates the expected and confirms Nilsson-Ehle’s conclusions.

The cultivated base of the grain is dominant to the wild and segre-
gates independently of the color genes. The heterozygous condition in
the lower grain ean be recognized in the majority of plants.

In this cross seven pairs of characters are completely correlated with
the character of the base. The characters associated with the wild base
are (1) heavy awn on the lower grain, (2) awns on the upper grain,
(3) wild base on the upper grain, (4) pubescence on the pedicel on the
lower, and (5) on the upper grain, (6) pubescence on all sides of the
an of the lower grain and (7) pubescence on the base of the upper

The gene for pubescence on the back of the lower grain is partially
linked with the black color factor. The F, generation is too small to
determine the exact degree of linkage but dinates that there are about
0.7 per cent. of crossovers.

5 Surface, on page 258 of Genetics, Vol. 1, No. 3, 1916, says, ““the ma-
jority of the lower grains show a weak, straight awn’’ and on page 265 says
““the majority of F, spikelets show no awn whatever.’’ It i is not elear just

ich i

Nos. 620-621] COLOR IN AVENA CROSSES 371

The gene for pubeseence on the back of the upper grain segregates
independently of the color factor except that in the absence of the gene
for pubescence on the lower grain the gene for pubescence on the upper
is unable to act. In this sense the gene for pubesence on the lower grain
is a basie pubescence factor similar to the color factor (C) found in
many animals and plants.

MATERIAL AND METHODS

The present authors have been working with oat
species erosses for several years. A number of dif-
ferent combinations have been made and studied, includ-
ing several species and many of their derivatives. It is
planned here to emphasize certain results obtained with
a cross between Avena fatua and Avena sativa var.
Sixty Day. This Sixty Day variety is identical with the
Kherson as used by Surface so far as general varietal
characteristics are concerned, yet no doubt there are
many strains of both sorts.

The Avena fatua in appearance was similar to the type
used by Surface and has the characters described above.
The Sixty Day oat is yellow and seldom do any awns
appear. There are no dorsal hairs but an occasional
basal hair may be found.

The plants used in making these crosses were grown
in the greenhouse since a greater number of successful
pollinations can be made than when the plants are grown
in the field. The Avena fatua was used as the female
parent and a number of flowers were emasculated and
pollinated. Three seeds developed and each produced a
plant. The F, generation plants were also grown in the
greenhouse since they may receive greater care and more
seed may be obtained. All later generations were grown
in the field, spacing the plants two to three inches in the
rows.

Discussion oF RESULTS

The F, type was as described by Surface, generally
intermediate in type. The color was brown somewhat '
lighter than the wild type. The large kernel of the
spikelet was often awned and was covered with dorsal

372 THE AMERICAN NATURALIST [ Vou. LIT

hairs. The small kernel of the spikelet was never awned
but had an occasional sprinkling of dorsal hairs. The
base was more like the sativa type, yet appeared to be
more intermediate in type with some basal hairs on either
side of the base but not at the back.

When the seeds from the first generation plants were
sown, a number of different types appeared in the second
generation. There were some that resembled the two
parent forms and also other types different in color,
amount of awning, pubescence, and the like.

As regards color there appeared four’ types, black, —
gray, yellow and white. The white ones, of which there
were only four, were tested later and proved to be gray,
thus leaving only the three color types. The black oats
were of two general types, those having the two strong
awns and pubescence on both kernels and the wild base,
and those having pubescence on the large kernel and
sometimes on the small one and with an intermediate or
sativa type of base. Some of these forms were awnless
and others possessed varying amounts of awn which were
in some cases strong and in others weak.

The gray colored oats were both pubescent and smooth,
some fully awned, some partially awned, and some awn-
less. They also segregated as to type of base.

The yellow oats, however, were all smooth and pos-
sessed very few or no awns. No yellow oats developed
the strong awns similar to the wild type.

The segregation as to color and percentage of awns of
the second generation plants is shown in Table I.

TABLE I
SHOWING THE SEGREGATION AS TO COLOR AND PERCENTAGE OF AWNS. SERIES
687, Avena fatua X Avena sativa VAR. Sixty Day
Percentage of Awns

Color |o | 1-9 10-10! 20-29] 30-39) 40-49 50-59 60-69 70-79 80-89| 90-99 100 Totals
. Black... 46 | 25. 20 | 23 12| 17} 21| 23| 14] 8| 2 | 90| 310
Gry. nS e w| ay el staja 28| 92
Yellow...| 9} 7| Iio | wad 18
a | | i 4 |
Totals -| 66 | a1 | 27| 34| 15| 27| 27 | 28| 15| 11 | 2 |127| 420
! E y 1 | |

Nos. 620-621] COLOR IN AVENA CROSSES 373

The percentage of awns was determined by taking a
typical head from each plant and counting the total num-
ber of spikelets and the number of awned spikelets, and
then determining the percentage of awned spikelets.
Since there are a number of plants having no awns and
also a number having 100 per cent. it seems best to ar-
range the classes as has been done in this table; that is,
separate classes have been made for the 0 and 100 per
cent. values.

From this table it is clear that the black oats possessed
varying degrees of the awned condition from awnless to
fully awned. The grays too showed about the same
range of distribution. The yellow types, however,
showed a tendency to be grouped in the lower classes.
No yellow types were found having more than 30 per
cent. of awns. It seems quite evident that there may be
some relation between the yellow oat and the lack of
_ awning.

Nilsson-Ehle® has discussed a case of a cross in oats
where there was an apparent inhibition of awning pro-
duced by a yellow oat. He concludes that there was some
inhibitory effect of the yellow color or that the yellow
color factor acted also as an inhibitor of awns. The
oats Nilsson-Ehle worked with were domestic types and
it is a well-known fact that the domestic sorts vary as to
the amount of awns with the change in environment.
For that reason it is felt that results obtained from wild
crosses will be more definite since the wild type produces
its fully awned condition under very diverse growth con-
ditions. This is all the more evident from a cross of
this same Sixty Day type used in these crosses with a
black cultivated oat. It was found that the yellow oats
segregating from this cross contained fewer awns than
did the blacks, whites, and grays but that the percentage
of awns on the black parent was so variable that the ru-
sults obtained are not definite. Such is not the case for
the awning of the Avena fatua as stated above.

6 Zeitschrift für induktive Abstammungs- und Vererbungslehre, Bd. XII,
Heft 1, 1914.

374 THE AMERICAN NATURALIST [ Vou. LIT

These results show that the segregation as to color, as
mentioned above, produces three types, black, gray and
yellow. On the assumption that the wild oat carries
genes for black, gray and yellow we would expect a segre-
gation of 12 black : 3 gray : 1 yellow. These figures ap-
proach this result, but there are too few yellows and too
many grays. Instead of the numbers obtained we expect
315.00 blacks : 78.75 grays:26.25 yellows. It is very diffi-
cult to always distinguish between pale grays, yellows
and whites, a fact which is well known to those working
with oats. This is especially true in unfavorable seasons
when oats are likely to weather badly. No doubt this
difficulty is one of the reasons for the deviation of the
gray and yellow classes. When we group the non-blacks
together we have a very fair approximation to the 3:1
ratio, which ratio would be expected on the above assump-
tion.

The relation between color and pubescence for these
same second generation plants is well illustrated by Table
IL.

TABLE II
SHOWING THE RELATION BETWEEN COLOR AND PUBESCENCE OF THE SECOND
GENERATION PLANTS OF SERIES 687, Avena fatua X Avena sativa
VAR. Sixty Day

Pubescence
j !
ad
Color | geo neg | oe ee Smooth | Totals

Biel Leer ce) Gk ke eitan orias | 310
AE OS 2 42 24 92
Vow re O | 18 | 18
Foa o SS USO e eee

Certain interesting facts are brought out by this table.
It is apparent that all of the black oats are pubescent,
some having both kernels pubescent and a larger number
having only the one kernel pubescent. It is very sig-
nificant that there are no smooth black oats. The gray
oats, on the other hand, have a certain number of which
both kernels are pubescent, a larger number with one

Nos. 620-621] COLOR IN AVENA CROSSES 375

kernel pubescent, and still another lot of smooth oats. It
seems that the gray oats segregate as to pubescence on
what may be a 1:2:1 ratio. Regarding the yellow oat it
is very significant that all of them are in the smooth class.
That is, no yellow oats are found which are pubescent.
Certain ones which were found appeared yellow but
which proved on testing to be gray instead of yellow, so
that at present no true yellow oats which are pubescent
have been obtained out of this cross.

From these data it seems without any doubt that yellow
oats in some way or another tend to inhibit the factor for
pubescent. It is also apparent that there is one factor
for pubescence which is linked with the black color fac-
tor. There sems to be another factor for pubescence
which is independent of any color factor and for this
reason we obtain gray oats in approximately the ratio of
1:2:1 so far as pubescence is concerned. Owing to the
inhibiting effect of the yellow oat, there are no pubescent
forms obtained. From this material it is clear then that
we are working with a fatua form which has two factors
for pubescence, one of which is linked with the black
color factor and one which is independent. More will be
said regarding these facts later in this paper, when an-
other cross will be mentioned.

Another interesting relationship is that shown between
the color in the F, generation and the segregation as to
type of base. As mentioned earlier in this paper, the
type of base differs from the wild form which is the typ-
ical sucker-mouth shape while that of the cultivated oat
is of the typical sativa form. The sativa-like form is
dominant or partially so to the wild type. A study of
this and a large number of other crosses in which the
wild type has been used as one of the parents indicates
that the segregation of the base follows the 1:3 ratio,
wild being the recessive type.

The relation between color and type of base for the
second generation plants of this cross is shown in Table

MI

376 THE AMERICAN NATURALIST [ Vou. LII

TABLE III
SHOWING THE SEGREGATION AS TO COLOR AND TYPE OF BASE. SERIES 687.
Aventa fatua X Avena sativa VAR. SIXTY Day

Type of Base
Color Wild | Sativa Totals
Wat ei ee rin n | 97 213 310
PRI a o E A a | 26 92
Yellow .. | 18 18
Wie 123 297 420

From this table it is clear that the black oats segre-
gates into the wild and sativa forms as also do the grays.
On the other hand, the yellow oats exhibit no wild type
of base but are all of the sativa class. It is also appar- .
ent from these data, then, that in addition to the inhibi-
tion of awn production and pubescence there is also some
factor or factors which inhibit the production of the wild
type of base when this particular cross is made. That
1t must be due to some factor or factors related to the
yellow oats is clear from the fact that with a large num-
ber of crosses in which white oats and other forms have
been used, the grays and whites as well as the blacks ex-
hibit the wild type of base in about the ratio that would
be expected.

TABLE IV
SHOWING THE RELATION BETWEEN COLOR AND PERCENTAGE OF AWNS FOR THE
THIRD GENERATION FAMILIES PRODUCED FROM THREE HETE
ZYGOUS PLANTS OF THE SECOND GENERATION, SERIES
687. Avena fatua X Avena sativa VAR.
IxTY Day
Percentage of Awns

Color 0 ¡1-9 10-19 | 20-29 | 30-39 | 40-49 | 50- 59 | 60-69 | 70- 79 80-89 90-99 100 | Totals

Back. PEAS AS a gg lg | 58 | 230

‘Grey. e BD OP alaa 1 18 | 57

Yellow ...| 10 7| 2 | tisi 21
|

Totals. .!104'65! 25/10 | 5 | 6|814{4141/ 76 | 308

_ Seed from a large number of F, plants were tested in
the F, generation but all of these will not be discussed
here. Three of these which exhibited the segregation

Nos. 620-621] COLOR IN AVENA CROSSES 377

similar to that obtained in the F, generation have been
brought together and the results are here shown for the
three classifications made on the second generation.

The relation between color and percentage of awns of
these three families mentioned above is shown in Table
IV.

Here again, it is apparent that the black oats ranged
all the way from awnless to 100 per cent. awns as do the
gray oats. The yellow oats, on the other hand, are
grouped near the lower percentage classes. Two indi-
viduals, however, exhibited about 50 per cent. of awns,
one being in class 40 to 49, one in 50 to 59. It may be
that these will be found to be grays instead of yellows.
However, in general, the tendency is for the yellows to
exhibit only a few awns. This is in accordance with the
results obtained on the second generation and substan-
tiates the conclusion drawn from the study of that ma-
terial. It may be worth while to call attention here to
the segregation as to color which fits the hypothesis more
closely than does the material of the second generation.
The segregation exhibits what is without doubt a 12:3:1
ratio. The observed numbers are 230 black: 57 gray: 21
yellows while the expected numbers are 231.00 blacks :
57.75 grays : 19.25 yellows.

The relation between color and pubescence on these
third generation families is shown by Table V.

TABLE V

SHOWING THE SEGREGATION AS TO COLOR AND PUBESCENCE OF THREE F,
FAMILIES GROWN FROM HETEROZYGOUS F, PLANTS. SERIES 687

Pubescence
Color aoan padron Smooth Totals
A A ee 91 139 — 230
Gray.. i 19 12 -s 57
Yellow — 5 16 21
TE AR 110 156 42 308

Here it is clear that again all the black oats are pubes-
cent while the gray oats fall into the three classes. The

378 THE AMERICAN NATURALIST [ Vou. LIT

segregation of the grays does not follow the 1:2:1 ratio
here but it is possible that some of those classed as non-
pubescent may have a few hairs when examined more
closely. It was found with the second generation ma-
terial that it was necessary to use a lens with certain ones,
especially where there was very little pubescence show-
ing. This has not been done with all those in this table
and, therefore, it is possible on later examination that
some of them may fall into the group of one kernel pubes-
cent. The yellow oats also, instead of all being in the
non-pubescent group, have a few in the class having one
kernel pubescent. It is likely that on later examination
these will be found to be gray oats. This can not be said
at present. In general, it may be said that this segrega-
tion agrees very closely with that of the second genera-
tion with the exception of the five yellow plants which
seem to exhibit some slight amount of pubescence.

The relations between color and type of base for these
same three third generation families is shown in Table

TABLE VI

SHOWING THE RELATION BETWEEN COLOR AND TYPE OF BASE IN THREE THIRD
GENERATION FAMILIES. SERIES 687, Avena fatua X Avena sativa
var. Sixty Day

Type of Base

Color | Wild | Sativa | Totals

Hk a 57 173 230
Gray. A IA 18 39 57
Yellow ze 21 21
Totals in T ee ee

IS AE ELSIE

_ On examining this table it is clear that the segregation
of these third generation families agrees very closely
with that of the second generation material. The black
and gray oats have both wild and sativa bases in appar-
ently a 1:3 ratio. The yellow oats, on the other hand,
have only the sativa base. This material tends to sub-
stantiate the conclusions drawn from the second genera-
tion material, which is to the effect that it does not seem

-

Nos. 620-621] COLOR IN AVENA CROSSES 379

possible to produce a yellow oat from this cross having
the type of base of the wild parent.

Further information may be had regarding pubescence
and color on examining the results on three other third
generation families which have been grouped according
to color and pubescence. The parent plants which pro-
duced these three families were black, pubescent on one
kernel and nearly awnless.

These results are shown in Table VII.

TABLE VII
SHOWING THE RELATION BETWEEN COLOR AND PUBESCENCE FROM THREE
THIRD GENERATION FAMILIES OF SERIES 687, Avena fatua X
Avena sativa VAR, SIXTY DAY

Pubescence
|
Color Qae mnel | Smooth fo Totals
Black nia A 231
Yellow | | 88
O A A Sas o ea 231 | 88 319

The segregation shows that no gray was present and
that the segregation is only for blacks and yellows so
far as color is concerned and follows an approximate 3:1
ratio. In regard to the pubescence it is clear that all of
the black oats are pubescent while all the non-blacks or
yellows are smooth. This material further substantiates
the statement made earlier in this discussion to the effect
that there is a pubescent factor linked with the black oat.

GENERAL DISCUSSION

The foregoing data show that there is a very definite
relation between color of glume and production of awns.
On the black and gray oats awns are produced in varying
amounts while few or no awns are produced on the yellow
oats. Regarding the inheritance of awns, it has been
shown” that the weak awn is inherited on a 1:3 ratio,
the fully awned condition being recessive. The data

“Love, H. H., and Fraser, A. C., ‘‘The Inheritance of the Weak Awn in
Certain Avena Crosses,” AMER. Nar., 51, No. 608, August, 1917.

380. THE AMERICAN NATURALIST [ Vou. LIT

that have already been collected show also that the strong
awn is inherited in the same manner. These data (Tables
I and IV) give 203 fully awned : 525 partially awned or
awnless, giving 2.92:1.12 per 4. We would expect on a
1:3 basis 182:546. The action of the yellow factor is to
reduce the amount of awns on the yellow glumed oats.

It was stated earlier that there was apparently a
pubescence factor linked with the black and also another
pubescence factor which was not linked with any color.
If this were true and there was no inhibitory effect pro-
duced by the yellow oats we would expect to obtain 15
pubescent to 1 non-pubescent form in the second gen-
eration. It may be well to state here that it has been
found by experiment that the wild form used in this cross
was of such a type that it had two factors for pubescence.
We have also found another form which has only one
factor for pubescence. When this form is crossed with
a white oat, all of the non-blacks are smooth, showing
that this form has the pubescence factor which is linked
with the black while the forms having the two factors for
pubescence give both pubescent and smooth non-blacks.
This is well brought out by the data presented in Table
VIII. Here the same white sort, Tartar King, was
erossed with two forms of Avena fatua.

This table is made up of data of the second generation.
It is possible that later experimentation may change the
relationship of the colors particularly so far as the grays
and whites are concerned. In these tables all those not
showing blacks and grays are classed as yellows and
whites. The further study of these has not proceeded
far enough to determine just the relation here. It seems,
without doubt, that we have two types of fatua, one giv-
ing the 15:1 ratio (Series 351a1) and one the 3:1 (Series
351b1). Again in the 3:1 distribution all the non-blacks
are smooth.

At will be of interest here also to state when the type
having one factor for pubescence was crossed with the
Sixty Day type similar to the one used in Series 687 that
all of the non-blacks, both grays and yellows, were

1

Nos. 620-621] COLOR IN AVENA CROSSES 381

TABLE VIII

SHOWING THE SEGREGATION AS TO COLOR AND PUBESCENCE ON A CROSS BE-
TWEEN Avena fatua and Avena sativa, VAR. TARTAR KING

Series 35lal Pubescence
Both Kernels One Kernel |
Color Pubescent Pubescent Smooth | Totals
Pake O 30 74 104
ay eben aes ces 9 19 3 | 31
White and yellow ...... 3 3 5 11
pe alba era ae 42 96 8 | 146
Se |
138 8
Series 351b1 Pubescence
Color Beo y cad Smooth Totais
rary wie aici ata E heen ed 72 144 216
Cae pee ee ee 61 61
White and yellow...... 18 18
TOTAL PE NA 72 144 79 295
216 79

smooth, showing that this form has the pubescence factor
closely linked with the black color gene.

The results as obtained from Series 687 so far as color
and pubescence are concerned may be explained in the fol-
lowing manner. We may assume the Avena fatua to be
represented by BBGGYYPP, that is, possessing the fac-
tors black (B) which also produces pubescence, gray
(G), yellow (Y) and another factor (P) for pubescence.
The Sixty Day oat then may be bbggYYpp. We then
assume Y to inhibit the production of pubescence in the
absence of B or G.

It may also be well to state that the results here found
can also be explained by assuming Y to have the effect
of producing pubescence in the presence of G. We
would then not assume any pubescence factor (P) for
this explanation. This assumption would account for
the results as well as the one chosen. On the other hand,
from data already obtained on other crosses it does not

382 THE AMERICAN NATURALIST [ Vou. LIL

seem that this latter explanation is the one which should
be used.

On the first hypothesis the F, individuals are, there-
fore, BbhGgYYPp forming eight kinds of gametes. From
this assumption then we would expect 48 black pubes-
cent:9 gray pubescent:3 gray smooth:4 yellow smooth.
The ratio of pubescent to non-pubescent would be 57:7.
The observed numbers in the second generation were
378 pubescent : 42 smooth. We would expect 374.06
pubescent : 45.94 smooth. The observed numbers from
the three third generation families gave 266 pubescent :
42 smooth, while we would expect 274.31: 33.69. Consid-
ering the two groups together we have 644 pubescent: 84
smooth, while we would expect 648,375:79.625. We see
that the observed facts agree very well with the theory.

While it is not intended to go into details regarding
the F, generation, it may be said that a number of ob-
served facts tend to substantiate this hypothesis. For
example, we should have some families segregating in
the third generation giving 15 pubescent : 1 smooth.
This we find to be so. Again we would, according to
theory, expect to find some F, families segregating into
blacks and grays where the grays would all be smooth.
This we find also. We would also expect some gray oats
to segregate into 9 pubescent : 7 smooth. This combi-
nation has also been found.

Regarding the base, it might be well to state that in
studying the segregation of this series into the third and
even into the fourth generation as yet no yellow oat has
been found exhibiting the wild type of base. These re-
sults do not agree with those obtained by Surface al-
though in general we might expect them to be similar
since the yellow Sixty Day oat and the Kherson type are
classed by some as the same variety of oat. Yet when
we know that it is possible to obtain different strains out
of a variety, particularly so far as the inheritance is
concerned, it is not surprising that these results should
not agree. Let. us illustrate this by some results we have

Nos. 620-621] COLOR IN AVENA CROSSES 383

obtained from crossing two black oats which by those
studying classification of varieties of oats have been
classed as the same variety and exhibit the same general
botanical characters. When these two forms were
crossed, both being black, the first generation plants were
black but when the second generation was grown it was
found that a segregation was obtained, giving 15 black
to 1 non-black. This point then illustrates the statement
made above that we shall probably obtain different segre-
gations even though we are supposed to be using the same
variety. This is also brought out by the fact that from
the wild form, Avena fatua, we have been able to obtain
different types so far as pubescence is concerned. How
many other types may exist in the wild form of fatua we
do not know, but experiments are underway to deter-
mine whether it is not possible to find other types as re-
gards color and certain other characteristics.

From what is here said we do not intend to convey the
idea that yellow color as found in oats will inhibit the pro-
duction of awns, pubescence and base but mean merely
that the yellow as exhibited in this series does that. In
fact, we know from the crosses we have already studied
where other yellow forms have been used that it is pos-
sible to obtain the yellow pubescent form and yellow ones
with the wild base. Therefore, the statements made here
hold only for the particular cross here reported.

CONCLUSION

The studies here presented show that we have some re-
lation between these yellow oats and the absence of awns,
- pubescence, and the wild base. We also find that there
are two types of pubescence, or better stated, two factors
for pubescence, one of which is linked with black and one
which is independent of any color factor. Owing to the
inhibitory effect, we do not get a definite Mendelian ratio
from these studies. It is also clear that the third gen-
eration material tends to substantiate the conclusion ar-
rived at from the study of the second generation plants.

STUDIES IN PALEOPATHOLOGY.

III. OPISTHOTONUS AND ALLIED PHENOMENA AMONG
Fosst. VERTEBRATES

PROFESSOR ROY L. MOODIE

DEPARTMENT OF ANATOMY, COLLEGE OF MEDICINE, UNIVERSITY
or ILLINOIS

Every student of the fossil vertebrates who is fortunate
enough to collect a number of complete or approximately
complete skeletons of fossil vertebrates is almost sure to
be impressed with the frequency of the peculiar curve to
the backwardly bent neck and the rigid appearance of the
limbs, if these members are preserved in anything like
the position assumed by the animal at death. This atti-
tude of the skeleton is very common in the petrified re-
mains of extinct animals and it is doubtless what is known
to medical men as opisthotonos. Williston! in describ-
ing the remains of Cimoliosaurus Snowii, a long-necked
plesiosaur, from the Cretaceous of Kansas, says:

The specimen comprises the skull and twenty-eight cervical vertebra,
all attached and with their relative positions but little disturbed. They
lie upon the right side, with the usual opisthotonie curve to the neck,
and are all laterally compressed.

The attitude has been noted among many other fossil
vertebrates, but its significance, so far as I am aware, has
never been commented upon.

Many of the beautifully complete skeletons of the smal!
pterodactyls (Fig. 1), Pterodactylus longirostris, P. brevi-
rostris, P. elegans, from the lithographic slate of Aich-
stadt, which were described many years ago by Goldfuss,
Cuvier, Wagner and Soemmering, exhibit a marked opis-
thotonic curve to the neck and a more rigid appearance
to the skeleton as a whole than is common among the
skeletons of these remarkable vertebrates. Pterodactylus

1 Trans. Kansas Acad. Sci., 1890, p. 1.

Nos. 620-621] STUDIES IN PALEOPATHOLOGY 385

longirostris Cuvier has the jaw gaping as if trismus was
not an accompaniment of opisthotonos, such as is usually
- the case in recent times, or else the jaw was secondarily
moved by the action of the water after the dissolution of

Pter zea a micronyx H. v. Meyer from the lithographic slates

of Eichstädt in Bayar: The original is in the paleontological collections at

unich. This specimen shows a typical ade position. One half natural
size. After Bro!

the muscles. Other pterodactyls, such as Pterodactylus
scolopaciceps, P. longicollum, and others described by `
Plieninger? from the Jura of Swabia show no indication
of any spastic distress.

The toothed bird (Fig. 2), Archeopteryx macroura,
from the lithographic slates, commonly figured in the
textbooks of geology, zoology, and paleontology, exhibits
a pronounced opisthotonos, which may be slightly exag-
gerated in all the slender-necked vertebrates having a rela-
tively heavy head. The weight of the head may have
added to the curve, but the position is none the less a
genuine opisthotonos. The skeleton of a small dinosaur,

2 Paleontographica, Bd. LIII, pp. 210-313, 6 Tafeln, 1907.

386 THE AMERICAN NATURALIST [VoL.LII

Fie. 2. Archaeopteryx macroura from the lithographic slates, showing a typical
: opisthotonos, x 4.

about the size of the modern turkey, described and figured
by Hoernes as Compsognathus longipes Wagn. (Fig. 3)
from the lithographic slate of Kelheim, exhibits an un-
usually well-developed opisthotonos, the skull lying far
back over the pelvis.

Probably the most complete representation of opis-
thotonos among fossil vertebrates is that seen in the skele-
ton of the small cursorial dinosaur, Struthiomimus altus

Nos. 620-621] STUDIES IN PALEOPATHOLOGY 387

(Fig. 4) described by Osborn? from the Belly River series.
The skeleton of this interesting dinosaur is mounted in
a panel mount where the skeletal parts are placed ap-
proximately as found (Fig. 4). The attitude is typically
opisthotonos, the jaws exhibiting trismus, with the head

Fig. 3. Compsognathus longipes Wagner, from the slates of Kelheim. The
position | va nis head and tail are characteristic expressions of a tetanic spasm.
After Hoe

thrown sharply back over the sacrum, the tail thrown
na up; the toes strongly contracted, with the
8 Osborn, H. F., 1917, nig ate Someone of Ornitholestes, Struthi-
of the Amer. Mus. Nat. Hist., Vol. 35

Vittles: on EA er.
Pp. 733-771, Pl, XXIV

388 THE AMERICAN NATURALIST [Vow. LIT

phalanges closely appressed. The whole attitude of the
body strongly suggests some severe spastic distress. The
animal may have been a plant feeder and its death and
spastic distress due to feeding on some poisonous plant,
such as to-day causes tetanic spasms in animals. It may
have suffered death from a severe cerebrospinal infection,
but whatever the cause of its death, the attitude of the

Fig. Skeleton of Struthiomimus altus, Genotype specimen, Amer
Mus. 5339. !/æ natural size. In this panel mount the animal is re ooog:
mately as found. The attitude is typically opisthotonos. After Osbo

animal strongly suggests the effect of disease, and in
discussing the history of disease among animals the opis-
thotonie position exhibited by fossil skeletons must be
- considered as indicating a possible diseased condition.
The correlative phenomenon, pleurothotonos, is less
common among the higher vertebrates, but is not uncom-
mon among the fishes. This is evidently the attitude as-
sumed by the skeleton of the plesiosaur, Plesiosaurus
macrocephalus (Fig. 5), collected by Miss Mary Anning
from the Lias of Lyme Regis, England, and figured by
William Buckland in his ‘‘Bridgewater Treatise.’’* It is
improbable that the head of this long-necked plesiosaur
could have been turned into its present attitude by a cur-
rent of water, since a force sufficiently strong to have
moved the heavy head to one side would doubtless have
+ Vol. II, Pl. 19, Fig. 1, 1837.

Nos. 620-621] STUDIES IN PALEOPATHOLOGY 389

disturbed other portions of the body, and there is no evi-
dence of this in the skeleton.

The remarkable specimen of Geosaurus gracilis H. von
Meyer, from the upper Jurassic lithographic slate of
Eichstádt, Bavaria, as described and illustrated by von
Ammon, shows a clearly marked instance of pleurotho-

Fic. 5. Plesiosaurus macrocephalus, a skeleton from the Lias of England, pre-
seryed in a pleurothotonic attitude. After Buckland.

tonos. The body, slightly twisted, is bent into a strong,
uniform arch toward the left, the animal having been pre-
served on its belly.

The fishes often assume at death and are fossilized in
the pleurothotonic attitude. This is clearly indicated in
the fishes from the Solenhofen, Leptolepis sprattiformis,
as figured by Dreverman, Gaudry and others, though
this attitude is also clearly that of fishes attempting to
flop out of the soft mud back into the water. It is nota
necessary sequence that all laterally compressed verte-

390 THE AMERICAN NATURALIST [ Vow. LIT

brates assume the pleurothotonic attitude, since often
the ganoid fishes (Fig. 6), especially, assume the opistho-
tonos. It is true that the majority of fishes which are
preserved in an approximately complete manner exhibit
no trace of either of these attitudes. The great series of
Triassic fishes from Connecticut seldom exhibit indica-

Fic. 6. Acanthodes gracilis F. Roemer, a ganoid fish from the Permian of
Klein-Neundorf, Lower Silesia, showing an opisthotonic position. After Hoernes.
Hoernes.

tions of either of these phenomena. A single specimen of
Catopterus gracilis, of those figured by Eastman,* exhibits
the opisthotonos, and a single one, Ptycholepis marshi,
exhibits pleurothotonos. Of the scores of specimens of
these fishes described by Newberry and Eastman a very
small percentage show any sign of spastic distress.

As a clinical manifestation of great severity, opistho-
tonos and the correlative phenomena, pleurothotonos and
emprosthotonos (episthotonos), have long been well
_ known in human beings as accompanying certain phases
of tetanus, abscesses of the brain, otitis media, hysteria.
cerebrospinal meningitis, strychnine poisoning, and other
afflictions, in which toxins affecting the nervous system
are liberated. In these manifestations the muscles of the
body, the spine and the extremities are strongly flexed.
This characteristic attitude of the spasm has been graph-
ically figured (Fig. 7) by Sir Charles Bell in his ““Anato-
my of the Expressions,” where he says:

I have here given a sketch of the true Opisthotonos, where it is seen
that all the muscles are rigidly contracted, the more powerful flexors
E s Poea 18, State of Connecticut, State Geol. and Natl. Hist. Survey,

Nos. 620-621] STUDIES IN PALEOPATHOLOGY 391

prevailing over the extensors. Were the painter to represent every
circumstance faithfully, the effect might be too painful, and something
must be left to his taste and imagination.

Opisthotonos has also been described by Falls* as occur-
ring in the fetus in utero, the cause for which is still
unknown.

It is a matter of great interest to find these same mani-
festations represented in the fossilized skeletons of
ancient vertebrates. The majority of the attitudes as-
sumed by the fossils may be due to the spasm usually inci-
dent to death, the Todeskampf of the Germans, or to acci-

Fie. 7. Charles Bell's drawing of a man in opisthotonos,

dental shifting after death. Many of the vertebrates
whose skeletons are found in anything like a complete
state of preservation do not show these manifestations.
It is on the whole unusual for fossil vertebrates to show
opisthotonos and much more common in the slender-
necked species. It is possible that the animals whose
skeletons are preserved in the above-mentioned attitudes
had suffered death owing to diseases similar to tetanus,
cerebrospinal meningitis, or similar disturbances.

The skeleton of Mesosaurus brasiliensis from the

6 Surgery, Gynecology and Obstetrics, January, 1917, pp. 65-67.

392 THE AMERICAN NATURALIST [ Vou. LII

Permian of Brazil’ exhibits a slight degree of opisthotonos
(Fig. 8) such as is common in the death struggle of many
modern vertebrates. There can be, I think, little doubt
that many of the opisthotonic attitudes assumed by fossil-
vertebrates are easily explained as a phenomenon accom-

panying the ‘‘Todeskampf,’’ but whether all can be sa

Fie. 8. aa brasiliensis MeGregor from the Permian: of Brazil, showing
slight opisthotonos. After McGregor

explained on this basis is extremely doubtful. It is cer-
tainly not true that all vertebrates exhibit indications ot
such spasms. While complete skeletal remains of fossil
vertebrates are relatively rare, yet there is a sufficient
number preserved which have been described to determine
the relative frequency of these positions.
In the many complete skeletons of fossil reptiles from
the Eocene of France described by L. Lortet® only four,
7 McGregor, J. H., 1908, ‘‘Commissao de Estudos das Mines, ete.,’’ Pl. IV.
8L. Lortet, 1892, ‘‘Les Reptiles du Bassin du Rhéne,’’ Archives du
Museum d’Histoire Naturelle de Eos: Tome V, pp. 3-139, Pl. I-XII.

Nos. 620-621] STUDIES IN PALEOPATHOLOGY 393

Alligatorium Meyeri (PL X), two specimens of Alliga-
torellus Beaumonti (Pl. XI), and Crocodileimus robustus .
(Pl. IX), exhibit any degree of the opisthotonie attitude.
Only one, Pleurosaurus Goldfussi (P1. VIT), exhibits the
pleurothotonos. The majority of the remaining skeletons
figured show no spastic distress whatever. So that while
we may say that these two positions are common they are
rather the unusual than the usual state of affairs.

On the other hand the dinosaurs Struthiomimus altus
and Compsognathus longipes, many specimens of small
pterodactyls and the fossil bird Archeopteryx exhibit
such a marked opisthotonic attitude as to lead one to infer
some cerebral-spinal or other intracranial infection which
would have been easily possible in the poorly protected
brain case of these early vertebrates. It requires but a
glance at the nature of the brain case of the early verte-
brates to see how poorly protected the cerebrospinal
spaces were. Ingress of infecting bacteria may have been
through any of the numerous nerve or vascular foramina,
through the thin cancellous walls separating the brain
case from the sphenoidal sinus, and through the anterior
end of the brain case which was often protected only by a
membranous covering, by cartilage, or by very thin bony
plates.

The possible presence of the infecting bacteria has been
so well established by the investigations of Walcott,’ van
Tieghem and Renault,” that little need be said here
concerning them. Walcott has described and figured bac-
teria from fossilized Pre-cambrian alge of central Mon-
tana, supposedly Micrococcus, the bacteria being ar-
ranged in groupings characteristic of the Staphylococcus
isolated from a case of Pemphigus neonatorum by Falls,

9 Walcott, C. D., 1915, ‘* Discovery of Algonkian Bacteria,’’ Proc. Natl.
Acad. Sci., Vol. 1, p. 258, Figs. 2 and 3; 1914, ‘‘ Pre-Cambrian Algonkian
Algal Flora,”? Smith. Misc. Coll., Vol. 64, No. 2 (Publication 2271).

10 Renault, B., 1900, **Mieroorganismes des combustibles fossiles.’’ Bul-
letin de la Société de 1’Industrie minerale Saint-Etienne, Serie III, 1899,
Tome 13, pp. 865-1161; 14 (1-2), pp. 5-159, 1900, with Atlas, 1898-1899,
Pl. X-XXV, Atlas, 1900-01, Pl. 1-V.

394 THE AMERICAN NATURALIST [ Von. LIT

by whose courtesy a photograph has been used for com-
. parison. ‘Renault has described a great variety of bac-
teria, many of which are apparently similar to the bacteria
of to-day.

In searching for evidences of disease among fossil
vertebrates I have been interested in making the above
comparisons. In the light of the above study it seems
probable that some of the instances of opisthotonos and
pleurothotonos among fossil vertebrates may be due to
acute cerebrospinal infections, the petrified skeletons
exhibiting trismus, rigidity of the limbs, and the peculiar
backward curvature of the vertebral column so common
to-day as clinical manifestations of spastic distress. This
is especially probable in the cases where the skeletons
exhibit such marked opisthotonos and pleurothotonos as
do many of the specimens above referred to. It may
then be said that opisthotonos as seen in the skeletons of
fossil vertebrates indicates disease only in those exag-
gerated cases of spastic distress as is evidenced by the
attitudes assumed by fossil vertebrates, such as the small
dinosaur, Struthiomimus altus, and the bird, Archeop-
teryx macroura. Not all vertebrates preserved in opis-
thotonus were victims of disease, but many of them sug-
gest a strong neuro-toxic condition.

CANCER’S PLACE IN GENERAL BIOLOGY

W. C. MacCARTY, M.D.
Mayo CLINIC, ROCHESTER, MINN.

Tae condition which has been called cancer by the laity
and the medical profession has been studied by the latter
largely from the standpoint of disease. Investigators
have considered its great destructive action, cause, pre-
vention and treatment, all of which study has been stimu-
lated by the urgent necessity of its eradication from the
ills of man, and not in its relation to the known biologic
facts concerning the universal conflict between living
normal cells and their natural enemies. In order to ap-
proach correctly this biologic phase of the condition it
will be necessary to answer the question : What is cancer?

To the pathologist, cancer is a cellular overgrowth
which occurs in some multicellular organism, especially
in man, and which is characterized by its anparently un-
limited proliferation, during which it destroys tissues,
and is fatal eventually to the whole organism. This, in
general, is the conception held by the members of the
medical profession, but to the scientific mind which is
interested in and trained in the fundamental or more
specific factors operating in living nature, it is neither
satisfactory nor sufficient.

An analysis of the condition from such a biologic point
of view necessitates also for its elucidation a study of the
facts relative to the evolution of multicellular organisms
from single cells as units of life. Biologists agree that
the cell is the visible unit of life, and that all cells have
certain fundamental structural and functional character-
istics which are common to all. They further agree that
all multicellular beings evolve by a process of division
or segmentation of a single cell which has been stimu-
lated automatically, or by the process of extrinsie fertili-
zation to such activity.

During the process of segmentation certain dominant
facts present themselves. A fertilized ovum, for exam-

395

396 THE AMERICAN NATURALIST [ Vou. LII

Fic. 1. Some of the differentiated cells (textocytes) of the o body:

iologic Terminology edical Termino

$ os cytes, Red blood corpuscles
2. Lymphocytes (small), ymphocytes (small).

phocytes (large), Lymphocytes en 2
4. Transiti Transitional

Leukocytes, Po ymorphonuclear leukocytes.
6. Eosinocytes, Eosino
T. Mastocytes, Mast
8. Fibrocytes, brous nective tissue cells
9. Rhabdomyocytes, Striated le E
10. Melan ; igme lls.
A 11.. Myx A aoe
12. Cardiomyocytes, art ‘decks cells,
13. Lipocyt Fat de
14. Leiomyocyt mo ells.
15; 19. Adenocytes, Glandular epithel
16, 24. Neurocy Nerve sty neurones)
TE cytes, Bone
ndotheliocytes, en ae cells
` ondrocytes, Cartilage cells

22. Epitheliocytes, Epithelial cells
23. Tendocytes, endon cells
25. Sudorocytes, Sweat popi 4
26. - 4 : Sebaceous
27. Gustocytes, j Taste pan ise a taste bud).

Nos. 620-621] CANCER’S PLACE IN BIOLOGY 397

ple, divides; the cellular divisions divide; these continue
to divide and form eventually, in a definite period, the
millions of cells which constitute the organism. This is
a simple statement of general facts, but coincidentally
with these facts there is an orderly sequence of cellular
changes which seems to be foreordained in the original
fertilized ovum; the cellular progeny does not retain, to
the same degree, all of the structural and functional char-
acteristics of the original cell (ovocyte). There is a
grouping of cells which is coincident with morphological

Fic. 2. In the embryonic evolution of adult tissues there are certain arbi-
trary stáges of differentiation in which the cells may be given certain names.

During segmentation of the fertilized ovum (ovocyte) e daughter cells do
not show any special morphologic characteristics of adult t =p) but are never-
theless sera ers of such tissues and may be called taal

The textocytes, or tissue cells, are represented in this a by symbols
derived soli the char ainia outlines of the cells of specific tissues (Fig. 1).

After the prot ¿cule s align themselves into the positions of subsequent
tissues they become the Ea e forebearers of the tissues and may be calle
textoblasts. These cells develop by differentiation and specialization into the

t life. S of

tissues (textocytes) of embryonic and p e. Some the cells remain undif-
ferentiated (textoblasts) in adult life S ions the reserve or aripa cells
for specific tissues when the latter are a ed.

and functional differences. Out of such differentiation
and specialization of cells, types of cells arise, groups of
which constitute what are called tissues (adenotex, chon-
drotex, endotheliotex, epitheliotex, erythrotex, fibrotex,
etc.) (Fig. 1). Two or more of the different tissues be-
come grouped to form organs (tongue, esophagus, stom-
ach, liver, kidneys, skin, ete.) which likewise are group

to build up structural and functional systems (respira-
tory, alimentary, nervous, osseous, etc.), the combined
qualities of which form the complete multicellular organ-

398 THE AMERICAN NATURALIST [ Vou. LI

ism or being. Such orderly evolutional facts apply to
the development of all animal and vegetable multicellu-
lar beings (Fig. 2). 5

Out of the essential living properties of a single cell
other cells develop which have in them an exaggeration
of some essential initial quality, each tissue representing
an exaggeration of some one quality. Such evolutionary
cytologic organization produces a communism of living
units, the combined apparent and dominant purpose of
which is to live and reproduce its kind.

A biologist, if asked the ultimate purpose of life, would
shake his head and say he did not know, but asked the
immediate and dominant purpose, would say the protec-
tion of life and this protection even at the expense of life,
an apparent contradiction which has been recognized but
not comprehended even by scientists. The evidence of
this great protective purpose of living matter is too uni-
versal to be called to the attention of the least observing;
it works automatically and in a large degree independ-
ently of the will of living beings.

This fundamental vital protective purpose forms the
basis of the following consideration of cancer’s place in
general biology: It presupposes, and observation sub-
stantiates it, that all living cells have natural antagonists
against which protection is necessary; and that there is
a conflict in nature during which there is constant build-
ing up and tearing down of things living. The human
body is no exception to this rule, as every physician and
layman knows. The tearing down is called disease, and
the rebuilding is called regeneration, repair and healing.

The partial destruction of a human tissue by animate
or inanimate antagonists may be followed by its regen-
eration; the complete destruction may be followed by
repair or replacement, but not by its regeneration! The

1 This statement may not seem to be true when applied to some of the
lower forms of life such as the earthworm (Allolobophora, fetida) and the

such. because the normal regenerative power resides in other cells,
which by the process of metaplasia build up tissues they do not form in the
normal sequence of evolution.

Nos. 620-621] CANCER'S PLACE IN BIOLOGY 399

degree of regeneration of tissues, according to the obser-
vations of biologists, is in an inverse ratio to the degree
of their specialization and differentiation. One finds,
therefore, this regenerative factor a variable and un-
equal quantity among tissues of the human body. The
protective tissue cells (epitheliocytes) of the skin, for
example, are readily regenerated if not completely de-
stroyed over a large area; the cells of the retina, in all
probability, are never regenerated even after partial
destruction. Fibrocytic, erythrocytic, epitheliocytic and
leucocytic tissues, in all probability, represent types, the
special functions of which show the highest degrees of
regeneration.

In many tissues of the body, coincidentally to normal
communistic activity, there is constant or periodic normal
destruction with constant or periodic regeneration, both
of which depend on communistic functional activity and
a constant or periodic destructive action of antagonistic
agents. The amount of regeneration depends on the
amount of destruction, which depends on the quality,
quantity and duration of action of the destructive agent
or agents.

Tissue destruction and regeneration were made the
subject of investigation by the writer, in the protective
cells of the human skin (Fig. 3) and in the secretory
epithelium of the human mammary gland (Fig. 4). One
finds in these organs that some unknown irritant or irri-
tants of an apparent low degree of virulence, acting over
a prolonged period of time, produce certain reactive cel-
lular phenomena; there is first a destruction of the spe-
cialized and differentiated cells (textocytes). This de-
struction is associated with an hypertrophy of the so-
called basal cells (cells of the stratum germinativum, or
textoblasts) and a lymphocytic infiltration in the sup-
porting stroma.

Space does not permit a consideration of the factor of |
lymphocytic infiltration. The hypertrophy of the so-
called basal cells, however, is of great importance from
the standpoint of the subject under consideration.

One sees clearly that nature, in building up the special-

Fic. 3. Three diagram-
matic histologic stages of re-
ction of tex

ion
epithelium showin
tive i

he hyperplastic undif-
ferentiated textoblasts,

THE AMERICAN NATURALIST

[ Vou. LIT

ized and differentiated protec-
tive cells of the skin and the se-
eretory cells of the mammary
gland, has also made provision
for an anticipated destruction,
an anticipation which is no more
remarkable in nature than that
of the butterfly which deposits
its eggs in a safe place and dies
with the inherent assuredness
that the eggs will some day de-
velop into caterpillars and event-
ually into butterflies to continue
the existence of the kind. In the
ease of the cells of the skin and
the mammary gland, if the irri-
tant is removed, complete regen-
eration of differentiated or spe-
cialized cells takes place, pro-
vided the basal cells themselves
have not been completely de-
stroyed.

Continuance of the action of
the destructive agent or agents
produces hypertrophy, hyper-
plasia and migration of the basal
or regenerative cells (Figs 3 and
4). Coincidentally with such a
hyperplasia the basal cells (text-
oblasts) do not always become
differentiated to the form of the

‘specialized squamous or secre-

ory cells according to their com-
munistic normal foreordination;

they retain their oval or spheroidal form, become larger
and produce a massed overgrowth of undifferentiated
cells (Figs. 3 and 4). The degree of hyperplasia and
migration varies under different and perhaps the same
irritative circumstances, depending on inherited variable

Nos. 620-621] CANCER’S PLACE IN BIOLOGY 401

factors in the basal cells the neighboring tissue cells,
their food supply, natural drainage, and perhaps some
unknown factors. The significant biologic facts rest

4 Mf, 4 s x a = >" Y aa
H A W
h yl = SR ¢ x
A9 N T. AANE a N op

Fic, 4. Diagrammatic relation of the glandular units to the other tissues.

in the attempted cellular regeneration by hypertrophy
and hyperplasia and the effort to change environment by
migration, all of which may be seen not only in the breast
and skin, but also in the specific cells of the hair follicle,
prostatic gland and stomach (Figs. 5, 6 and 7).

A change of environment through overgrowth or mi-
gration often stimulates or allows an attempt at differ-
entiation into the specific tissue-cells for which the origi-
nal reserve or regenerative cells (textoblasts) were
apparently foreordained in the normal evolution of tis-
sues. This is evident in cancerous new growths which
have migrated into other tissues, and in regional lym-
phatic glands, which are the favorite locations of envi-
ronmental change for such migrants.

Cellular regenerative reaction takes place in one or
both of two ways; there is hyperplasia with or without

402 THE AMERICAN NATURALIST [ Vou. LIT

differentiation. An hyperplasia with differentiation into
specific tissue-cells may be called texto-typic in contra-
distinction to that without differentiation, which may be

IG. 5. The reaction to destruction of textocytes in the hair follicle. cd

pasa pilo-cytoplasia. B. Secondary pilo-cytoplasia. ©, Tertiary pilo-cyto- .
. 6. The reaction to destruction of textocytes in the prostatic acinus.

A. primary adeno-cytoplasia. B. Secondary adeno-cytoplasia. ©. Tertiary adeno-

cytoplasia

termed cytotypic. This occurs before and after mi-

gration.

According to writers on the subject of cancer the prin-
cipal criterion for the denotation of a condition by this
term consists of the destructive migration of tissue-cells.
Such neoplasms have been often considered to be direct
derivatives of tissue-cells, because their cells sometimes
resemble those of specific tissues of organs from which
they have arisen. As a matter of fact the malignant
neoplastic cells (neocytes) do sometimes resemble the

Nos. 620-621] CANCERS PLACE IN BIOLOGY 403

A
Y

Ce

23000000

Pepere

S

=

Primary Cytoplasia
rg. o The E

condary Cytoplasia sia
eaction to destruction of textocytes in the gastric tubules
mary adeno-cytoplasia. B. Secondary adeno-cytoplasia. C. Terti
adeno-cytoplasia.

Tertiary Cytopla:

ary

specific tissue-cells of the site of origin, but this is not
evidence that the tissue cells themselves have been con-
verted into the cells, which at best never are morpho-
logically and functionally identical with the original
tissue-cells.

According to the observations of the writer, the cells

which constitute cancer are the progeny of the partially
differentiated or reserve cells (textoblasts) which have
for their natural communistic function the protective
restoration of the specific tissues when the latter have
been destroyed. It may be asked, how can a condition
which will certainly destroy the whole organism be the
result of a protective principle? This perfectly natural
question can only be answered by stating a general prin-
ciple in biology, namely, that regenerative changes do
not always consider the communistic adaptation of the
whole organism. It is a manifestation of a principle
which is inherent in cells, cytologic life being primary,
and tissue or organic life secondary. Thus, the plana-
rian in response to certain stimuli produces a new head
when it already possesses one; the actinian produces a
new mouth on the side of its body under certain regen-
erative conditions. Protective migration of animals as
a result of food famine leads to their complete destruc-
tion not infrequently.

404 THE AMERICAN NATURALIST [ Von. LI

In the case of the human being there is no more fitting
example of a fatal, protective communistic action of cells
than that which occurs when a human being obtains a
severe and destructive burn about the mouth or in the
esophagus. Under such circumstances the fibroblasts in
the region become hypertrophic, hyperplastic, differen-
tiated and specialized into dense, contracting scar-tissue,
which, if the destruction has been great enough, may, as
a communistic, regenerative, protective process, inherent
in the fibroblasts, completely close the orifice of the mouth
or esophagus, the result of which is starvation and de-
struction of the whole organism. The fibroblast’s evo-
lutionary duty in the communism is that of replacing
losses of other tissues, and the duty is performed in this
incidence at the expense of its own life and the life of
the organism. Thus, it may be seen that communistic
life is secondary to the life of the cells even in such a
wonderful and complex organization as the human body.

The hyperplasia or neoplasia does not even have to be
migratory from a cellular standpoint to destroy the
whole organism and thereby be clinically malignant; a
term which has been utilized by the medical profession
largely to differentiate cancer from other neoplastic con-
ditions which are generally conceived of as benign.
Thus, a fibroid tumor of the uterus may be clinically ma-
lignant and still not show cytologie signs of the malig-
nancy so characteristic of cancer.

Biologically speaking, protection may be divided into
types—cytotypic, textotypic, organotypic, systemotypic,
organismotypic, familiotypic, raciotypic and speciotypic,
ete. Cancer represents the cytotypic protection which
is of primary importance in all protection of living pro-
toplasm.

From a biological standpoint the three reactions of
regenerative cells of tissues to antagonistic influences are
hypertrophy and hyperplasia with differentiation, hy-
perplasia without differentiation, and hyperplasia with
migration, with or without partial differentiation. These
three conditions have been termed eytoplasias (condi-

Nos. 620-621] CANCER’S PLACE IN BIOLOGY 405

- tions of cells) and have been numerically classified by

the writer as primary (restauro-), secondary (expando-),
and tertiary (migro-) cytoplasia. This classification is
applicable to the regenerative cells of epithelial tissue of
the mammary gland, prostatic gland, skin, hair follicle,
stomach, fibrous connective tissue, erythrocytic tissue
(red blood corpuscles) and lymphocytic tissue. These
represent eight tissues out of possible nineteen or more
known specific tissues in the human body. Doubtless
there are other specific tissues in the human economy and
perhaps some of those already mentioned may be even-
tually divided into other specific tissues. In all of the
following tissues, adenotex, cardiomyotex, chondrotex,
endotheliotex, epitheliotex, erythrotex, fibrotex, leucotex,
leiomyotex, lipotex, lymphotex,, melanotex, myxotex,
neurotex, osteotex, pilotex, rhabdomyotex and tendotex,
with the exception of the neurotex and perhaps the myxo-
tex, the fact has been demonstrated that all are regener-
ated after loss, the degree of THE seen varying con-
siderably in the human body.
The following classification of the three biologic reac-
tive phenomena which take place in the regenerative cells
of tissues may be made:
adeno-
cardiomyo-
chondro-

| endothelio-
epithelio-

Primary (restauro-) $ :
Secondary (expando-) } leiomyo- | eytoplasia.
Tertiary (migro-) lympho-

p -
rhabdomyo-
tendo-

ete..

406 THE AMERICAN NATURALIST [ Vou. LIT

Such a nomenclature of known reactive facts has
served the writer as a convenient, simple and practical,
biologic, histologic and clinical terminology. The writer
does not mean that these terms should be utilized to des-
ignate neoplasms. They apply only to the tissue-reac-
tion, which is, after all, the essential thing to be consid-
ered. Animal and vegetable neoplasms represent only
phases of such reaction.

From a clinical standpoint, a hypertrophy and hyper-
plasia with complete tissue differentiation (restauro-
eytoplasia) represents tissue regeneration, which is a
benign condition, since it is normally reconstructive from
a communistic standpoint instead of destructive. A
hyperplasia without tissue differentiation (expando-
cytoplasia) represents a condition of the cells in which
no one can foretell whether the cells will become differ-
entiated into tissues, and thereby be constructive, or mi-
grate and become destructive.

Such a condition is, therefore, in the presence of sci-
ence, a questionable condition. Its benignancy or malig-
nancy, in so far as the organization is concerned, with
our present knowledge, can not be forecast. The proba-
bility of possible migration may be suspected from the
frequent morphologic identity of these undifferentiated
cells to the migratory cells of a known malignant or can-
cerous condition, the only difference being their location.

Biologically considered, primary cytoplasia represents
a tissue regenerative condition, the secondary cytoplasia
represents a neoplastic condition, and tertiary cytoplasia
represents a neoplastic migration to regions foreign to
the cells in question. The whole field of tissue replace-
ment, tissue regeneration and benign and malignant (can-
cerous) neoplasmata (new growths) is comprehended in
these three groups.

The following diagram represents the relation of ma-
lignant (cancerous) and so-called benign neoplasms to
the evolution and organization of the human body:

Nos. 620-621] CANCER’S PLACE IN BIOLOGY 407

Fertilization (Ovoeyte and spermatocyte).
|
Segmentation (Protexto-blasts).

Pro-ditferentiation? (Texto-blasts).

Differentiation dom Acacia neocytes)
(textocytes). (Benign? or malignant?).
Differentiation (textocytes) (Migrat
(Benign nopploam ms). Bi merg neoplasms).
Incomplete differentiation Undifferentiation (neocytes
(Less malignant) (Pseudo-textocytes). (More malignant).

Briefly, in conclusion, the writer makes the following
generalizations from his experience in his studies of can-
cer’s place in general biology:

1. All multicellular organisms represent communisms of cells which have
divided their labors and become specialized and differentiated to form

tissues.
2. Nature has provided for the regeneration of most, if not all, tissues
a

3. In many animal an
to aR Oa in three degrees, i. ¢., hypertrophy, hyperplasia and
atio:

: hyperplasia with or without migration the cells sometimes at-
tempt to aen tiate.

5. Limited hyperplasia with complete apaa produces tissue re-
placement. Unlimited hyperplasia without complete differentiation pro-
gnan neoplasms

E nu ited pb he of regenerative cells of
tissues plus migration without complete differentiation
7. neration (hyperplasia) without &ierentintin’ is a eytotypic pro-
tective process.

8. Regeneration (hyperplasia) with differentiation is a textotypic pro-
tective proces

9. Cancer is a nee instead of a ae abi renn process.

i es sometimes fatal to the

11. All of the reactions may be designated Wa a simple eaa termi-
nology which standardizes clinical, histologie and biologie facts.
REFERENCES
MacCarty, W. C., and Willis, Ba
1911. Carcinoma of the Breast. Old Dominion Jour. Med. and Surg.,
XII, 189-198,
2 The stage of pro-differentiation exists prenatally and postnatally and
may be ished terminologically by calling the prenatal cells of-
segmentation protextoblasts and the regenerative postnatal cel ells textoblasts.

408 THE AMERICAN NATURALIST . [Vou.LII

MacCarty, W. C., and Blackford, J. M.
1912. Yavolichoat of Regional Lymphatic Glands in Carcinoma of the
‘Sto . Ann. Surg., a 811-843.
T W. C., and Sistrunk, W.
913. Dara pe Malignant cade Cysts. Surg., Gynec. and Obst.,
XVII, 41-50,

MacCarty, W. C., and Broders, A. C.

1914. ced => Ulcer and its ea to Gastric Carcinoma.
Arch, Int. Med., XIII, 208-22.

MacCaty WwW. C

913. The eninge of Cancer of the Breast and pe voga Sig-
Surg., Gynec. and Obst., XVI, 441-4,

1914. Pare Suggestions Based on a Study of econ Secondary
(Carcinoma?) and Tertiary or Migratory (Carcinoma) Epi-
thelial ibi of the Breast. Surg., Gynec. and Obst.,

289.

XVIII
1913. The Histogenesis of Carcinoma in Ovarian Simple Cysts and
ystadenoma. Collected Papers, Mayo Clinic, V, 380-390.
MacCarty, W. C., and. McGrath, B. F.
1914

- The Frequency of Carcinoma of the Appendix. Ann. Surg.,
I 5-678,

MacCarty, W. C.
1915. The Biological Mini of the Carcinoma-cell. Pan-Am. Surg.
and

Med. Jou
1914-15, _Proeaneerons Conditions. Jour. Iowa State Med. Soc., IV,
1915. ne ibi of Cancer of the Stomach. Am. Jour. Med.
9476.

1914, Note on the ‘Beguiacity and Similarity of Cancer Cells, Col-
ected Papers, Mayo Clinic, VI, 600-602.
1915. No Facts about Cancer o their Clinical Significance. Surg.,
Gynec. and ee XXI,

Irwin, H. C., and MacCarty, W. C.

1915. Papiloma of the dde Ann, Surg., LXI, 725-729.
MacCarty, W.

1915. Evolution of Cancer. Collected Papers, Mayo Clinie, VII, 903-

Broders, A. o a MacCarty, W. C
A 1916. Me ladosplihidlisine. A réport of 70 cases, Surg., Gynec. and
st., XXIII, 28-32.

MacCarty, W. C.
1915. ae versus Speculation in the Professional Conception of
ncer. Teras State Jou r. Med., XI, 165-1
1916. A New ca of Neoplasms. and it. Clinical Value. Am.
06.

1916. The Relation between Chronic Mastitis and nigan of the
Breast. St. Paul Med. Jour., XVIII, 1 167.

x)

A SURVEY OF THE HAWAIIAN CORAL REEFS

VAUGHAN MacCAUGHEY

PROFESSOR oF Botany, COLLEGE or Hawan, HoxoLuLu, HAWAN

During a residence of nearly ten years in the Hawaiian
Archipelago the writer has had opportunity of visiting
and exploring many of the coral reefs, and has been much
interested in thetr formation, flora, and fauna. The pres-
ent paper aims to present the salient and significant facts
relating to the natural history of these remarkable reefs,
in compact and largely non-technical form. There is a
large scattered literature (inaccessible to the average
reader), dealing with the coral reefs and their life, but
the writer believes this to be the first time that the follow-
ing data have appeared within the confines of a single
paper.

The Hawaiian Archipelago is situated in the midst of
the North Pacific Ocean. It lies between latitudes 18°
34' and 22° 14’ and 154° 48’ and 160° 13’ West Longitude,
being about 2,020 miles southwest of San Francisco. Its
east and west extension is nearly two thousand miles.
the islands are but the apices of a titanic mountain range
that rises to heights of from three to five miles from the
floor of the ocean. l

This long archipelago, comprising about twenty-two
islands, is remarkable for the simplicity of its geologic
formations. Only two classes of rock material are known
in the entire group—lava and coral. There are numerous
subdivisions of these groups (for example, there are many
varieties of lava), but all the known rock-formations give
conclusive evidence of having originated from either one

of two sources—voleanic or coralline. It is extremely

interesting to consider that all of these islands are com-

pounded of two such diverse elements—one from a roar-

ing lake of incandescent lava; the other from the age-long
409

410 THE AMERICAN NATURALIST [ Vou. LIT

labors of coral polyps. A strange ‘‘partnership,’’ with-
out parallel in the annals of natural history.

The islands of the Hawaiian group may be classified on
this basis. The large, high islands of the eastern por-
tion of the archipelago are composed almost wholly of
lava, with small fringing reefs. The low, small islets that
comprise the western extension of the archipelago are
made almost wholly of coral, that i is, in so far as their ex-
posed portions are concerned. The coral formations un-
doubtedly rest upon a voleanic substratum. The group-
ing may be expressed as follows:

I. Large mountainous lava islands, forming a compact
group at eastern end of archipelago; elevations
over 1,000 ft.

A. With well-developed fringing reef:
Niihau, 1,300 ft.; Kauai, 5,250 ft.
Oahu, 4,040 ft.
Molokai, 4,958 ft.; Lanai, 3,400 ft.; Kahoolawe,
1,472 ft.
B. With scanty fringing reef:
Hawaii, 13,825 ft.; Maui, 10,032 ft.
II. Small, low islets, scattered along the western axis of
the archipelago; elevations below 1,000 ft.
C. Eroded volcanic blocks, 120-900 ft., with fringing
coral reef: Nihoa, French Frigates Shoals, Gard-
ner I.
D. Elevated coral islands, 45-55 ft., with fringing
reef: Laysan, Lisianski.
E. Typical coral atolls: Pearl-and-Hermes, Midway,
Ocean.
D. Reefs with visible surf, but no exposed coral:
Maro Reef, Dowsett’s Reef.

The entire series, named in sequence from east to west,
is: Hawaii, Maui, Kahoolawe, Lanai, Molokai, Oahu,
Kauai, Ni TOR Nihoa, Necker, French Frigates Shoal,
_ Gardner, Dowsett’s Reef, Maro Reef, Laysan, Lisianski,
Porland: Hermes, Midway, Ocean. The two extremes—
Hawaii and Ocean Island—present a contrast of wonder-

Nos. 620-621] THE HAWAIIAN CORAL REEFS 411

ful vividness. Mauna Loa, the greatest active voleano
on the planet, dominates the island of Hawaii. Its colos-
sal dome is crowned by a summit lake of reverberating
liquid lava, with spectacular displays of high-jetting fire
fountains. The bulk of the island is still growing,
through intermittent outpourings of lava. Ocean Island,
on the other hand, is the last white fragment of a subsid-
ing coral-crowned mountain—perhaps a dead volcano,
that may have resembled Loa in many respects, but which
has been drawn inexorably into the abysses of the Pacific.
One represents the culmination of the voleanic forces;
the other the climax of coral work—an atoll on a tropic
sea,
Of the larger eastward islands, Kauai and Oahu are of
particular interest, as they have the largest coral reefs,
and support the most luxuriant marine life. The reefs
are all of the fringing or platform type, and vary in width
from a few hundred feet to half a mile. Reefs are well
developed along the southern or leeward shores of the two
Islands mentioned, and also, to a lesser degree, along the
northern coasts. Oahu is almost encircled by coral reefs,
whereas Kauai, Molokai, and Maui have numerous coastal
stretches wholly free from coral. The little island of
Niihau, to the west of Kauai, has considerable coral reef.
It is significant to note that although the majority of
corals, particularly the more massive reef-building forms,
occur only in the shallow waters of tropic seas, there are
a number of species that inhabit deep, cold waters.
Lophohelia prolifera and Dendrophyllia ramea, for ex-
ample, form dense beds at depths of from 600 to 1,200
ft. off the coasts of Norway, Scotland, and Portugal.
The general requirement, however, is shallow water whose
mean temperature does not fall below 68° F., and the
reef-building species do not flourish unless the tempera-
ture is considerably higher. Although a single Hawaiian
species of mushroom-like coral (Bathyactes Hawaitensis)
was dredged by the Albatross from a depth of nearly
7,000 ft., most of the Hawaiian forms live in waters of

412 THE AMERICAN NATURALIST [ Vou. LII

6-150 ft. depth. Of the 34 Hawaiian genera, 14 habituate

this shallow-water zone throughout the archipelago, and

10 of these oceur on the leeward reefs of Oahu between

Leahi and Pearl Harbor.

Dana’s! comprehensive statement concerning the reef-
building corals may be compactly summarized. He
states that it is important to have a correct apprehension
of what are those reef species as distinct from those of
colder and deeper seas. The coral-reef species of corals
are the following :

In the Astrea tribe, all the many known species.

In the Fungia tribe, almost all the known species.

In the Oculina tribe, all of the Orbicellids; part of the
Oculinids and Stylasterids; some of the Caryophyl-
lids, Astrangids, and Stylophorids; all of the Pocillo-
porids.

4. In the Madrepora tribe, all of the Madreporids and

Poritids; many of the Dendrophyllia family.

5. Among Alcyonoids, numerous species of the Aleyontum

and Gorgonia tribes and some of the Pennatulacea.

6. Among Hydroids, the Millepores and Heliopores.

7. Among Alge, many Nullipores and Corallines.

He further states that
Through the torrid region, in the central and western Pacific, that is,

within 15° to 18° of the equator, where the temperature of the surface

is never below 74° F. for any month of the year, all the prominent
genera of reef-forming species are abundantly represented. The Ha-
waiian Islands . . . are outside of the torrid zone of oceanic tempera-
ture, in the subtorrid, and the eorals are consequently less Juxuriant and
much fewer in species. There are no Madrepores, and but few of the

Astraea and Fungia tribes; while there is a profusion of the corals of

the hardier genera, Porites and Pocillipore.

The more abundant reef builders, at moderate depths,
are the madrepores, astreids, porites and meandrines.
At depths of 90-120 ft. the millepores and seriatopores
predominate. The great field of coral development thus
lies between low water and 120 ft.

Dana’s classification of reef-formations is useful in sur-
veying the Hawaiian reefs:

1 James D. Dana, ‘‘Corals and Coral Islands,’ 1872.

cl de el

Nos. 620-621] THE HAWAIIAN CORAL REEFS 413

1. Outer reefs, or reefs formed from the growth of
corals exposed to the open seas. Of this character are all
proper barrier reefs, and such fringing reefs as are un-
protected by a barrier. All of the larger Hawaiian reefs
are of this character.

2 Y < 3

Nautical Miles
Fic, 1. Midway Island. A nearly complete circular coral atoll, about 16
miles in circumference; area of lagoon about 38 sq. miles; several low sand
islets in the lagoon

2. Inner reefs, or reefs formed in quiet waters between
a barrier and the shores of an island. The reefs of this
type are very rare in Hawaii; usually they are mere hum-
mocks in the lagoon of the fringing reef.

Kaneohe Bay and Pearl Harbor, on Oahu, are essentially
large drowned valley regions, converted by subsidence
into land-locked bays, which have become more or less
completely barred and filled by coral growths. Were
there not such large quantities of fresh, mud-laden water
poured into these bays, they would be veritable coral
wonderlands, for it is in protected waters of inner chan-
nels or lagoons that corals attain their finest develop-
ment, and the Parhon. so are presented to the: ex-
plorer of coral scenery.’

The marine flora and fauna in these bays presents

a

414 THE AMERICAN NATURALIST [ Von. LIT

many contrasts with those of the true lagoons and outer
reef rims. All of the pure-sea-water-requiring organisms
are wholly absent or rare, and in their places one finds a
large series of brackish water and silt-loving forms. The
generalization is quite accurate for the Kaneohe and Pearl
Harbor inner reefs that

The main distinction between the inner and outer reefs consists in the
less fragmentary character of the rocks in the former ease, the less fre-
quent accumulations of débris on their upper surface, and the more
varied features and slopes of the margins. . . . There is to be found
about inner reefs, over large areas, solid white limestone, showing
internally no evidence of its coral origin, and containing rarely a shell
or other imbedded fossil. It is a result of the consolidation of the fine
coral sand or mud that is made and accumulated through the action of
the light waves that work over the inner reefs. Other portions of reef .
consist of branching corals, with the intervals filled in by sand and
small fragments; for even in the stiller waters fragments are to some
extent produced. A rock of this kind is often used for buildings and
walls on the island of Oahu. It consists mainly of Porites, and in
many parts is still cavernous, or but imperfectly cemented.?

3. Channels or seas within barriers, which may receive
detritus either from the reefs, or from the shores, or from
both these sources combined. These channels correspond
to the lagoons of the fringing reef, except that the chan-
nels are much larger. The Hawaiian lagoons are gen-
erally floored with coral sand, indicating that reef erosion
is more rapid than coast erosion.

4. Beaches and beach formations, produced by coral
accumulations on the Shore through the action of the sea
and winds. Beaches and dunes of coral sand are com-
mon on the islands of Molokai, Oahu, Kauai, Laysan, Mid-
way, Ocean, etc.

Of the three great classes of coral reefs—fringing, bar-
rier, and atoll—the first and last only have representation
in the Hawaiian Archipelago. The fringing reefs are
platforms of coral limestone which extend but a relatively
short distance from the shore. The seaward edge of the
platform is characteristically somewhat higher than the
inner portion, and is usually awash at low tide. The reef

2 Dana, loc. cit.

Nos. 620-621] THE HAWAIIAN CORAL REEFS 415

_ is cut by more or less numerous channels, which mark
those places where streams flow down from the land.
There is usually a lagoon—of sufficient depth to be navi-
gable by canoes or small boats—between the reef rim and
the shore. The outer wall of the fringing reef is steep,
and in the Hawaiian Islands descends abruptly into deep
water. The reef rim is the region of most active coral
growth, the shoreward coral being gradually killed by
fresh water and the deposition of mud and sand.

Pe et O MR y
| pe o SA
sl .
ib £76 SS cos
e — TED E N
” x S
je $
- .”
a ra oe
ee asi eer A (Gz
i ai St (ES Se
q A,
E A
Ed
| f A J
`
uw
ee ER, N
MO À
~Y a £ 1
Ñ A A
N ` aD ese El
Y ae ia e aah pee at
2m"
176108" 176100" a mie

Fic. 2. Pearl and Hermes Reef. An irregular, oval coral atoll, 42 miles
circumference; area of lagoon about 80 sq. miles; numerous low sand islets
in the lagoon. The soundings (2) are in fathoms.

Barrier reefs may be considered as fringing reefs upon
a large scale. Although rare in the North Pacific Ocean,
there are many fine examples in the South Pacific. The
grandest in the world is the Great Australian barrier reef,
which is 1,250 miles long, and supports a wonderfully
rich marine life.

An atoll is an annular or ring-shaped reef, either awash
at low tide or surmounted by several islets, or less fre-
quently by a complete circle of dry land surrounding a
central lagoon. The outer wall of the atoll generally de-
scends with a very steep but irregular slope to a depth of

416 THE AMERICAN NATURALIST [ Vou. LII

3,000 ft. or more. The central lagoon is seldom more
than 60 ft. deep, and is often much less. There are
usually one or more navigable passages leading from the
lagoon to the open sea.

The thickness of the Hawaiian reefs is an engaging
subject for speculation. Many of the reefs are undoubt-
edly several thousand feet thick at their seaward margins.
Dana writes:

Could we raise one of these coral-bound islands from the waves, we
should find that the reefs stand upon submarine slopes, like massy struc-
tures of artificial masonry; some forming a broad flat platform or shelf
ranging around the land, and others encircling it like vast ramparts,
perhaps a hundred miles or more in circuit.

The late Dr. S. E. Bishop, of Honolulu, estimated the
depth of the coral at Barber’s Point, Oahu, to be 2,300 ft.

Our first exploration of a Hawaiian coral reef, some
ten years ago, made a lasting impression, so novel and
vivid were those initiatory experiences. The tropic
morning was fine and clear, with the clouds heaped along
the mountains, and the seaward sky flawless. The trade
winds were unusually quiet and the tide was at lowest ebb.
All conditions were most favorable for a detailed exam-
ination of the reef. My comrade and I embarked in a
native outrigger canoe and paddled from the well-known
Waikiki Beach, near Honolulu, to the white surf-lines of
the reef-rim. This is one of the richest portions of the
Oahu fringing reef, from the biological standpoint. We
were clad in bathing suits and provided with suitable col-
lecting apparatus and water-boxes —glass-bottomed boxes
by means of which the sunlit translucent waters are easily
surveyed.

Arriving at a suitable location, a thousand feet from
the shore, where the water was scarcely two feet deep, we
anchored the canoe and prepared for wading. We were
equipped with old shoes to protect our feet from the coral
(which can cause very painful and slow-healing wounds) ;
with broad-rimmed hats to protect eyes, face, and neck
from the intense glare of sun and water; with hammers

Nos. 620-621] THE HAWAIIAN CORAL REEFS 417

for breaking up the coral blocks and for loosening ma-
terial; and with sundry haversacks, wide-mouth bottles,
formalin, etc. For three entrancing hours we wandered `
over the ledges, knolls, and sandy pockets of the reef,
bewildered by the luxuriant diversity of marine life.
Fantastic clumps of living coral, a large number of
strange molluscan species; bright-spotted crabs and other

N
9
'%
8%
¡0% 5
10%
If
H
Ee eee Mi
i Sea mile.

Fie. 3. Laysan Island. An elevated coral island, with a central lagoon.
The soundings are in fathoms, as in all the maps. The dotted line indicates the
reef rim; this also applies to all the maps.

crustaceans in an array of shapes and sizes; colonies of
sea-urchins ; spidery-armed brittle-stars ; exquisitely beau-
tiful hydroid colonies; bizarre-hued holothurians; and
everywhere marine alge of many tints and shapes, repre-
senting a long list of interesting genera. Gorgeously col-
ored fishes, small and large, lurked in the shadowy reef
pools, and evaded prolonged inspection. It is impossible
to describe the profound impression produced by one’s
first sight of the strange and fascinating reef-world.

The coral fauna of the Hawaiian reefs, although not as
rich nor as diversified as those of more tropical waters, is
not to be regarded as scanty. Dr. T. Wayland Vaughan,
who thoroughly investigated the Albatross and Bishop

418 THE AMERICAN NATURALIST [ Vou. LIT

Museum collections, reports 15 families, 34 genera, and
123 species, varieties and forms. As Bryan? states,

Some idea of the richness of the coral fauna of any given locality
can be gathered from the fact that the reef and shallow waters along
the south side of Oahu, but especially at Waikiki, yielded examples of
thirty-four of the species enumerated.

Of the Hawaiian stony corals (Madreporarians) the
genus Porites is the most abundant and is represented by
the largest number of species and varieties. Pocillopora
ranks next in importance, followed by Montipora, Pa-
vonia, Favia, Leptastrea, Cyphastrea, and Fungia. The
last-named genus merits special mention because of the
unique shape of the skeleton, which closely resembles the
inverted head of a fully expanded mushroom, hence the .
‘name mushroom coral. These are solitary, and fairly
common. They are usually found lying flat on the floor
of little pools or pockets along the outer edge of the reef.

The corals, like many other groups of marine or-
ganisms, are remarkable for the variety and brilliancy of
their color during life. Those who know only the bleached
museum specimens have little conception of the living
tints, some of rare delicacy, others of brilliant hue. The
Hawaiian reefs, although they do not show colors as
striking as those of the South Pacific and Indian Oceans,
are not lacking in color, and the ‘‘Coral Gardens’’ are
becoming far-famed as tourist attractions. Pink, yellow,
green, brown, purple and scarlet are represented in many
shades and combinations.

One of the most beautiful of the Hawaiian corals is a
highly precinctive species, Dendrophillia Manni, which is
known only from Kaneohe Bay, on the island of Oahu.
The living coral is a rich deep orange red. There are
numerous short branches, each of which is surmounted by
a single bright orange polyp. When fully expanded the
polyp is about three quarters of an inch long, and re-
sembles a miniature sea-anemone. The polyp can with-
draw completely within its cup. This species is rare,
e Alanson Bryan, ‘‘The Natural History of Hawaii,’’ Honolulu,

Nos. 620-621] THE HAWAIIAN CORAL REEFS 419

occurring only here and there along the margins of the
little coral islands in Kaneohe Bay.

espia bed" N
15% A ema p

ki ‘Ba. a

K ais
16 f NGE to: opt
16
14 $
15%
15%

1%
ze
We

l 2 Bs 4

Nautical miles

. 4. Lisiansky Island. A low, oval island of coral sand, two miles by

three miles; the lagoon empty of water. The surrounding reef extends six or
seven miles from the isle

The famous black coral, Antipathes abies, is absent
from the Hawaiian reefs, although it has a wide distribu-
tion in the Indian and South Pacific Ocean. It grows to
considerable size in the tropical waters of the Great Bar-

* rier Reef of Australia.

The eight-rayed corals (Alcyonaria) are very rare on
the reefs, but occur in fair abundance in the deep offshore
waters. The Albatross collected about 70 new species, 40

` of which were new to science. The Red or Precious Coral
of commerce, Corallium rubrum, does not occur in the
Hawaiian Archipelago. It is most abundant in the Medi-
terranean Sea, although also occurring off the coasts of
Treland and Africa. Related species of slight commer-
cial value have been obtained off Mauritius and near
Japan.

420 THE AMERICAN NATURALIST [ Vou. LIT

The Alcyonacea, organ-pipe and blue corals, are repre-
sented in Hawaii by only five species; the Gorgonacea,
sea-fans, by 16 species; and the Pennatulacea, sea-pens,
by 48 species. Many of the Hawaiian members of these
groups are of great beauty, but are never found in situ
on the reefs, and when rarely washed ashore are badly
mutilated by the waves. Some of the species are phos-
phorescent.

The typical Hawaiian fringing reef exhibits five dis-
tinctive biological zones. This zonation parallels the
shore-line, and is best developed on those reefs which
possess wide lagoons and a well-defined outer margin or
rim.

1. Beach or Inshore Waters.—The shallow inshore
waters, varying in depth from 6 to 36 inches, sustain a
number of the quiet-water alge, such as Enteromorpha
spp., Hypnea nidifica, Gracilaria, Chetomorpha, Ulva,
Chondria, Liagora, ete. The bottom is of coral sand or
mud, more or less contaminated by voleanic soil washed
from the lowlands. The water is often mingled with rela-
tively high percentages of fresh water. The nature of the
bottom depends largely upon the proximity of fresh-
water streams and of the reef-rim. In many places
where the surf is heavy and reef material abundant, the
bottom is pure white coral sand, with practically no rock
or mud. In other districts there are large mud-flats ex-
posed at low tide; the limestone pavement is covered with
a thin sheet of mud, with little sand. Every gradation
may be found between these two extremes. At the mouths
of streams and at numerous other places along the coasts
where fresh-water springs exist below tide-level, the in-
shore water is so fresh as to prohibit the development of
the strictly marine species,

2. Partially Submerged Rocks.—In some places the
_ beach and shallow waters are devoid of rock masses, but
in general one finds partially submerged rocks scattered
all along the coasts. These may be either close inshore,
in the form of ledges or detached fragments, or may lie

Nos. 620-621] THE HAWAIIAN CORAL REEFS 421

at varying distances from the shore. In any case they
distinctly indicate, by their horizontal banding of algal
and hydroid life, the ranges of high and low tide. The
rocks are either of consolidated reef coral or of black
basaltic lava; tufa rocks, and sedimentary coral sand-
stone are infrequent. Some groups of marine organisms
show strong preference for the coral rock, others for the
lava rock. The rocks may be in somewhat protected
situation or may be exposed to the full force of the surf.
The following genera contain alge which are representa-
tive of the kinds that withstand the constant battering of
the waves: Gymnogongrus, Codium, Haliseris, Aspara-
gopsis, Dictyota, Gelidium, Ahnfeldtia, Porphyra. The
controlling factor in the alga-flora of the partially sub-
merged rocks seems to be the circulation of pure, well-
oxygenated sea water. Rocks in stagnant or impure
water support a scanty flora as compared with those of
the surf-swept localities.

3. Pools.—Beyond the rock litter, although sometimes
interspersed by it, lies the zone characterized by numerous
pools or pockets. These cup-like depressions in the
lagoon floor vary in size from little pockets two or three.
feet in depth and diameter to large pools twenty or thirty
feet in depth and diameter. In wading or paddling over
the reef, the pools are easily distinguished by the darker
tint of their waters as contrasted with that of the shallow
lagoon. These pools in the floor of the lagoon are not to
_ be confused with the tidal pools, that lie along the beaches,
and are entirely detached at low tide. The lagoon pools
are inhabited by a great variety of alge and animals that
prefer these shadowy havens to the exposure of the
shallows or the outer reef. The bottom of the pool may
be covered with clear coral sand, or coral débris, or
masses of living coral; its alga-flora will depend upon its
depth and the resultant intensity of illumination.

The following are typical alga genera that have repre-
sentatives in the lagoon pools: Corallina, Peysonnelia,
Grateloupia, Ceramium, Amansia, Polysiphonia, Chon-

LA

422 THE AMERICAN NATURALIST [ Vor. LIT

dria, Laurencia, Martensia, Champia, Wrangelia, Galaz-
aura, Padina, Sphacelaria, Hydroclathrus, ete.

4. The Lagoon.—The entire region between the beach
line or strand and the seaward rim of the reef is properly
designated as the lagoon, but for the purposes of this de-
scription the term will be restricted to the deeper waters,
which are usually located about midway between the beach
and the reef-rim. As one approaches the lagoon wading
becomes impossible, the water deepens to eight, twelve or.
twenty feet, but again becomes shallow as the outer edge

E caii ASA N
Teba ED ea AET ae
4 i z if
4 AN ER N
ies a
505 ise Lar e
e
34 100 ¡64 138 | di
16 M 2 10
se . 42 } >
! oe.
f E
$ al
y1 j x
IT 16161615 ioe F
—— AE
cu Cari ltda
5Sea miles

5. French Frigates Shoal. Crescentic atoll, with numerous low sand
islands, and several high rocky volcanic isles; area of shoal about 30 sq. miles.
The reefs are extensive.

of the reef is reached. The water of the lagoon is placid,
clear, and in normal weather very translucent, so that the
bottom receives good illumination. Although a number
of the smaller alge grow upon the floor of the lagoon, the
region is comparatively barren as compared with the shal-
lower waters on either side. The lagoon floor is a region
of coralline and animal life, rather than of plant life.
The quantities of sand and silt that are constantly washed
over the floor from the disintegrating reef-rim render it
difficult for plants to maintain themselves. The floor is
so irregular in topography that collecting is very difficult;

Nos. 620-621] THE HAWAIIAN CORAL REEFS 423

dredging is almost impossible, and diving is both labori-
ous and unsatisfactory.

5. Reef-Rim.—Upon paddling across the lagoon to the |
outer rim of the reef, one comes to shallow water, where
the heavy combers break and where wading is again pos-
sible. This zone is a favorite fishing-ground of the native
Hawaiians, as it abounds with plant and animal life. The
highest portions of the rim are usually exposed at low
tide; at high tide they are covered by 18-24 inches of
water. There are many table-rocks or shoals, with deep
channelways between. The rim is not regular or sym-
metrical; there are many indentations, crags, débris
slopes, pools, hoammocks and sandy spots. Almost all of
the visible coral of this region is living coral, associated
with an abundance of corallines, bryozoans, hydroids and
red and brown alge. Some of the algal genera that are
confined largely to the outer reef-rim are: Codium, As-
paragopsis, Gymnogongrus, Porphyra, Turbinaria, Dic-
tyota, Haliseris, Gelidium, ete. Many of the speties that
inhabit these turbulent and surf-churned waters are not
the tough, cartilaginous forms, but are very delicate and
fragile species, that apparently survive the wave action
because of their very delicacy. This is particularly true
of some of the finer red alge.*

Highly important on the Hawaiian reefs are the coral-
line or stony algæ or nullipores. A number of genera—

4Some of the representative marine alge of Hawaii that are common on
the coral reefs and ‘shallows are: Oscillatoria bonnemaisonii, Phorm: idiu
crosbyanum, Lyngbya semiplena, L. majuseula, Hydrocoleus cant 'OSMUS,
Nodularia Hawaiiensis, Hormothamnion is. Scytonema fuliginosum,
Calo tihe æruginea, Ulva spp., Enteromorpha spp., Chetomorpha pacifica,
Cladaphora spp., Bryopsis poneros peo se Halimeda spp.,
Codium spp., Valonia spp., Dictyo ia favulosa, Microdictyon umbili- :
catum, Ectocarpus spp., po aaa rei SPP., eo cancellatus, Aspero-
coccus aii te dabas ornata, Sargassum spp., Padina pavonia, Dic-
tyota spp., gora decussata, Galaxaura lapidescens, Scinaia furcellata,
Gelidium ain ici Diada Gymnogongrus spp, Ahnfeltia concinna,
Gracilaria spp., Hypnea nidij . armata, Plocamiwm sandvicense, Mar-
tensia. flabelliformis, ocd ed ea guia spp., Chondria
tenuissima, Polysiphonia spp., , Ceramium spp., Grate-
loupia filicina, Fiii ihe mere het spp., or hothamnion

424 THE AMERICAN NATURALIST [ Vou. LIT

Lithothamnion, Corallina, Mastophora, and others—are
abundant on the reefs, and undoubtedly have been active as
reef-builders. The importance of the lime-secreting alge
was overlooked by the earlier students of coral reefs, but
is now receiving adequate consideration. Howe? shows
that these forms work effectively at greater depths, and
at lower temperatures than do the true corals, and that
they are much more generally and widely distributed
than the latter. >

The Hawaiian coralline alge inhabit the shallow waters,
as well as occurring at considerable depths. In the
former situations they form beautiful rose, purple and
lavender incrustations. On the faces of cliffs that are
washed by the sea these incrustations appear as con-
spicuous bands, extending from high-tide mark or the up-
permost wash of the surf down to the zone of minimum
illumination. The lower margin of the coralline zone has
not been investigated in the Hawaiian Islands, but in
other island groups they flourish at 1,000 ft. depth. In
the coralline zone are also many of the calcareous hydro-
zoa.

Sponges of many species, sizes and colors abound in all
protected portions of the reefs, but have never been made
the subject of critical taxonomic study. They range from
tiny, fragile forms, the size of a shoe-button, up to coarse
horny masses as large as a man’s head. The lesser
species are common on the coral-rock litter in the lagoons.
The larger forms inhabit the deeper waters, and are torn
from their anchorage only by the occasional severe storms.
After a period of southerly storms, for example, the lee-
ward beaches are littered with these large, tough sponges,
_ which average eight inches in diameter.

The range of color is bizarre and striking. In a single
afternoon’s collecting one may pick up, in the shallow
water, species of bright red, pale yellow, rich purple, dull
brown, creamy white, green, and dead black pigmentation.
Dredging reveals others which add to the chromatic series.
Most of the sponges are of the encrusting type, the body

5M. A. Howe, ‘‘Building of Coral Reefs,’’ Science, 36: 837-842, 1912.

Nos. 620-621] THE HAWAIIAN CORAL REEFS 425
conforming to the substratum and having no definite
shape. The Calcarea are not uncommon in the littoral
region, especially in sheltered situations among rocks and
seaweed. These and the true horny sponges (Ceratosa)
have not been found below 2,700 ft. The sponges found
at the greatest oceanic depths are members of the Hexac-
tinellida and Choristida of the Non-Calcarea.

EF EZ
A 3
== N
SS
4
>
y a 2
E,
š a
ss
ne
ES
A Zz FEF By
> sz Ti
zz 2
y 279 F
EZ.
3
` rN
` 3
= 3
SS 2,
mAN A
SS SS <A
4 SSS TAN
== Ws SS Fe == v3
<< XA ZA Ek 2
- Pr SB AA ERE
¿<= Sr EFS FZ CATRE AL AS
> eS Z
“E nn
LES = oe
TTT de

Fig. 6. Island of Oahu. Showing extensive development of coral reefs
and elevated coral limestone We Note the abundance of coral in the vicinity
of Pearl Harbor (the fan-shaped bay h

uth coast), Kane-ohe Bay
northeast coast) and along the southern shores.

The Hawaiian sponges have few or no natural enemies,
and do not appear to be edible to fishes, crustaceans or
molluses. Innumerable lowly forms, however, inhabit
their tissues, for shelter, if not for food. The interior
of any one of our common reef sponges is almost sure to
be found teeming with minute crustaceans, annelids, mol-

luses and other invertebrates.

None of the Hawaiian species have been utilized com-
mercially and no serious attempts have been made to in-

426 THE AMERICAN NATURALIST [ Vou. LII

troduce and establish the valuable species from other
parts of the world. All of the commercial sponges be-
long to the two genera Euspongia and Hippospongia,
which do not occur on the Hawaiian reefs. Such an en-
terprise, if undertaken with thorough scientific supervi-
sion, would unquestionably meet with success. There are
many areas along our reefs where the sponges could be
established. With adequate labor and marketing ar-
rangements a steady development of the industry would
be assured.

It is of interest to note that of the fresh-water sponges,
Spongilline, a group which is widely distributed in «all
parts of the world, no representatives have been taken
in the Hawaiian Islands.

Jellyfish are of casual occurrence along our reefs. The
smaller forms are chiefly Hydrozoan meduse; the larger
ones are Seyphozoans. A relatively few species are
known and the life-cycles of these are not known in detail.
A number of the tiny species are phosphorescent, and on
clear nights when the sea is calm and other conditions are
favorable, they give beautiful luminous effects. In pad-
dling along the reef in an outrigger canoe, on such a night,
the paddles, at every stroke, drip with tiny stars. Many
of the larger species have gonads, tentacles, radial canals,
or other organs brilliantly pigmented.

The large forms attain diameters of 8-12 in. and some-
times appear in great numbers in quiet, protected waters.
Pearl Harbor, for example, which is almost wholly land-
locked, is a favorite habitat. At low tide, in other parts
of the islands, along the coral beaches one sometimes finds
great numbers of jellyfish stranded and slowly deliquese-
ing

En addition to the true jellyfish the reefs support a rich
hydrozoan« or marine hydroid fauna. The littoral species
have not been studied taxonomically ; the Albatross collec-
tions were made at depths of 60-3,000 ft. These latter
comprised 49 species, representing 27 genera and 11 fami-
lies. The shallow-water zoophytes or hydroids are abun-

Nos. 620-621] THE HAWAIIAN CORAL REEFS 427

dant in all protected situations; many forms also inhabit
the surf-beaten rim. Sertularia, Plumularia and Cam-
panularia are well-known genera. The species are all of
small size and superficially resemble in habit, color and
habitat the more delicate marine alge.

The false corallines or Hydrocoralline are also very
abundant and have played an important rôle, as have the
coralline alge, in the construction of the Hawaiian reefs.
These colonial animals resemble delicately branching
corals; their bleached and rather fragile skeletons are
common on the beaches. When alive the corallines are
of various tints of pink, orange and salmon, and add
bright touches of color to the brilliant ensemble of the
reef. The Hydrocoralline occur only in tropical seas;
Millepora and Stylaster are typical genera.

That remarkable order of free-floating colonial hy-
droids, the Siphonophora, is well represented in all trop-
ical waters, and has numerous forms in the Hawaiian
marine fauna. This group exhibits the greatest diversity
of form. The common ‘‘Portugese man-of-war,’’ Phy-
salia utriculua, with its brilliant peacock-blue float and
long retractile tentacles, is abundant along the reefs and
shallows, and like the jellyfish, is often cast ashore in
enormous numbers. The tentacles contain powerful bat-
teries of stinging capsules ; the wounds are intensely pain-
ful, and so this lovely evanescent creature is dreaded by
bathers. Other well-known genera are Halistemma,
Diphyes, Porpita and Vellelo. Porpita pacifica, the sea-
money, is a beautiful blue-fringed dise about 13 in. in
diameter. Vellela pacifica is also abundant at certain
seasons. It resembles Physalia, but has much shorter
tentacles.

Sea-anemones, Actiniaria, are abundant along. the
Hawaiian reefs, but no taxonomic studies have been made.
A number of species inhabit the inshore pools whose
waters are periodically renewed by waves or tides; others
may be found on the floor of the lagoon, and still others
on the protected sides of rocks which stand in the heavy

428 THE AMERICAN NATURALIST [ Vou. LIT

surf. The colors most frequently observed are shades of
tan, olive and purple; some forms have tentacles which
are beautifully pigmented. The size varied from species
so minute as to almost escape detection up to fine showy
forms 1-2 in. in diameter. They form considerable colo-
nies, sometimes covering areas of several square feet.
Isolated individuals, particularly of the larger species,
are not rare. Usually their rosette of tentacles and bril-
liant color renders them quite conspicuous, but many
kinds are embedded more or less completely in the sub-
stratum, and upon the slightest alarm contract into shape-
less lumps, and are thus easily overlooked.

p

of repeated subsidence and elevation, The crater on the right is Diamond Head
(Leahi) ; the channel to the left is Kalihi Channel. Honolulu Harbor is the
middle channel.

The Ctenophore have about 20 known species in
Hawaiian waters, but these are so rare and fragile that
they are practically unknown to the reef-collector. They
are all pelagic, delicate, transparent creatures, with long
tentacles and peculiar comb-like locomotor organs. As
they swim gently through the sunlit waters their trans-
parent bodies and tentacles yield beautiful iridescent re-
flections. Hormiphora, Cestus, and Beroé are well-
known genera. All the members of this highly specialized
group are solitary and do not form skeletons.

F

Nos. 620-621] THE HAWAIIAN CORAL REEFS 429

The fauna which inhabits the innumerable small cavi-
ties in the coral, and which drills countless tunnels
through the soft rock, is of much interest. This fauna
comprises chiefly the worm-like animals or sea-worms.
Important groups are Turbellaria, Nemertinea, and An-
nelids. Some species creep about in the interstices;
others construct covered passageways on the surface of
the coral. Others burrow in the sand and mud on the
floor of the lagoon. Some tunnel deviously through the
coral rock itself. Many of the sea-worms are brightly
colored. Little is known concerning the relationships or
life-histories of the Hawaiian forms. Nereis, Serpula,
Terebella, Tubifex, Sipunculus, and Echiurus are charac-
teristic annelid genera.

The true corallines (Polyzoa) or sea-mats bear a close
resemblance to the hydroid zoophytes, and only upon
microscopic inspection show that their organization is
much higher than that of the hydroids. The skeleton is
not exclusivély calcareous; in many forms it is chitinous
or even gelatinous. These corallines are abundant on
the Hawaiian reefs.

The true starfishes, Asterioidea, are comparatively rare
on the reefs themselves, although fairly common in the
offshore waters. The brittle-stars, Ophiuridea, are the
common reef forms, and lurk in every cranny. The Alba-
tross expedition collected 60 Hawaiian species of true
starfish during its dredging operations in the island chan-
nels; they were taken at depths of 60-6,000 ft. These
rapranta 46 genera and 20 families; 52 species were
new to science. According to Bryan, _

Large specimens of an eight-rayed starfish, Luidia hystrix, are occa-
sionally captured at Pearl Harbor. They are often a foot and a half
in diameter. A similar but very small species is to be found abundantly
in the coarse green sponges in Kalihi Bay and at Pearl Harbor. A
small, stiff, irregularly developed, pink, leather-like species, Linckia sp.

. without spines, is occasionally found crowded into small holes in the
coral reef.

The common brittle-star, Phiema sp., is blue-black in
color, with small body and long snaky arms. It is gre-

430 THE AMERICAN NATURALIST [ Vou. LII

garious in habit, and the collector frequently finds a dozen
or more congregated beneath a half-buried stone or coral
mass. A tiny pink species, Ophiothrix sp., with remark-
ably long arms, inhabits crevices in the coral. It is very
difficult to capture intact, because it, like most of the
Ophiuroids, possesses to a remarkable degree the faculty
of self-mutilation. Many of the Hawaiian brittle-stars,
when disturbed or removed from the water, sever por-
tions of their arms piece by piece until finally nothing is
left but the central dise. This is capable of developing a
new set of arms; and a detached arm can, under favorable
conditions, develop a new dise and a completed. series of
arms. The basket-stars, Cladiophiure, have never been
collected on the Hawaiian reefs.

The sea-urchins, Echinoidea, are richly represented, but
most of the species inhabit the deep offshore waters.
The inshore species are gregarious and common in all
rocky situations along the coasts, as well as on the reefs
themselves. Podophora pedifera, for example, prefers
the black lava rocks and cliffs exposed to the full force of
the surf, and is so abundant that the zone of massive
basalt which it inhabits is literally honeyeombed with its
burrows. Several species of Echinometra are also very
abundant; these prefer the shallow waters of the lagoons.
In the deep holes and caverns along the outer edge of the
reef is a large purple-black species, Diadema paucispinum,
with slender, awl-shaped spines. In the same situations
occurs Echinothrix desori, a large form whose long spine§
are beautifully banded with gray and black. The curious
club-spined urchins, Heterocentrotes spp., occur here and
there along the reef, and are frequently exhibited in the
Honolulu Aquarium. The sea-biscuit, Brissus carinatus,
is a large, heart-shaped urchin, covered with short, brown
hair-like spines, and is occasionally found along the reef
rim. A number of the Hawaiian urchins are known to the
natives as wana, or ““sea-eggs,”” and are habitually used —
by them for food. They may be purchased in the local
fish markets.

Nos. 620-621] THE HAWAIIAN CORAL REEFS 431

Numerous species of holothurians (known as sea-cu-
cumbers, sea-squirts, and béche-de-mer) are common in
the shallow waters. There are over 40 described species,
representing 4 families and 21 genera. A large, worm-
like form, Opheodesoma spectabilis, is common at Pearl
Harbor and Kaneohe Bay, in quiet water. It is about
2 ft. long and 13 inch diameter, reddish brown mottled
with brown. A large, dark greenish-brown species,
Stichopus tropicalis, is plentiful in the large pools of the
outer reef, near Honolulu. Inhabiting the tidal pools in
the lava rocks is another large form, Holothuria atra;
dark brown, and with ambulacral feet scattered all over
its body. Frequently associated with it is a heliotrope-
purple species, Holothuria cinerascens. There are about
600 known species of holothurians, varying in size from
3 inch to 2 or 3 feet. They are found in all seas, but are
particularly abundant in the West Indies, and between
Asia and Australia. They feed chiefly on Foraminifera.
The movements of all the Hawaiian species are very slug-
gish; they seem to have few enemies. All are harmless,
although of unpleasant aspect. They are capable of the
most extraordinary regeneration of parts, even of the
most important organs. Many species show the curious
habit of evisceration—when alarmed they dispel from
the anal opening the viscera either wholly or in part.
In the course of a few weeks all of the lost organs are
replaced by a new set. | |

The Crinoids or sea-lilies do not exist in the shallow
waters of the Hawaiian reefs. A dozen forms were col-
lected by the Albatross at depths of about 600 ft. These
all proved to be new species, although representing 8
wide-ranging genera in 4 families of the non-stalked Neo-
Crinoidea. Crinoidal fossils have not been found in the
uplifted coral limestone beds of the Hawaiian Archi-
pelago. These forms made important contributions to
the Silurian and Devonian rock strata in other parts of
the world, during which epochs the crinoids were enor- _
mously abundant.

432 THE AMERICAN NATURALIST [ Von. LII

Molluses abound on all the reefs. There is a tremen-
dous range of size, structure, habitat and generic repre-
sentation. The marine molluscan fauna has never re-
ceived adequate attention, as scientific interest has cen-
tered upon the unique terrestrial and arboreal forms.
There are about 20 species of bivalves (Pelecypoda) that
are fairly common. These include such genera as Mytilus,
Perna, Arca, Ostrea, Anomia, Pecten, Tellina, Cadokia,
Cytherea, Venus, Cardium and Chama. Tellina sugosa,
the Olepe, is, according to Bryan, ‘‘the most important
shell-bearing mollusc”* in the islands. The famous pearl
shell, Avicula margaritifera, of the South Pacific, does not
occur in Hawaiian waters. The Hawaiian pearl oyster,
pa, Margaritifera fimbriata, has a shell often 3 or 4 inches
broad, with a brilliant iridescent interior. It is the
species which gave Pearl Harbor its name. In the early
days the collecting of pa was a royal monopoly, like the
collecting of sandalwood. The pearl-shell was used by
the Hawaiians chiefly for making fish-hooks, and for the
curious shell-eyes of their wooden gods. A true pearl-
bearing species also occurs at Pearl Harbor and other
localities in the group in the deeper offshore waters. The
_ edible oysters are represented by Ostrea rosea, which is

not of sufficient abundance to permit commercial exploita-
tion.

The chitons and their allies, Amphineura, are uncom-
mon in the shallows, but a thorough systematic survey
would undoubtedly bring to light many additional forms.
The true chitons, Placophora, are confined largely to the
shallows, and apparently are herbivorous, feeding on
minute alge and diatoms. The Aplacophora as worm-
like, shell-less creatures, with the body beset with cal-
careous spicules. They are wholly absent from the lit-
toral zone, occurring only at considerable depths—3,000
ft. and in some instances down to 7,500 ft. They are car-
nivorous and subsist on such small animals as hydroids
and coral polyps.

- The univalves or Gasteropods are by far the most

Nos. 620-621] THE HAWAIIAN CORAL REEFS 433

abundant molluscs on the Hawaiian reefs and aggregate
several hundred species. Space forbids any detailed
treatment of this huge and highly diversified class. A`
bare enumeration of important and common families and
genera, adapted from Bryan, will indicate the richness of
the marine univalve fauna (the number of species is in
each instance an approximation) :

Family Species Typical Genera
Tiitoge; Thitonide (iio 2415; 12 Triton, Ranella.
Spiny Rock Shells; Muricide .. 30 Purpura, ees Vexilla, Sistrum.
Spindle Shells; Fuside ........ 6 Fusus, Latirus, Peristernia.
olks; Bueeinidm. ici. 6 Pisania, pc US.
Dog Whelks; Nasside ........ 4 Pb
Mitre Shells; Mitride ........ 26 Mitra, Imbricaria, Turricula.
Margin Shells; Marginellide ... 4 Erato, Mar ch inella.
Olive Shells; Olivide ......... 4 Harpa,
Dove Shells; Columbellidee .... 15 Columbella, Puen
Cone Shells; Conide .......... 25 Conus.
r ‘Shells; Terebride ...... 25 Terebra
Conch Shells; Strombidæ ...... 9 trombus
owry Shells; Cypreide ...... 40 Cyprea, Trivia
Tun Shells; Doliide .......... 3- p
Cameo Shells; Cassisæ ........ 5 sis
Moon Shells; Naticide ....... 10 Natica.
asim ge Calypreide .... 12 Crepidula, Crucibulum, Hipponyz.
Eulim PATI oe se vas 1 i

ri lima.
Prs Shells; aie saa 3 Pyramidella.
6 Solarium.

Violet Snails; Aripa Uneen: 3 Ianthina.

Ladder Aor; Sealariide ..... 10 Scalaria.

Herald’s-horn Shells; Cerithiide 20 Cerithiwm

Periminides: : isola VERS 3 Littorina; cepto Risella,
Sea-Snails; Neritide .......... 10 Nerita, Ner

Turban Shells; Turbinidæ ..... 18 Turbo, Prasianeila, Astralium.
ajad penne Trechide. cc. Trochus

12
; Fissurellide and Patel- H Jilowa
PEE A E ule ok Wakes 10

Sea Ea, Nudibranchiata .... 10 Aplysia.

The highest and most highly specialized class of mol-
luses, the Cephalopods, have an abundant and familiar
representative on our reefs, in the form of the common
octopus or he’e. This is popularly called ‘‘squid’’
although it is a true cuttlefish, with a small round sac-like
body and eight arms. It is very common in holes and

434 . THE AMERICAN NATURALIST [ Vou. LIT

pools on the rocky platform of the reef, and in caverns
along the reef-rim. During the day it hides in cavities;
at night it creeps about over the rocks of the bottom.
The natives are very fond of the flesh, which they prepare
for food in a variety of forms. Dried ‘‘squid’’ is com-
mon in the fish-markets. Our cuttlefish is rarely more
than 18-26 inches in length. The true ‘‘devil-fish’’ of
many a sea-tale is a giant squid, Architeuthis, which in-
habits the Newfoundland banks and often attains the gi-
gantic proportions of an over-all length of 50 ft., with a
body 6 by 9 ft., and enormous arms 40 ft. long.

N

A

“se - paT a —

ER em AT ER o pa ee e
Fig. 8. Coral Reefs near Pearl Harbor, Oahu. The inlet to Pearl Harbor
is shown to the right. To the left is the southwest point (Barber’s Point) of

the island. The lowland is a plain of coral limestone; the reef is rich in bio-
logical material.

Our reefs support a characteristic crustacean fauna.
In the growing coral at the reef-edge are found a number
of small Cyclometopous crabs, which are often beauti-
fully sculptured and colored to harmonize with the coral.
The Alpheide, which are shrimp-like Macrura with highly
asymmetrical claws, are commonly found in pools on the
reef. In the coral rubble formed by the disintegration
of the reef-rim, in relatively shallow water, numerous
Leucosiid crabs are found. Many lowly forms of Ento-
mostraca are abundant, but have never been surveyed

Nos. 620-621] THE HAWAIIAN CORAL REEFS 435

taxonomically. The Phyllopods, Ostracods and Cope-
pods are all plentiful. The Cirripedia include the most
aberrant of the crustaceans, and are represented by the
common barnacles, including both the stalked (Lepadid:e)
or goose barnacles, and the sessile (Balanide) or acorn-

. 9. Kane-ohe Bay, Oahu. This bay, which is a drowned valley com-
plex, ee a great variety of coral formations. There are many small coral isles
and atolls; some are of notable perfection. The exact boundaries of the outer

reefs are not known. The crater and little isles to the lower right are secondary
volcanic products.

shells. The latter are exceedingly abundant along the
shores and reefs; there are also numerous deep-water bar-
nacles.

436 THE AMERICAN NATURALIST [Vou. LII

The most commonly known, the largest and the most
highly organized crustaceans, the Malacostraca, are very
common. Space does not permit even a general sketch
of the many crabs, prawns, crayfish and other interesting
forms that teem in the Hawaiian littoral. The so-called
Hawaiian ‘‘lobster,’’ ula, Panulirus japonicus, is really a
large marine crayfish, and not closely related to the true
lobster. It is brilliantly colored and ornamented, with
spiny carapace and long antenne. The ula is common in
the fish markets, as are also species of Scyllarides, Ocy-
poda, and many crabs. Hermit crabs (Onomura) are
common and in great variety. They make their homes in
empty sea-shells, and have many interesting habits.°

The last great division of the reef fauna comprises the
fishes, a group that could easily occupy the space of an
extended monograph. There are several hundred reef
species, occupying a wide range of habitats, and varying
in size from minute species up to huge food-fish weighing
a hundred pounds each. Like the fish of many tropical
waters, the Hawaiian species are famous for their bril-
liant coloration, fantastic patterns, and strange shapes.
Many are grotesque; many are exceedingly beautiful;
many are consummate embodiment of that riot of gor-
geous color that is so characteristic of the reef and its
life. The reef fishes, like the other littoral forms of life,
were an important item in the dietary of the primitive
Hawaiians, and continue so to the present day. Most of
the common species are offered for sale in the fish-markets.
Space is not available for any detailed account of the in-

6 The following list of common littoral and reef species and genera of
malacostraceans will indicate the richness of this portion of Hawaii’s re-
markable r fau B

eef na: L —Ocypode ceratophthalma, O. levis,
grapsus, Metopgrapsus messor, Paci p plicatus, Cyclograp-
sus, Perenon, Carpilius, Pl podia, Lophozozymus, X A odius,

ci ez,
Parribacus, Devito Stenopus, Peneus, Hippolysmata, Spirontocaris,

Nos. 620-621] THE HAWAIIAN CORAL REEFS 437

shore” fishes, as contrasted with the pelagic and abyssal
species.

7 Some of the more important groups and species may be listed as follows:

Blarke 9s. Os Carcharias melanopterus; C. nesiotes; Sphyrna zy-
gena; seis glanon:
Pek ccd ORs pees Stoasodon nari
'Tarpons-........... Elops saurus
öhöhos *. si... Albula vulpes
PERSEA curiae sre > hanos
ANCHOVIES .:-.. r-r: Anchovia purpu
tral a Trachinocephalus myops; Synodus varius; Saurida
racilis,
Conger Eels ........ Leptocephalus marginatus,
DOTA VSG bee es Murena, Enchelynas EE Eurymyctera,
Echidna, Propia die: Scuticaria.
Trumpetfishes ....... Aulostomus valentini.
Cornetfishes ........Fistularia petim
Needlefishes. .......- Athlennes hians.
TniP-DERRA Ls Hemiramphus depauperatus; Euleptorhamphus longi-
rostris.
Flyingfishes ........ Parexocetus brachypterus; Cypsilurus simus; C. ba-
haiensis.
Be eal E Atherina insularum.
Mulets ee a ee Mugil p
oe ga Tee ge Sphyrena helleri
Thresanns Glas. aos Polydactylus sexfilis.
hunen E olotrachys lima; Myripristis Bp, ; Flammeo sam-
; mara; F. scythrops; Holocentrus spp.
Big-eyed pg Pe Trachurops PAS Carangus.
Threadfishes ....... — ci ng
idad oes oa, Amia.
POUPOPE Opa ccs Bpinepaue detain
Cathlafas >.. ipoe Priacan
Snappers a.s: prani pai Bowersia, Aprion, Etelis.
DOLOR es Monotaxis grandiculis.
Rudderfishes ....... Kyph
Surnmullets .. 0. Mulloides.
bos a Pseudupeneus ; Upeneus
Demoiselles ........ Dascyllus; Chromis; Po Caiena, Abudefduf.
Weert ots PA Lepidaplois; Stethojulis; PENES Gomphosus ;
Anampses assomaā ; oris; Cheilio;
Cheitinns's Novaculichthys; Titis, Hemiptero-
notus; icht
is one Calotomus ; Cally
Butterflyfishes ...... Forcipiger ; Chaton: Holocanthus.
Moorish Idols ...... Zanclus canescens.
Surgeonfish

E Hepatus, Zebrasoma, Ctenochetus, Acanthurus, Calli-
canthus.

438 THE AMERICAN NATURALIST [ Vou. LII

The gorgeous colors of many of our reef fishes are very
evanescent, and undergo rapid deterioration when the
fish is taken from the water. Hence the coloration of
those offered for sale in the markets often conveys but
little idea of their living hues. Preserved specimens and
printed descriptions are of even less value.

In concluding this condensed sketch of the Hawaiian
reefs, the writer desires to emphasize his impression of
the struggle for life which goes on there unceasingly,
The reef is a region of intense competition. It is com-
parable in many of its ecologic relations to the montane
rain-forest. The excessive illumination of the reef is
perhaps as constraining an influence as is the excessive
humidity of the rain-forest. The diversity of organisms
which inhabit the reef is far greater than that of any other
island habitat. The competition for food is keen and
unremittent.

The reef as a food supply for human beings has been
a dominant factor in the lives of the primitive Poly-
nesians. Through the experiences of thousands of years
they have acquired a very intimate knowledge of the reef
and its life, but this has never been given adequate scien-
tific investigation. One of the great forward steps in the
economic history of the world will be the scientific utiliza-
tion of coral reefs and their products.

Triggerfishes ....... Balistes, pinar ai
Pafos ici ows Tetraodon, Can
Trunkfshes os. e Ostracion,

fis) nia
Cirrhitoid fishes kesic irikita,
Mail-cheeked tia . Caracanthus.
Scorpenids

OTPIENTOS ias... Sebastapistes, Sebastopsis, Scorpenopsis.
Gobies AA ie cee otris sandwicensis, Asterropteryx semipunctatus,
keut epiphanus iomorphus eugenius, Mapo

US Gobiichthys, Gnatholepis knighti.

cus,
ee Enneapterygius atriceps, Al ticus, Enchelyurus, Sala-
rias.

CONTINUOUS AND DISCONTINUOUS VARIA-
TIONS AND THEIR INHERITANCE IN
PEROMYSCUS. III

DR. F. B. SUMNER

Scripps Institution, La JOLLA, CALIF.

VII. Mutations

In a recent paper (1917a) I have described two widely
aberrant color types which have appeared in my cultures,
together with certain minor deviations, which likewise
seem to behave as discontinuous variations. I am pre-
pared to add considerably to the data thus far published.

1. “Partial albinos’’

The term albino, as applied to these mice, admittedly
does not conform to current usage, and this has become
especially evident with the appearance of the mature pel-
age. I do not think, however, that any of the various
names given to fancy races of Mus musculus apply to
these animals. Having at hand no specimens or even
satisfactory colored plates of fancy mice, I am unable to
make the comparisons.? As previously stated, this mu-
tant strain has red eyes, and lacks pigment wholly on the
ears and tail. The fur, on the colored region of the body,
is a very pale gray, rather strongly tinged with a shade
of yellow approaching Ridgway’s ‘‘ochraceous buff,’’ or
perhaps ‘‘ochraceous orange,’’ on the most highly col-
ored areas. As a convenient non-committal expression,
I shall henceforth employ the term ‘‘pallid’’ for these
mice.”

A microscopic examination of the hairs of these mice
reveals some interesting departures from the normal con-
dition:?S (1) a considerable proportion of the hairs are
practically devoid of pigment in the zone which is ordi-
narily yellow, while the rest are normal in this respect;

27 It may be that the factional modifications of my mice are the same as
those of Castle’s ‘‘red-eyed yellow’’ rats (see Castle and Wright, 1915).

28 Cf. Morgan’s account (1911) .of the hair of some ‘‘modified’’ indi-
viduals of Peromyscus leucopus ammodytes.

439

440 THE AMERICAN NATURALIST [Vou. LIL

(2) the surface pigment of the terminal portion of the
hairs is nearly or quite lacking; (3) in the basal zone, the
black pigment bodies are represented by small flocculent
dark masses. Thus, we are not, as in the next ‘‘mutant”’
to be described, merely dealing with changed proportions
of perfectly normal types of hair. These red-eyed mice
possess types which I have not found in any of the rest
of my stock.

At the time of my earlier description of these pale
sports, no young had been obtained, but their pedigree
suggested that they were simple Mendelian recessives.
This conjecture has thus far been sustained. The two
‘‘mutants,’’ bred to one another, have given six pale
young, like themselves, and no others. When bred to
dark mates, of the same stock as themselves (sonoriensis-
rubidus hybrids), the pallid animals gave only dark
young, except in a single instance where the dark parent
was known to be heterozygous. In this case, one pallid
mouse was the outcome. Of the dark progeny, three
broods, aggregating eleven individuals, have thus far
been born.

This clear-cut and typical example of Mendelian seg-
regation, in respect to these mutant characters, is in
striking contrast to the complete lack of segregation—
so far as is obvious—in respect to the subspecific charac-
ters which have entered into the germinal constitution of
these same individuals.

As I have previously stated, these parent ‘‘mutants’”’
were the offspring of F, sonoriensis-rubidus hybrids. In
a recent article (1917) the Hagedoorns have described a
number of strongly aberrant types of rats (including
some waltzers!) which appeared in a mongrel strain re-
sulting from the crossing of Mus alexandrinus, M. tecto-
rum and M. rattus. The authors recognize in these aber-
rant derivations some entirely new products, though they
do not attribute their origin to real mutation. In the
opinion of the Hagedoorns, as I understand it, these ap-
parently ‘‘mutant’’ characters have resulted, in each
ease, from the chance coming together of two recessive
factors (or two ‘‘absences,’’ according to the prevailing

Nos. 620-621] INHERITANCE IN PEROMYSCUS 441

theory). No two of these ‘‘absences’’ coexisted in the
gametes of any one of the parent species, and no single

‘‘absence”’ by itself is believed to be adequate to produce
one of the abnormalities. Since the average number of
each kind of ““mutant”” in their stock of 37 was approxi-
mately 1 in 16, they assert:

These numbers make it clear that we are not dealing with a sort of
period of mutation; it was easy to see that the new types were already
given in the A of the three species erossed (p. 415).

And in later passages the de generalize this con-
jecture, as for example:

The only cause for inheritable variability in multicellular organisms
which can be of any account in evolution is mating between individuals
of unequal genotype, crossing in the widest sense (Amphimixis) (p.
418).

That the Hagedoorns’s explanation does not fit the case
of my pallid Peromyscus is evident from the history of
the stock. I have obtained, in all, 47 F, offspring from
the mating of F, sonoriensis-rubidus hybrids, counting
only those animals which lived long enough to reveal their
essential color characters. These were the progeny- of
six different fathers and eleven different mothers. Just
four of these very pale sports have appeared in my F,
stock. They are the offspring of a single father by two
mothers, both his own sisters. These mothers, by the
same futher: also produced seven dark young.

It seems plain, therefore, that the mutation in question
is not due to any recombination of factors (or their ‘‘ab-
sences’’) regularly occurring in the parent races. If it
were, we should reasonably have expected similar aberra-
tions among the offspring of other parents. It is hard to
determine from their published statement the exact pedi-
gree of the Hagedoorn’s aberrant rats. But one thing
seems plain. All were the progeny of a single father by
two mothers, the latter apparently being sisters. The au-
thors are certainly not warranted, therefore, in assuming
that such results would have been obtained by mating any
animals of the same racial composition.

I am inclined to think that my pale red-eyed mice are
true mutants, i. e., that they appeared de novo in my cul-

442 THE AMERICAN NATURALIST [ Von. LII

tures. It is more than possible, likewise, that the hybridi-
zation of such diverse strains was the disturbing element
that led to the loss or modification of a ‘‘gene.’’? The
latter possibility is strengthened by a consideration of
the Hagedoorns’s waltzing rats and abnormalities of coat
color. But this is a very different view from the hypothe-
sis that ‘‘the new types were already given in the geno-
type of the . . . species crossed.””

2. Yellow gambeli

The five normally colored progeny of a single pair of
normally colored Peromyscus maniculatus gambeli (La
Jolla race) became the parents of 21 offspring, of which
14 were normally colored and 7 were of a peculiar yel-
lowish-brown color. These ‘‘mutants,’? which I have
called ‘‘yellows,’’ are of a shade not very far removed
from Ridgway’s ‘‘clay color.” They are considerably
darker than some, at least, of the yellow races of Mus
musculus. Microscopic examination of the hair of these
aberrant gambeli shows that it is closely similar to that
found upon the more highly colored parts of P. m. sono-
riensis. In comparison with normal specimens of its sub-
species the mutant strain is found to have a larger num-
ber of the yellow-banded hairs, in proportion to those
which are black throughout their entire length. The
latter type of hair is, however, by no means wanting. In
the second place, the yellow zone of each hair, on the col-
ored parts of the body, occupies, on the average, a con-
siderably larger proportion of its length. On the mid-
ventral surface, the basal, plumbeous zone is quite lacking,
the hairs being entirely white. Besides the differences
stated, I can not be certain of any hair characters which
distinguish this type of sports from the normal stock.
Moreover, the eyes, ears, tail, ete., carry a normal amount
of black pigment.

It is to be noted that these ““yellow”” mice, unlike the
““partial albinos,”” are not distinguished by any types of
hair which are lacking in normal individuals. We may,
however, very justly regard the yellow condition as having
arisen through ‘discontinuous variation.” Though due

Nos. 620-621] INHERITANCE IN PEROMYSCUS 443

merely to a change in the proportion of elements pre-
viously present, the new type has arisen abruptly and has
diverged so widely that its range of variation does not
overlap that of the normal race. Among the many hun-
dreds of individuals which I have dealt with, I have never
found any mice which would serve in a true sense to
bridge the gap between these two types. Nor have any
other yellows appeared in my cultures, except among the
descendants of the single pair in question.

As stated in an earlier paper (1917a), I trapped several
years ago a mouse which I feel fairly certain was a juve-
nile yellow gambeli. It is possible that this character, in
a heterozygous condition, may be of not uncommon occur-
rence among the mice of this vicinity. Thus, the muta-
tion through which my stock came into existence may
have taken place among the wild ancestors, many gen-
erations earlier. On the other hand, the same genetic
instability which led to such a factorial loss or modifica-
tion in one case may be responsible for its occurrence on
many independent occasions. I have no data by which
to decide between these two alternatives.

As regards the genetic behavior of these yellow mice, I
have fairly satisfactory evidence that they are simple
Mendelian recessives. As was stated above, 7 yellows
and 14 normal animals constituted the fraternities in
which they first appeared. The departure from Men-
delian expectation may well have been accidental here,
though a differential mortality may possibly have been
responsible. The first yellows, bred to their (presumably
heterozygous) parents, have given 5 dark and 5 yellow
offspring. Bred to homozygous dark animals, they have
thus far produced only a single brood, consisting of three
dark individuals. Yellows bred to yellows have produced
young of the aberrant type only (thus far 10). These fre-
quently do not attain the full yellow color until they as-

29Mr. H. H. Collins has, however, found a number of sports of this
gene pearance among the offspring of a single pair of normally
colored individuals which were trapped at La Jolla. Mr. Collins’ mice
differ somewhat in shade, however, from my ‘‘yellows,’’ and may represent

et ‘‘mutation.’’ His experiments have not been carried far enough
to test the genetic behavior of this character.

444 THE AMERICAN NATURALIST [ Vou. LIT

sume the mature pelage, but I no longer have reason to
doubt that the yellow type ‘‘breeds true.’

A yellow female gambeli mated to a ““pallid”” male of
the strain discussed above, has given birth to a single
offspring, having abundant dark pigment in the skin,
eyes and hair. In other words, these two pale, recessive
mutants seem to be ““complementary”” to one another, as
were Castle’s two yellow races of rats (Castle and
Wright, 1915).

3. Discontinuous Variation in Restricted Pigment Areas.

I have discussed briefly elsewhere several sorts of color
markings, along with limited data which seemed to show
that some of ‘these were inherited in alternative fashion.
Other characters of the same type have been added to the
list. For example, in the second cage-born generation of
gambeli I have found three mice with faces strongly
““grizzled”” by the presence of white hairs. It is prob-
ably no mere coincidence that these three grizzled speci-
mens, while not belonging to a single fraternity, are all
descended from the same grandparents. Neither the
parents nor the grandparents were recorded as having
the peculiarity in question, which would hardly have been
overlooked if present. Furthermore, the single off-
spring which I have obtained from a ““grizzled”” pair ex-
hibits this character plainly, though in a reduced degree.
One specimen showing the white-faced condition likewise
appeared in the C, generation of sonoriensis.

Again, occasional mice of perhaps all of the races are
characterized by having considerable pigment in the skin
of the tail. Normally, the skin of this member is nearly
or quite devoid of pigment, the dorsal tail-stripe being
due to black hairs. Examination of two successive gen-
erations of rubidus makes it probable that this caudal skin
pigmentation is likewise a hereditary character. __

I shall here discuss only one of the localized pigment
variations which were dealt with in my earlier report on
color ‘‘mutations.’’ This is the occurrence of a white-
_ tipped snout, due partly to the absence of skin pigment
and partly to the presence in this region of white hairs.

Nos. 620-621] INHERITANCE IN PEROMYSCUS 445

I am now able to indicate rather more definitely the mode
of transmission of this character. I wish to lay some
stress here upon its genetic behavior, since I regard it as
an interesting case in its bearings upon certain theoretical
questions.

The pale-nosed condition has been studied only in the
darkest of my races, rubidus. It was not noticed in the
_ Original wild stock, but it may well have been overlooked,
as it is not a conspicuous character, and I was not search-
ing for this type of variations when the wild generation
was examined. In the first cage-born (‘‘C,’’) generation
twelve mice were recorded as having white-tipped snouts,
four of the cases being entered as doubtful. At the time
of examining these animals I had no idea as to the par-
entage of the individuals, so that there was no bias in my
selection. Upon looking up their pedigrees, I found that
ten of the twelve cases (eight certain and two doubtful)
were the offspring (indeed, the only offspring) of two
mothers of the wild generation (P $ 40 and 41) by a single
father (P ¿ 15). The other two cases (both doubtful)
were of other parentage. In connection with the figures
just given, it must be stated that the C, generation con-
sisted altogether of 60 (surviving) individuals, these
being the progeny of twelve females and nine males.

Only 38 mice were obtained in the C, generation, 6 of
which had white-tipped snouts. As before, the count was
made without my being aware of the parentage of the
individuals examined. Of the six ‘‘mutants,’’ it turned
out that four belonged to a fraternity of five, the fifth
member of which was normal. This fraternity was the
offspring of C, 2 61 (normal) by C, ¢ 10 (white-nosed).
The other two mutants were the offspring of this same
C, 2 61, by one of her brothers (¢ 3), the latter being like-
wise normally pigmented. These parent animals we may
believe to have been heterozygous.

- Unfortunately, none of the matings of the pale-nosed C,
individuals inter se proved fertile, and indeed the only
one of these aberrant mice which left descendants was
the J 10 referred to above.

The relationships here indicated, and the incidence of

446 THE AMERICAN NATURALIST [ Von. LIT

the aberrant condition, are quite intelligible on the as-
sumption that we have to do with a monohybrid recessive
character. The character can not be dominant, for we
had a case of white-nosed young from two dark-nosed
parents. It can not well be due to more than one factor,
owing to the relatively large number of recessive indi-
viduals,

VIII. Discusston

Any one approaching the data dealt with in the fore-
going pages, unhampered by theoretical considerations,
would, I think, conclude that we had to do with two types
of variation and two types of inheritance, differing from
one another in rather fundamental ways. In the one
class we have the continuously graduated differences, oc-
curring within the limits of one of our ‘‘subspecies,’’ as
well as the differences in average or modal condition
which distinguish the various subspecies from one an-
other. Here we find a sensible continuity, both within
and between these rather artificial assemblages of indi-
viduals, and distinct taxonomic units can be recognized
only if we erect more or less arbitrary boundaries. In
heredity, likewise, we have no indication of a dominance
of one step or grade in this series over another, and little
to suggest that two of these grades, once united or blended
in the offspring, tend to reassert their independence in
subsequent generations.

In the other class we have the ‘‘sports’’ or ‘‘muta-
tions.” These are distinctly discontinuous, in relation to
the parent stock, either in the sense that one of the two
possesses elements which are altogether lacking in the
other, or at least in the sense that the new form has under-
gone such a change in the proportions of existing elements
that its range of variation does not overlap that of the
normal race. Looked at in another light, it is of interest
to note that all the mutations which I have discussed, with
a single exception, are dependent upon the loss of some-
thing. The red-eyed ‘‘pallid’’ mice have lost most of
their black and some of their yellow pigment, the ‘‘yel-
lows” have lost much of their black. The white-tipped
tails are due to a loss of part of the dorsal tail-stripe, the

Nos. 620-621] INHERITANCE IN PEROMYSCUS 447

““grizzled”” heads likewise to the local loss of hair pig-
ment, while the white snouts have resulted from a loss of
pigment both in the skin and hair of the latter region.
The single exception among the ‘‘mutations’’ which I
have observed in Peromyscus is the occasional presence
of skin pigment in the tail. Here something has been
definitely added to the usual condition.*

In heredity, too, these mutant characters, whether nega-
tive or positive, behave in distinctly discontinuous fash-
ion. They do not blend, but are either present or absent
in their entirety.

Taken at face value, I say, the evidence shows that we
have to do here with two different types of variation and
two different types of heredity. Now admittedly, the
naive view of such a situation is not necessarily the cor-
rect one, else we should be forced to return to the geocen-
tric theory of the solar system. But even in this last in-
stance, the burden of proof most assuredly rested on the
man who first asserted that the sun did not move around
the earth. And to-day the same burden rests upon those
who claim—possibly with truth—that heritable variations
are all discontinuous and that blended inheritance is an
illusion.

In the few remaining pages of this paper, it is obviously
impossible to discuss the various lines of evidence which
have been advanced in favor of the Mendelian-mutation-
pure-line scheme of things. I think that few would be
enthusiastic enough to assert that the case had yet been
really proved on evidential grounds. The considerations
which are chiefly effective in determining one’s adherence
to this system of beliefs are doubtless of a more general
nature. Thus it is argued that Mendelian inheritance has
been shown to hold rigidly throughout a vast range of
material, and that, therefore, the ‘‘unit-factor’’ concep-
tion is probably of universal application. Or, it is con-
tended that the scheme of things here considered is more

30 Even in this case, it is possible that we have to do merely with a
““reversion,”” or return to an ancestral condition. Some other Muride (e. g.,
Mus musculus) normally have abundant pigment throughout the skin of
their tails.

448 THE AMERICAN NATURALIST [ Von. LII

in harmony with the atomistic principles of physics and
chemistry. ‘‘Unit-factors’’ have even been identified
with molecules.

In respect to the pigmental characters of our geo-
graphic races, it has been shown to be probable that in-
sensible gradations occur throughout considerable ranges
of territory. There results a series in which marked
contrasts can be found only by comparing individuals
from widely separated localities. The hypothesis that
. the variations in this case are of the Mendelian type
involves the assumption that the subspecific differences
have arisen by a whole succession of small mutations in
the same direction, the number of these mutations being
a function of the distance from some hypothetical center
of dispersal. In a recent paper (1917) Morgan has con-
sidered the question whether there are ““any connections
between the gradations of character in allelomorphic .
series and the order in which the characters appear,’’
i. e., Whether ‘‘the assumed fluctuation of factors is a
sequential process.” He concludes that, ‘‘as a matter
of fact, there is no such relation known . . « for the ac-
tual evidence from multiple allelomorphs shows that
genes may mutate in all directions and also that extreme
mutations such as white eyes arise suddenly from red
and not by graded steps”” (p. 524). These assertions,
which, it is true, were primarily concerned with the
effects of selection, lend little support to the view that
graded geographic variations have arisen through mu-
tation.

The attempt to explain away the substantial mass of
evidence for permanent gametic blending and the indefi-
nite efficacy of selection by invoking the theory of ‘‘mul-
tiple factors” is too well known to be reviewed here.
Castle has been the most able and vigorous opponent of
this theory. At present I will merely refer to certain
evidence of my own which, I think, supports such an ex-
planation no better than Castle's.

The dorsal tail-stripe is entirely lacking in a certain
strain of my mutants. This stripeless condition is reces-
Sive to the striped one. In crosses with normal mice, the

Nos. 620-621] INHERITANCE IN PEROMYSCUS 449

stripe appears in its full size and intensity.* Neverthe-
less, the stripe itself was shown in the preceding pages
to vary from race to race and from one individual to
another. And these variations, both racial and individ-
ual, were found to be hereditary.

The case, of course, is parallel to that of Castle’s
hooded rats. Since “Iroodedness”” is recessive to “self-
color”” and reappears in one fourth of the F, generation,
Castle argues that it is dependent upon a single unit fac-
tor. Nevertheless, this factor itself presents hereditary
variations in ‘‘potency,’’ since it can be modified indefi-
nitely by selection. The Mendelian counter-argument is
that ““hoodedness”” behaves as a unit character in certain
crosses merely because there is some one factor without
which it can not manifest itself at all. The variability in
its degree of manifestation is due to the fact that the

¿Looded pattern is modified by the action of a number of
independent cumulative factors. The argument seems a
bit scholastic, but we must admit that it is logical and
consistent.

Take another instance. Here I admit that my evidence
is to some extent inferential at present. I have given
good grounds for believing that the pigmentless con-
dition of the snout in certain strains of rubidus is a
simple recessive trait, dependent upon a single factor (or
its absence). By this I mean that the pigmentless con-
dition is probably allelomorphic to any degree of pig-
mentation whatever,

Now we find, in examining a series of mice, all grada-
tions from those with heavily pigmented snouts to those
in which no pigment is to be discovered with the aid of a
hand lens. Indeed, there are a few ‘‘borderland’’ cases,
which can be only doubtfully distinguished as pigmented
or unpigmented. Unfortunately, I have no data showing
wLether or not these various gradations are hereditary.
Analogy with the case of the tail-stripe would make it
probable that they are. Moreover, we do know that those

31 The fact that this condition of the tail stripe is but one manifestation
of a mutation which has affected the hair pigment of the entire body does
not affect the argument. It is generally believed that most ‘‘ unit factors’’
manifest themselves in diverse ways.

450 THE AMERICAN NATURALIST [ Vou. LII

differences in the mean condition of the snout which dis-
tinguish the various local races from one another are
hereditary.

Here, too, I am aware that we could argue, with flaw-
less logic, that the pigmentless condition was due to the .
dropping out of some single factor, without which the for-
mation of snout pigment in any quantity was impossible.
Each member of the graded series of pigmented snouts
we might suppose to be conditioned by the presence of
this color factor, together with one to many cumulative
factors determining the degree of its manifestation.

Johannsen (1913), Morgan (1915) and others have
made much of the increased range of variability which
has frequently been met with in the F, and subsequent
generations, even when appearances otherwise pointed to
apermanent blending of types. Recently several writers,
particularly MacDowell (1916) and Little (1917), haves
analyzed some of Castle’s data and have reached conclu-
sions directly opposed to his. All of these authors (Cas-
tle excepted) hold that increasing range of variability in
successive hybrid generations is strong evidence for the
hypothesis of multiple factors, and we must grant that a
pretty good case can be made out along these lines. The
theory runs smoothly until we encounter the awkward
class of facts which Johannsen has called by the name of
““transgressive splitting,” i. e., the ultimate extension of
the range of hybrid variability beyond that of both of the
parent races combined. These facts would seem to prove
too much, despite the ingenious explanation which has
been offered by the pan-Mendelians to account for them.

An analysis of my quite limited data furnishes no evi-
dence of an increased variability in the F, generation,
except where it pretty plainly results from an increase
in the amount of actual abnormality, due to the conditions
of captivity: In the largest, as well as the most normal
of the series, the range of variation actually diminishes
when we pass from the F, to the F, generation. I do not,
however, offer the present evidence as conclusive, even
for the single case of subspecific hybridization in Pero-
myscus. It should be confirmed by data derived from

Nos. 620-621] INHERITANCE IN PEROMYSCUS 451

more extensive series, consisting of animals which are
free from any pathological modifications.

e must urge in passing, however, that evidence of
segregation, even if valid, is not necessarily to be accepted
as evidence of complete segregation. There is no reason
why we might not have partial segregation, combined with
partial gametic blending, as Castle maintains.

In two recent illuminating articles (1917, 1917a), Jen-
nings points out how Mendelian-mutationists of the most
extreme school have been driven by their own researches
into a position that does not differ, according to any prag-
matic test, from the one which they so long have vehe-
mently opposed. More and ever more minute hereditary
differences in the manifestation of a given character are
recognized, until the limit of distinguishability is ap-
proached. This state of affairs has been attributed to two
causes: (1) hereditary modifications in the constitution
of single factors, resulting in the formation of series of
gradations, allelomorphie to one another; and (2) the
existence of series of independent modifying factors,
cumulative in their effects.

As remarked earlier in this paper, the contest has lat-
terly come to resemble that allegorical one of the two
knights, fighting upon the opposite sides of the same bi-
colored shield. And yet there would seem to be a differ-
ence. The two knights in the legend were both equally
right. In the present case, if we may judge by every
pragmatic consideration, the larger measure of right be-
longs to those who have contended for the frequent per-
manent blending of hereditary characters in erossing and
the continuous modifiability of these characters through
selection. The finely spun theories of their opponents
may help us to symbolize the machinery underlying these
phenomena, but. the phenomena themselves, and not the
theories, are the indubitable realities in the case.

IX. Summary a

1. In the preceding pages, the differences, structural
and pigmental, which distinguish four geographical races
of deer-mice are discussed in some detail. The pigmental

452 THE AMERICAN NATURALIST [VoL. LII

differences relate to a considerable range of more or less
independently varying characters, affecting both the in-
tensity and the extensity of the pigment in the hair and
skin. They are found to be, in a general way, correlated
with certain elements of the physical environment, while
the structural differences do not appear to be so corre-
lated.

2. All of these differences, structural and pigmental,
are found to be differences of degree, revealed through a
comparison of mean or modal conditions rather than of
individual animals. In comparing the less divergent of
these races with one another, the frequency polygons for
any given character overlap broadly.

3. These subspecific differences, and even the minor
differences which distinguish one narrowly localized sub-
race from the parent form, are found to be hereditary, as
evidenced by their persistence when environmental con-
ditions are interchanged.

4. The gradations in certain of these characters by
which individuals’ of the same race differ from one an-
other are found to be strongly hereditary.

5. Hybrids between even the most divergent of these
four races are predominantly intermediate in character,
both in the F, and the F, generations. In both of these
generations a wide range of variability is exhibited, which,
however, is little if any greater in the F, than in the F..

` 6. In contrast to the sensibly continuous variation and
sensibly blended inheritance shown in respect to these
subspecific characters, is the behavior of certain ‘‘muta-
tions.”? Here we meet with typical discontinuous varia-
tion, and inheritance of the strictly alternative or Men-
delian type. It is insisted that the burden of proof rests
upon those who contend that these two types of variation
and inheritance are reducible to a single category, that of
discontinuity. Anything like a proof of this contention
appears to be thus far lacking. `

SUPPLEMENTARY Note (Jury 23, 1918).
It gives me pleasure to call attention to points of close
pus sel between certain of the views expressed in the

Nos. 620-621] INHERITANCE IN PEROMYSCUS 453

foregoing pages and ones which have recently been ad-
vanced by Gates (1917) and by Goldschmidt (1918) ; like-
wise to the resemblance between some of the features of
geographic variation which I have described for Pero-
myscus and those which have been observed by Swarth
(1918) for certain birds. . None of these articles had
been published at the time the present paper was written.

LITERATURE CITED
Bateson, W.
1894, renee for Mae Study of Variation. London: Macmillan and

v + 598 pp.
1903. The ti be of Knowledge of Color-Heredity in Mice and
ats. Proceedings Zoological Society of London, Vol. II,
98:

1913. Problems of Genetics. New miia: Yale University Press,
viii + 258 pp.
Castle, W. E.
1916. Genetics and Eugenics. Cambridge: Harvard University Press,

pats ii E. and Wright, S
Two Color Mutations of Rats which Show Partial a
Science, N. S., Vol. XLII, No. 1075, Aug. 6, pp. 193-1

F.
1889. Natural Inheritance. London: Macmillan and Co., ix + 254 pp.
R.

1917. The mutation theory and the species T American Nat-

uralist, Vol. LI, No. 610, Oct., pp.
Goldman, E. A.

1910. Revision of the Wood Rats of the Genus Neotoma. Washing-
ton: Bureau of Biological Survey, ioia American Fauna
no. 31, 124 pp.

Goldschmidt, R.

1918. A preliminary report on some genetic experiments concerning

evolution. American Naturalist, Vol. LII, No. 613, Jan., pp.

28-50.
Hagedoorn, A. C., and A. L
1917. Rats and Evolution. AMERICAN NATURALIST, Vol. LI, no. 607,
o pp. 385-418.

Jennings, H.
1917. as Changes in Hereditary Characters in Relation to Evo-
lution. Journal a eta Academy of Sciences, Vol. VII,
lay 19, pp. 281-300. :
1917a. M Factors and Multiple Allelomorphs in Relation to

patio
the Results of Selection. AMERICAN NATURALIST, Vol. LI,
No. 605, May, pp. 301-306
Johannsen, W.
1913. — der exacten Erblichkeitslehre. (Zweite Auflage.)
Jena: Gustav Fischer, 723 pp.

454 THE AMERICAN NATURALIST [ Vou. LIT

To + EE
1911. The Influence of Heredity and Environment in Determining
the Coat Colors in Mice. Annals of the New York rar
of S tences, Vol. XXI, July 5, pp. 87-117, pls. VII-
1915. The Mee hanism of oe privar Heredity. New York: eal
olt and Co., ix + 2
-1917. The Theory of the ae Amas NATURALIST, Vol. LI,
o. 609, September, PP- 513
Nelson, E. W.
1909. The Rabbits of North America. Washington: quo of Bio-
logical Survey, North American Fauna, No. ‘29, p.
Osborn, H. F.
1915. Origin of Single Characters as Observed in Fossil and Living
Animals and Plants. AMERICAN NaTURALIST, Vol. XLIX,
No. 580, April, pp. 193-239. ;
Osgood, W. H.
1909. Revision of the Mice of the American Genus Peromyscus.
ashington: Bureau of Biological Survey, North American
Fauna, No. 28, 285 pp.

Pearl, R.
1911. Biometrie Ideals and Methods in Biology. Scientia, Vol. X,
pp. 9.
Pearson, K.
1900. The Grammar of Science. London: Adam and Charles Black,
548 pp.
Ridgway, R

1912. Color Standards and Color pa Washington: pub-
lished by the author, 43 pp.,
Sumner, F. B,
1915. ¡Some Studies of Faros Influence, Heredity, Correlation
PR rowth, in the White Mouse. Journal of Experimental
oology, Vol. 18, p 3, April, pp. 325-432.
1915a. fei Studies ‘of Several Geographic Races of California
Dee mbag ce. pera NATURALIST, Vol. XLIX, No. 587, No

er, Pp.
1917. The a of us: in the Formation of a el Localized
Race of Deer-mice (Perom sii oie a ee CAN NATURALIST,
Vol. LI, No. 603, March, pp.
1917a. Several Color ‘‘Mutations’’ in e fe the Genus Peromyscus.
Genetics, Vol. 2, May, pp. 291-300.
19176. Modern Conceptions of Heredity and Genetic Studies at the
Scripps Institution. Bulletin of the Scripps I ios for
Biological Research, No. 3, October 19, 24 pp.
Swarth, H. S.
1918. The Pacifice Coast jays of TOR genus, Aphelocoma. University
of California Publications in Zoology, Vol. 17, No. 13, Feb.
e : 23, pp. 405-422,

JOAN BAPTISTA PORTA

THEO. HOLM

BROOKLAND, D. C.

Like a cemetery with costly monuments for the rich,
modest wooden crosses for the poor, and for others sites
unmarked, hidden beneath brambles and weeds, a pic-
ture of death and oblivion—so history of botany has
dealt with records of the past, with life and labors
crowned with success or hopelessly ignored and forgot-

en.

For years, nay centuries unchallenged some works
have braved the everchanging hands of time, guiding
human thought into a highway with increasing light, con-
fronting nature, its laws and problems; great steps have
been taken forwards, new facts have been born, militat-
ing against former, old conceptions and resulting in com-
plete revolution. Coupled with intense sincerity great
skill has conquered, paving the way for future research,
culminating in success, or suddenly, without a warning,
crushed with defeat. Many a brilliant thought, but dis-
guised by a less powerful style, has remained obscure
and unnoticed, until at some proper time, as if surviving
itself, it has arisen and gained due homage, even though
late and in foreign soil.

Inclement fate has doomed to silence names of great
men, more fortunate thus than labors of merit that have
been misunderstood, carelessly weighed, and exposed in
unfavorable light. Who knows Porta? His work was
soon forgotten and in history it stands among those ridi-
culed or silently passed by. He was born in classic Italy,
in the middle of the sixteenth century, an era of scientific
research, marked by rapidly increasing interest in bot-
any, with splendid results laid down in precious volumes,
copiously and carefully illustrated. They were the days
of Cesalpino, Dodoneus, Conrad Gesner, Fuchs, Clusius,

455

456 THE AMERICAN NATURALIST [Von. LIT

Lobelius, Caspar Bauhin, all workers in botany, seeking
the same source for solving the problems of the plant
world, through demonstrating the relations between the
plants themselves, and beginning with classification first
of all. At that time literature was less than scant; there
was actually nothing to distract the views of investiga-
tors; it was an era of thought original, with room only
for the gifted and talented, none for the mediocre. And
strange to say, Cesalpino, though secluded from the bo-
tanical centers of Holland, France, and Germany, rose to
hold the palm as the most brilliant of his contemporaries.
In modern times he is still revered as the father of sys-
tematic botany. 'To these men Porta was not known, and
presumably they were not known to him either. To de-
scribe the principal episodes in the life of Porta, the
various phases of his character and labor, history has
little to say, fame less.

He was born in Naples, in the year 1545, and he did
succeed in gaining reputation as a noted naturalist, phi-
losopher, physician and pharmacologist. His home was
a favorite gathering place for men of learning; meetings
were held, dubbed ‘‘ Accademia dei Secreti,’’ and themes
were discussed delving into all the mysteries of nature,
principally the chimerical secrets of magic. That Porta
held an eminent place among his associates seems proven
by the fact that they regarded him as ‘‘a new prophet,”
and as such Porta was summoned to the court of Rome to
defend himself. He must have made a bold stand, tor
instead of meting out some punishment for his suspected,
supernatural power, the court exonerated him and elected
him a member of the Accademia dei Lincei. After that
time he lived in Rome for several years, and he died in
February 1615. The only botanical work written by
Porta is ‘‘Phytognomonica,”’’ published in Naples, 1588.
Three subsequent editions, 1591, 1608, and 1650, were pub-
lished in Germany. In the later years of his life Porta
acquired no small notoriety as an author of dramas and
tragedies. Considering the profuse material treated in
‘‘Phytognomonica,’’ and the fact that Porta was only in

Nos. 620-621] JOAN BAPTISTA PORTA 457

his forty-fourth year when the work appeared in print,
he must have begun his botanical career at a very early
age.

In this work, ‘‘Phytognomonica,’’ Porta puitecsdriaed a
new system for plants, but far different from those estab-
lished by his contemporaries or predecessors. His mind,
evilly influenced by the extravagancies of the Paracelsistes,
dwelt mostly upon such singular phases, remote from nat-
ural history, as similarity between parts of plants and or-
gans of man and animals, or the resemblance of parts of
plants with diseases of man and animals, furthermore
the habit or aspect of plants as being analogous to those
of man, and finally the relation of plants to the stars, the
sun, and the moon. Nevertheless Porta was a botanist,
and a very learned one. His studies of plants reveal
more than a superficial knowledge of their parts, and he
must have known many. But from beginning to end
the system, or better the method, proposed by him was
too enigmatic to conform with the requirements of nat-
ural science, founded as it was on principles so contrary
to nature as they possibly could be; and so the system
never reached beyond being considered the product of
““l'imagination brilliante mais déréglée.’”

It is, indeed, difficult to understand how a man so in-
tellectually gifted as Porta would ever waste his time
and labor on such problems as to demonstrate the secret
virtues of plants by merely observing the forms of their
parts and the color of their flowers. Thus according to
Porta certain species of Orchis with the roots palmate,
and grasses with the spikes in fives (Cynodon Dactylon)
would be a safe remedy for diseases in foot or hand, for
gout, ete.; plants with heartshaped roots or fruits (Va-
leriana, Persea) for heart disease; plants with the flowers
resembling eyes (Aster, Sedum) for eye diseases. Fur-
thermore, plants with spotted stems (Aracee) would
on account of their likeness to snake-skin be useful as

1 Compare Planchon, J. E., ‘‘Des limites de la concordance entre les
formes, la structure, les affinités des plantes et leurs propriétés medicinales,’’
Thèse, Montpellier, 1851.

458 THE AMERICAN NATURALIST [ Vou. LII

antidote for snake bites, ete. Forty-two sections of this
type are described by Porta, and some are fairly well
illustrated. Much attention is given to citing and ex-
plaining descriptions and names of species known to and
mentioned by the old authors, Pliny, Dioscorides, Colu-
mella, and others, and from this particular viewpoint the
book is quite interesting and useful.

But even if the greater part of this book is devoted to
considerations of the secret virtues of plants, some chap-
ters and remarks, scattered here and there, reveal the
indisputable talent of Porta as an observer of plant life.
To do full justice to this part of his work let us briefly
consider the status of botany in the sixteenth century.
It was an era of classification or attempted classification;
the plants were described and arranged in some way as
an expression of their mutual relationship. By Bock
(1560) they were divided into herbs, shrubs, and trees;
by Clusius (1576) the system became enlarged so as to
comprise bulbous plants, plants with the flower fragrant
or inodorous, plants with milky juice, ete. The descrip-
tions furnished by Clusius have always been regarded
as most excellent, but he gave much more attention to
the foliar structure than to the floral; in this point of
view he was followed by Lobelius and Dodoneus. Cas-
par Bauhin (1550-1624) established a system covering
twelve books, and he began with the grasses and grass-
like plants, including Iris, Acorus, ete.; after these came
the bulbous plants, then those with large, edible roots,
ete. ; the genera were not described, only the species with
a number of synonyms. Bauhin was the earliest author
to use binary names; but in describing the plants he did
not consider the structure of the flower, nor the fruit.
Finally Cesalpino (1583) not only established a system
principally based upon the organs of fructification, hith-
erto ignored, but he added a large number of new obser-
vations of great importance to the study of botany. The
introduction to his work contains a discussion of theo-
retical botany in general. With regard to his classifica-
tion of the plants, into arborescent and herbaceous, the

Nos. 620-621] JOAN BAPTISTA PORTA 459

minor groups, characterized by the structure of fruit and
seed, are not natural, except the sixth, which comprises
the Umbellifere, the tenth, Boraginex and Labiate, and
the fifteenth, plants destitute of flowers and fruits, ferns,
mosses, and fungi.

Naturally the tendency to classify governed botanical
research during as early a period as the sixteenth cen-
tury, and the various systems proposed were all purely
artificial. Not until the year 1703 were the Phanero-
gams distinguished as mono- and dicotyledonous by Ray.
We remember that so late as the middle of the eighteenth
century Linné established his artificial system, based
upon the floral structure, and at about the same time
Antoine Laurent de Jussieu undertook the task to de-
scribe the families. In other words, Bauhin wrote the
diagnoses of the species, Tournefort (about 1700) char-
acterized the genera, Linné arranged the genera in
groups, which he named only, and finally Jussieu estab-
lished a natural system with family-diagnoses. But re-
turning to the sixteenth century, the actual knowledge
of the plants was embodied in systems, and beyond the
mere classification no attempts were made to consider
the plants from a biological viewpoint, as members of a
living world adapted to environments of highly different
nature as to climate and soil, at least not in accordance
with the history of botany. The treatment of this par-
ticular phase of plant life was reserved to the very close
of the nineteenth century, when Warming? introduced a
supposed new branch of botanical science, dealing with
plant societies, now universally recognized as plant ecol-
ogy. The appearance of this work has brought about a
fuller valuation of the factors that govern plant life, a
purely biologic consideration of the plants on morpho-
logical and physiological grounds. However, twenty
years of experience has taught us that this branch of
botanical science lacks organization and is yet rather to
be compared with a speculation having run far in ad-
vance of facts. Nevertheless the supposed new doctrine

2**Plantesamfund,’’ Kjebenhavn, 1895.

460 THE AMERICAN NATURALIST [ Von. LIT

does exist, and has existed for more than three centuries,
founded by Porta, and amply discussed in several chap-
ters of his *“Phytognomonica.”? No mention is made by
Warming of Porta’s book. And, strange to say, no men-
tion is made of Planchon’s either. By Sachs3 Porta is
passed by in silence. —

Planchon (l. c.), Guy de la Brosse* and Adanson® refer
only to the part dealing with the secret virtues of plants;
the chapters on plant societies are ignored, or let us say
not appreciated.

To Porta the method of classifying plants as instituted
by his contemporaries must have been absolutely un-
known; of rendering the knowledge of plants more ac-
cessible by means of a system he had no thought. His
principal object was to demonstrate the virtues or prop-
erties possessed by plants, and, as stated above, Porta
combined these with the general aspect of plants, the
shape of their leaves, stems, etc. While making these
observations in the field, as he did, Porta became aware
of the distribution of a number of species under condi-
tions very variable, and especially with regard to the
soil. He noticed the fact that the general aspect of the
plants, their shape, color, odor, hairiness or smoothness,
at least to some extent, depended upon the environment
in which they grew, and from this point of view, we
might say ‘‘biologic,’’ did Porta elaborate the introduc-
tion to his ‘‘Phytognomonica.’’ He divided the plants
in two groups, aquatic and terrestrial, each with several
subdivisions. Of the former, examples are given of
species characteristic of lakes, Swamps, rivers, brackish
marshes, etc., and he described the habit of several plants,
in most cases very correctly. With respect to the ter-
restrial plants Porta distinguished between those that
occupy a rich, a dry, or a sandy soil, illustrated by Malva,

Lithospermum, and Feniculum. Furthermore, some am-
phibious species are described, such as are terrestrial but

3**Geschichte der Botanik,’’ Miinchen, 1875,

#**De la nature des plantes, >?

5 **Familles des plantes,’’ Vol. I, préf., p. ii, 1763.

Nos. 620-621] JOAN BAPTISTA PORTA 461

adapted also to live in the water. In several chapters
descriptions are given of the habits of plants, of species
characteristic of the mountains, the lowlands, the hill-
sides, and the shady valleys; to these were added some
brief remarks upon the vegetation of the northern, the
temperate, and the torrid zones. Among the cultivated
plants Porta mentions Zea, which, however, is not the
plant known now under this name (maize), but a kind of
wheat (Triticum Spelta) as demonstrated by De Can-
dolle. In bringing these facts together Porta certainly
laid the foundation of plant ecology, and the classifica-
tion, proposed by Warming (1. c.), of the various plant
societies: ‘‘Hydrophytes, Xerophytes, Halophytes and
Mesophytes’’ is not much more instructive than the one
introduced by Porta: ‘‘plante palustres, fluviatiles, mar-
inæ, salse aque, silvestres,”” ete.

Naturally these groups have received a more elaborate
treatment by authors of a recent date, especially with
reference to the internal structure, which often, but very
far from always, is in correlation with the respective en-
vironment. However, the weakness of modern ecology
rests on the belief that the structures may be explained
as caused by the natural surroundings. Experience has
taught that the genera and species do possess some char-
acter of their own, which they never give up. To a cer-
tain limit the plants may allow themselves to submit to
changes, but beyond that they will sooner die.

So far as Porta considered the biologic question of
plant life, dealing only with the superficial aspect, more
or less comparable to the surroundings, he committed no
errors of consequence. For as a matter of fact the prin-
cipal features exhibited by members of plant societies
are mainly external, such as the shape of leaves, their
relative size, the organs of vegetative reproduction, and
the general habit; the internal structure cannot be relied
upon, at least not at the present stage of our knowledge
of plant life.

Thus already in the sixteenth century the first essay on
plant ecology appeared, and Porta was the author. |

SHORTER ARTICLES AND DISCUSSION

AN AUTOSOMAL BRISTLE MODIFIER, AFFECTING A
SEX-LINKED CHARACTER

A RECESSIVE gene in the third chromosome of Drosophila mela-
nogaster (ampelophila) affects the bristles on the thorax and
seutellum of females which are heterozygous for a recessive sex-
linked character, forked in such a way as to make forked semi-
dominant. This latter character has been described by Morgan
and Bridges (Carnegie Publ. 237). The bristles of ‘‘stock’’
forked flies are shortened, twisted, and heavier than normal.
This applies to the bristles of the head, thorax and scutellum.
Flies heterozygous for forked, that is, females, since the gene is
in the X chromosome, have normal bristles unless the fly is also
homozygous for the third chromosome modifier here recorded.
Females with one forked gene and one normal allelomorph,
which are homozygous for a modifying gene in the third chromo-
some, are intermediate in appearance between forked and normal
flies, and are designated as ‘‘semiforked.’’ The males never
show the character since they are never heterozygous for sex-
linked factors. The forked appearance is limited to a few of
the thoracic and scutellar bristles and the bristles in general are
less affected than are those of the homozygous forked flies as
regards thickening of bristles and twisting. Both of the third
chromosomes must bear the modifying gene in order to affect
the bristles. In the absence of the forked gene, the semiforked
genes, even when homozygous, are nearly always without effect,
but occasionally a few individuals may be detected which have
shorter and heavier bristles, but this is not pronounced and is
rarely found. Flies which are known to be pure both for forked
and for the modifier, semiforked, can not be distinguished from
the simple forked individuals without the modifier.

ORIGIN OF SEMIFORKED

The semiforked character was first observed in February,
1918, in the heterozygous Bar forked daughters resulting from
culture 668, a cross of a Bar male from stock to a forked female
of a non-disjunction strain which had been used to observe non-

462

Nos. 620-621] SHORTER ARTICLES AND DISCUSSIONS 463

disjunction for about three months previous to this mating.
The semiforked females were not noticed until the bottle was
half through hatching, probably being overlooked. The counts
show that after the new character was found there were 15 such
females and 42 of the expected normal bristle type. This is
clearly a 3:1 ratio and both parents must have been heterozy-
gous for the semiforked gene. Although they were not brother
and sister, this is not improbable, because the gene seems to have
existed originally in the Bar stock, which, however, was not pure
for it. The forked female doubtless obtained the gene from the
Bar stock also, as her pedigree contains many Bar stock males
used in the non-disjunction experiments. Since attention was
paid only to the behavior of the X chromosomes, it is easy to see
that the autosomes would be interchanged from generation to
generation and the forked female could have a third chromo-
some which came originally from the Bar stock.

The strain was kept going by brother-sister matings. One
F, culture (712) of a forked male to a heterozygous Bar forked
female, which was semiforked, produced all the heterozygous
forked: females with semiforked bristles. Here both parents
were pure for the modifier. In Culture 722, which was an F,
from 668, half of the females heterozygous for forked were semi-
forked. In this case, one parent was pure and the other hetero-
zygous for the modifier. One case was observed where a forked
male crossed to a semiforked female produced no semiforked
daughters. - The explanation is that both the third chromosomes
of the father carried the normal genes. The reverse case of
this was shown when a forked male was crossed to a hetero-

gous Bar forked female with normal bristles. Of the hetero-
zygous forked females produced, approximately half were semi-
forked and half normal bristled. Here the father was pure for
the modifier but the mother was heterozygous for it and the 1:1
ratio resulted.

Location OF MODIFIER

The presence of the modifying gene in the third chromosome
was demonstrated by the following method, which has been used
before in work on Drosophila. A semiforked female was out-
crossed to a star dichete male from stock. Forked star dichete
males were selected from the offspring and back-crossed to the
semiforked females from stock cultures. Star and dichete are
dominants in the second and third chromosomes respectively

464 THE AMERICAN NATURALIST [ Vou. LII

and are used because they can be easily detected in the hetero-
zygous condition. Since there is no crossing over in the male in
melanogaster, any dichete fly in the offspring of the back-cross
must have obtained one third chromosome from the dichæte
stock and one from the semiforked stock. Any fly which was
not dichete traces back both its third chromosomes to the semi-
forked stock.

Examination of the offspring from the back-cross in three
cultures showed that no dichete fly was ever semiforked and,
conversely, all not-dichete females were semiforked, provided
they were heterozygous for forked, and this includes all those
not homozygous forked. About 500 individuals were obtained
from these three cultures and the above statement is based upon
them. The result is absolutely clear-cut and shows that the
modifying factor is recessive and in the third chromosome. The
` presence of the star chromosome (II) did not affect the appear-
ance of the semiforked character in any way. The location of
the gene within the chromosome by its linkage relations to other
third chromosome genes has not been carried out.

SUMMARY

1. A recessive third chromosome modifying gene converts
heterozygous forked females into intermediate semiforked in-
dividuals.

2. Homozygous forked flies are not visibly affected by the
modifier.

3. The semiforked modifier rarely produces any visible effect
when homozygous unless the forked gene is present.

D. E. LANCEFIELD
COLUMBIA UNIVERSITY

THE
AMERICAN NATURALIST

Vou. LII. October-November, 1918 Nos. 622-623

MIGRATION AS A FACTOR IN EVOLUTION:
ITS ECOLOGICAL DYNAMICS!

CHARLES C. ADAMS
PROFESSOR OF Forest Zo0LOGY, THE New YORK STATE COLLEGE OF
Forestry AT SYRACUSE UNIVERSITY

CONTENTS

. INTRODUCTION,
. THE PROCESS METHOD OF ANIMAL RESPONSES.

=
paj

d. Interaction of Sys
111... THE ormon FACTORS IN pan
1. Introduet:
Atmospheric Agencies in 1. Transportation.
n

2.
3
A, Lithospherie Agencies in Transportation.
5. ee Agencies in Transportation and Migration.
. Plants in nis open
>. Animal Migrat:
IV. SUMMARY AND CONCLUSIONS.
V. BIBLIOGRAPHY.

I. IxtroDUCTION

My subject, ‘‘ Migration as a Factor in Evolution,”” is,
in other words, the function or róle of migration in evo-
is paper was prepared, by invitation, for the symposium on the Factors
of O at the Pittsburgh meeting of the rir REE Te of Natural-
ists and an abstract of this paper, entitled ‘‘ Migration as a Factor in Evo-
lution,’’ was read January 1,1918. Dr. G. H. Shull, peada of the society,
kindly consented to its publication in advance of the volume which is

to contain all the papers of the symposium.

465

466 THE AMERICAN NATURALIST [Vou. LII

lution.? In view of the recent concentration of interest
on heredity, my subject has the flavor of an old-fashioned
one, which calls back to the days when Darwin and Wal-
lace were living, and when Wallace’s ‘‘Island Life’’ was
frequently read with enthusiasm, and when there was
possibly a more general belief that natural selection was
one of the large factors in evolution. But progress has
not been limited to the studies of heredity, for with the
rapid rise of certain phases of general physiology, ani-
mal behavior and animal ecology, a newer orientation is
now possible with regard to the migration of animals by
both the active and passive methods. For our knowledge
of animal responses, as well as the influence of the vege-
tational and physical environment, have made consider-
able progress, and we now probably see more clearly than
ever before the intimate relation existing between the
animals and the conditions which influence their migra-
tions. The present occasion has thus furnished an op-
portunity to make a preliminary reorganization of the
accumulated materials from a somewhat different stand-
point than was formerly current in discussing migration.
And although some of these ideas are widespread and
even commonplace in certain limited fields, yet they are
not yet in as general use as is desirable, and they are in
urgent need of extended application and critical study.
In the following discussion of migration as an evolu-
tionary factor, I wish to emphasize two points in particu-
lar. One is the discussion of the process of analysis and
the other is to suggest some methods of applying this
method to the problem of migration. It may seem aside
from the main thesis to give this emphasis to the process
method of evolution, but after striving several years for
the conscious application of this method in an allied field
(Adams ’13, 715), and seeing its beneficial results there,
2 By evolution, I mean to use the term in the broadest possible sense—to
include all changes within the organism and its necessary environmental im-
plications, and not as limited to the ‘species problem.’’ This term must
be as thoroughgoing as metabolism in physiology, or as metamorphism as
ated e rocks by Van Hise, ‘‘any change in the constitution of any kind

Nos. 622-623] MIGRATION A FACTOR IN EVOLUTION 467

and furthermore, not having seen this formulated and
applied to evolution as here presented, I feel that the im-
portance of the subject merits this treatment.

Special attention is called to the fact that this discus-
sion is not intended as a complete, scientific explanation
of migrational facts, but as the presentation of a point
of view, or working hypothesis, which it is believed will
aid in explaining many well-known facts and relations,
and will aid in the discovery of new ones. An effort has
been made to frame this hypothesis in such a manner as
not to prejudice the problems investigated, or to inter-
fere seriously with the various constructive schools of
investigation, although I am well aware that this hy-
pothesis, like all others, is built upon certain assumptions.

During the preparation of this paper (which amplifies
certain ideas which I have previously outlined) I was
much impressed by finding so much confirmatory evi-
dence of the general validity of the dynamic standpoint,
in fields relatively remote from migration. The inde-
pendent growth of such conceptions in diverse fields is
indicative that many subjects are independently reaching
a certain common stage of development and that spon-
taneously such ideas are becoming independent organiz-
ing centers of activity. By interaction and regulation
among these ideas, new higher systems of unity and cor-
relation are developing, which are producing important
effects in zoology as well as in other sciences. The slow-
ness seen in the application of dynamic ideas to biology
is perhaps rather natural as is evident when we recall the
fact that even in the simpler physical sciences we have
as yet no complete dynamical theory or system, although
much progress has been made, not, however, toward a
complete system, but toward that dynamic equilibrium
which characterizes a growing subject.

Il. Tue Process METHOD or Anrmat RESPONSES
1. Introduction

The fundamental assumption upon which this discus-
sion is founded is that the animal should be looked upon

468 THE AMERICAN NATURALIST [Vou. LIH

as an entity or agent whose system of activity or re-
sponses to internal and external influences are its most
fundamental characteristic. The activity of the animal,
as an agent, is its process of change or its process of
activity. Broadly speaking, this is a study of influences,
of response or behavior, a study of what animals do and
how they do it. Throughout it is the dynamic aspect of
the animal; the pressure it exerts upon the environment;
and the pressure exerted by the environment on it, that
-is of greatest importance. I see no reason why these
assumptions can not be of universal application, and why
they can not be accepted in all investigation. This view
appears to be so well established that no detailed evi-
dence and discussion of it seems necessary at this time.
The animal agent itself is not a fixed thing, but one which
runs through a cycle; it originates, develops and disinte-
grates and is thus in its maintenance subject to all the ebb
and flow of other processes, and has similar dynamic rela-
tions. There is thus valid reason for assuming a
thoroughgoing process or dynamic program for dealing
with all animal problems. The same is equally true of
all plants which form a part of the animal environment,
and the physical environment lends itself, in fact, easily
leads, in such a treatment, and fits into this scheme har-
moniously, and makes it possible to give not only a uni-
form treatment to all phases of animal relations, but en-
ables the student of animals to make a perfect contact
with all the allied sciences, and to draw from each one
all possible support, with the least possible friction and
interference.
2. Dynamic Principles

(a) Activity of Agent.—In discussing animal problems
from the process standpoint there are several concep-
tions which are fundamental. These ideas can be illus-
trated in simple form by an example from physical sci-
ence. Running water is a substance combined with
energy (gravitation) which exerts stress or pressure,
which it expends upon other substances, and it is there-
fore an agent. An agent thus exerting stress and expend-

Nos. 622-623] MIGRATION A FACTOR IN EVOLUTION 469

ing gravitational energy upon other substances is in the
process of activity. Thus, running water, by the general
process of erosion, including the subsidiary processes of
weathering and transportation, wears down the land and
results in the formation of many features such as brooks,
creeks, plains, deltas, and a variety of other physio-
graphic products. An organism is also an agent which
expends physical and chemical energy, producing stress
and exerting pressure and expending energy on other
substances, exhibits its process of response or its process
of behavior. An animal, by the process of predation runs
down another animal and devours it, by its process of di-
gestion dissolves it, and by the process of assimilation
makes muscle, bone, feathers or fur out of it, and these
are all products of its activity. The process of response
is here strictly comparable to the process of erosion of
running water, and their products are similarly compar-
able. The general process is generic and includes many
species and varieties of subsidiary processes, ad infinitum.
As Keyes (’98) has well said, ‘‘Processes are merely
operative. If coupled with products at all... they must '
be regarded as formative or constructive. The product’s
destruction, its loss of identity, is wholly immaterial.
The action of agencies is merely to produce constant
change.” It is, therefore, to the process of living, to the
process of evolution, rather than to its products such as
species, varieties, etc., which are of fundamental impor-
tance. For this reason the products must be subor-
dinated to the agencies and processes, because the laws
of change are in reality the object sought.

The physiographer is not content to rest with the idea
that the agent, as, for example, running water, is the fin-
ished product of his analysis, for he also applies the same
methods of investigation to the agent itself, in order to
know its method of origin, the process by which it origi-
nates in the stream, whether indirectly from a spring, or
directly from the clouds. Thus the same methods which
the physiographer uses in studying the activity of the
agent, he again uses to explain the origin or derivation

470 THE AMERICAN NATURALIST [Vou. LII

of the agent iself. The investigator of animals follows
the same plan. He likewise uses the same kind of meth-
ods in the investigation of the processes of functional
and structural development, not only as applied to the
actions of the agent, but to its origin as well, and thus
again we are justified in concluding that the process
method is of universal application.

(b) Cycle of Activity.—The activity of agents is always
accompanied by the expenditure of energy. This expen-
diture does not take place at a uniform rate, there is a
pulsation, an ebb and flow, a rising and a falling. Periods
of activity are followed by periods of repose and a
rhythm is seen which can often be resolved into cycles.
The importance of determining such cycles has been well
expressed by Lockyer? as follows:

Surely in meteorology, as in astronomy, the thing to hunt down is a
eyele, and if that is not to be found in the temperate zone, then go to
the frigid zones, or the torrid zones and look for it, and if found, then
above all things, and in whatever manner, lay hold of, study it, record
it, and see what it means. If there is no cycle, then despair for a time
_ if you will, but yet plant firmly your science on a physical basis... and |
having gotten such a basis as this, wait for results.

There are innumerable cycles in the responses of ani-
mals, and of these the life-history cycle is perhaps the
most generally recognized; but activity and response,
hunger and satiety, stimulation and response, are other
familiar expressions of these conditions. During these
cycles of change the relative amount of energy set free
varies greatly, in other words, its dynamic status
changes. As expressed elsewhere, I have stated (715,
p. 10):

lated, its normal activities are interfered with, and a physiological con-
dition of stress is produced which lasts until, by repeated responses or
3 ‘í Solar Physics,’ ? 1874, pp. 424-425,

Nos. 622-623] MIGRATION A FACTOR IN EVOLUTION 471

“trials,” the animal escapes stimulation or succumbs and a relative
equilibrium is established. An area becomes overpopulated and conse-
quently there may be established a condition of stress which results in
an adjustment by a reduction (through many causes) in | the excess of

of relative equilibrium. The cycle of change may be considered to begin
at any point. I have taken as the initial stage of the cycle the condi-
tion of stress and pressure and have indicated how this condition tends
to change in response to pressure, bringing about the process of adjust-
ment to strain, and leading to the condition of adjustment to strain, or

condition of stress, the process of adjustment to the strain follows, and
this leads to the Ta establishment of the condition of adjust-
ment or of relative equilibriu

These ideas can be si applied to the life eycle, as is
indicated from the following statement by Sedgwick
(710, p. 177), who says:

The life-cycle, of which the embryonic and larval periods are a part,
consists of the orderly interaction between the organism and its environ-
ment. The action of environment produces certain morphological
changes in its organism. These changes enable the organism to come
into relation with new external forces, to move into what is practically
a new environment, which in its turn produces further structural
changes in the organism. These in turn enable, indeed necessitate, the
organism to move again into a new environment, and so the process con-
tinues until the structural changes are of such a nature that the organism
is unable to adapt itself to the environment in which it finds itself. The
essential ete of success in this process is that the organism should
always shift into the environment to which the new structure is suited—
any failure in pe leading to impairment of the organism. In mo
cases the shifting of the environment is a very gradual process (whether
consisting in the very slight and gradual alteration in the relation of
the embryo as a whole to the egg-shell or uterine wall, or in the relations
of its parts to each other, or in the successive phases of adult life),
and the morphological changes in connection with each step of it are
but slight. But in some cases jumps are made such as we find in the
phenomena known as hatching, birth, and metamorphosis. . . . And with
this property of reacting to the environment goes the further property
of undergoing a change which alters the relation of the organism to the
old environment and places it in a new environment.

It is seldom indeed that one finds ontogeny so clearly

expressed in terms of an active agent which is ünder-

f

472 THE AMERICAN NATURALIST [Vou. LH

going a cycle of changes—both in structure and function
—and is being stimulated, responding, behaving, and
even migrating into new environments, in response to
internal and external stimulation. This is indicative of
the dawn of a new era in the study of ontology (ef.
Thompson, *17). As Bancroft (’11, p. 178) suggested,
Sedgwick (*10, p. 177) saw clearly for the moment, as it
were, but not in practice and concretely, the dynamie
conception of individual development, although it is very
evident that he saw the unity or continuity of the on-
togenetic cycle. However, it has remained for Child (715,
"15a, *15b) who, apparently adapting largely the dy-
namic conceptions of the plant and animal ecologists, and
to a lesser degree those of the physiologists, has now
given expression, in a clear and concrete manner, to the
dynamic ideas in individual developmental responses, and
special attention is called to his important work.

That the life cycle varies in its degree of susceptibility
to environmental influence has been pointed out by Ver-
non (’99, p. 199), DeVries (1900), Bancroft CIE p 179)
and others (Woods, ’10; Pike and Soott, LOs Puto: 17).
Vernon’s law is expressed (’03, p. 199) as follows: ‘‘In
fact, it would seem to be a law of general application that
the permanent effect of environment on the growth of a
developing organism diminishes rapidly and regularly
from the time of impregnation onwards.” Bancroft was
the first to see that his law included Vernon’s. He said
CIL p. 175):

We know that, as we get older, our tendeney to resist change in-
creases; our habits of body and mind become more fixed.* We should
therefore be tempted to conclude that the resistance to change increases
as the organism becomes mature and that a given stimulus would prob-
ably have the most effect if applied at or before the earliest stages of
development.

4 In this connection it is interesting to recall the influence which this law
may have upon scientifie research. Once Clerk-Maxwell wrote to Herbert
Spencer about some point in his “First Principles’’ as follows: ‘‘It is
seldom that any man who tries to form a system can prevent the system
from forming around him; and closing him in before he is forty. Hence
me ingredient to prevent erystallization and
condition.’? (Footnote by C. C., A.)

Nos. 622-623] MIGRATION A FACTOR IN EVOLUTION 473

He then quotes Vernon and continues:

If the pressure on a liquid is made less than the vapor pressure for
that liquid at that temperature, some of the liquid vaporizes, the tem-
perature falls, and the liquid may be said to adapt itself to the new con-
ditions. What would happen if the liquid were not adaptable? The
easiest way to obtain non-adaptable liquid is to place a Bunsen burner
under it. The temperature rises until the boiling point is reached. The
liquid then ceases to be adaptable. It volatilizes, it disappears, it be-
comes extinct so far as that particular region or flask is concerned. If
a species can not adapt itself to changed climate or other conditions, it
does not volatilize; but it disappears, it becomes extinct. It may be a
new point of view to consider the extinction of the mastodon as anal-
ogous to the distillation of water; but the two cases are really parallel,
except in time.

These facts are of the greatest importance because
they indicate the critical stage or condition at which, in
the migration of animals into new localities and condi-
tions, organisms are most likely to be modified, and thus
influence their evolution. This furnishes a new reason
for stressing the importance of the breeding conditions
and habitat in ecology.

An exception to Vernon’s law is to be seen in the case
of the Protozoans (and probably to other kinds of non-
sexual reproduction), as is indicated by Jennings (712,
113) and Woodruff’s investigations, which show that in
a proper environment the inertia of the life cycle tends to
continue on indefinitely and doesnot rundown. Jennings
(712, pp. 573-574) shows that conjugation in Protozoa
and sexual reproduction in the metazoa cause diverse and
new combinations of characters. In other words, this
means that the processes of conjugation (favored by ad-
verse conditions) and sexual reproduction, tends to break
up the stability and crystallization into which the on-
togenetic system tends to develop, and tends to restore,
as it were, flexibility and a colloidal state to the race.
This changes the system so as to minimize the interfer-
ence with its processes. In the metazoan the number of
systems is so large, that in spite of its chemical integra-
tion and regulation, interference with one or more of

474 THE AMERICAN NATURALIST [Von. LII

them, possibly the ‘‘slowest,’’ limits action and causes
* death. As indicated later, according to the phase rule
the greater the number of ‘‘phases’’ interacting the
lesser the number of possibilities of change. This is
not a condition limited to organisms, but is a general law.
It is perhaps in some sense as this that we can concede
““differentiation”” as a cause of death.

The animal as an agent, or individual, behaves accord-
ing to its own system, to the extent that it is an independ-
ent unit, and these activities are cyclical. All systems
tend to perpetuate themselves. Bancroft’s law for all
systems is that: “The broadest definition of it is that a
system tends to change so as to minimize an external dis-
turbance.’’ In other words this is a perpetuating tend-
ency, a method of assimilation, of which reproduction
may be considered but a special phase; it is not solely a
peculiarity of organisms, as is often stated, but of all
systems. Sedgwick (*10, p. 177) has said: ‘‘It is a prop-
erty of living matter to react in a remarkable way to ex-
ternal forces without undergoing destruction. . . . This
property of reacting to the environment without under-
going destruction is, as has been stated, a fundamental
property of organisms.’’ In these features the animal
acts only as other systems tend; as a catalyzer, it hastens
changes and maintains itself. The activity of the animal,
its centrifugal stress, causes it to collide with its environ-
ment, while, on the other hand, there is the environmental
bombardment, both of which, within certain limits, tend
to interfere or destroy the animal. On the other hand,
the tendency of the system is to ‘‘minimize disturbance,’’
to change within, to minimize, to ‘‘retreat’’ from inter-
ference (absolutely or relatively), and in this manner to
a large degree, the system is perpetuated. To be sure,
many individuals perish, but the system of the species
continues. The rate of change of the system can be modi-
fied only as fast as its slowest member can change, and
on this account many individual systems are destroyed.

In addition to influences which ‘‘interfere’’ with sys-

Nos. 622-623] MIGRATION A FACTOR IN EVOLUTION 475

tems, as expressed by Bancroft, there are those which
reinforce or accelerate (tend to continue or hasten activ-
ity) and do not change its character, but only the inten-
sity of the response (temperature, enzymes, repetition,
etc.). By this method also systems tend to be perpetu-
ated, and organisms in ‘‘favorable’’ (non-interfering)
conditions, tend to continue their normal activities.”
- This law appears to be a corollary of Bancroft’s law
which is concerned with interference or retardation.
Thus when a system is reinforced, rather than disturbed,
the system continues onward in its normal cycle without
interference, and may even be accelerated in its activities.
This is a condition which may maintain a relative equi-
librium, or increase stress. The intensity of interference,
or reinforcement, and its repetition, hastens or retards
the rate of change of a system. We thus have the quali-
tative and quantitative relations applying to the law of
reinforcement or acceleration of the equilibrium, and
Bancroft’s law of interference with its development.

Even relatively fixed and automatic responses of be-
havior may be looked upon much as the relatively stable -
structural characters, so that every sort of behavior, even
to the process of higher learning, shows this regulatory
influence which tends to change in such a manner as to
eliminate all disturbance with its systems, even to the in-
consistencies of our ideals.

Thorndike ('11, p. 244) in summarizing the laws of
““acquired behavior or learning’’ formulates two laws.
The first is essentially a statement of Bancroft’s law. of
response to interference (discomfort or satisfaction),
and the second (exercise or repetition), is that of rein-
forcement. This means that the kind and intensity of
stimulation, and its repetition are the laws of establish-
ing associations, or of changing the system, and that in-
tensity and repetition act as the catalyzers which influ-
ence the speed of modification of the system; at bottom it
therefore appears we have qualitative and a quantitative

5 Cf. Jennings, ’06, p. 295, ‘‘ positive reactions.’’

476 THE AMERICAN NATURALIST [VoL. LIT

expression of them. Limiting factors, because of their
intensity and repetition, tend to change the animal sys-
tem so as to minimize the external disturbance, or the
animal system tends to change in such a manner as to
minimize external disturbance, at a speed determined by
the intensity and repetition of the disturbance. The so-
called ‘‘trial and error,”” or, better, trial method of be-
havior, is also an independent formulation of Bancroft’s
law. It seems probable that the modifiability of be-
havior, and even all methods of animal regulation, are
expressions of these laws of interaction.

The ‘‘balance of nature’’ is a culminating phase of the
eycle of adjustment to strain. As expressed elsewhere, I
have said (Adams, 715, p. 14): “When a balanced con-
dition, or relative equilibrium, in nature is referred to,
we must not assume that all balances are alike, for some
are disturbed with little effort and others are exceedingly
difficult to change. This distinction is an important one.
Once the balance is disturbed, the process of readjust-
ment begins. This is a phase in the balancing of a com-
plex of forces. Just what stages this process will pass
through will depend, to an important degree, upon the
extent of the disturbance. Slight disturbances are tak-
ing place all the time and grade imperceptibly into the
normal process of maintenance, as when a tree dies in the
forest and its neighbors or suppressed trees expand and
take possession of the vacancy thus formed. Disturb-
ances of a greater degree, on the other hand, may only be
adjusted by a long cumulative process. This change can
progress no faster than the rate at which its slowest
member can advance. Thus a forest association of ani-
mals may be destroyed by a fire so severe that all the lit-
ter and humus of the forest floor is burned. The animals
which live in the moist humic layer as a habitat, such as
many land snails, diplopods, and certain insects, can not
maintain themselves upon a mineral soil, rock or clay.
As such a forest area becomes reforested, these animals
can only find the optimum conditions when the slow proc-

Nos. 622-623] MIGRATION A FACTOR IN EVOLUTION 417

ess of humus formation reaches a certain degree of cumu-
lative development. Under such circumstances this later
stage must be preceded by antecedent processes, and the
restoration of the balance is long delayed. Some adjust-
ments take place so quickly that little can be learned of
the stages through which they pass. There are, however,
many slow processes which afford an abundance of time
for study; in fact some are too slow to study during a life
time. The processes which are moderately slow are often
particularly illuminating because all stages are fre-
quently so well preserved that comparison is a very use-
ful method of study; the slowness of a process has a cer-
tain resolving power, as it were, recalling the influence of
a prism upon a beam of white light, which reveals many
characteristics obscure to direct vision. A study of the
processes of adjustment among animals is a study of an
important phase of the problem of maintenance. The
continued process of response will, if circumstances per-
mit, lead to a condition of relative adjustment, or bal-
ancing among all the factors in operation.”? The de-
termination of the dynamic status and its application to
cycles is seen to be a method or criterion which may be
used for the determination of cycles of activity, and the
repetition of these determinations will indicate the direc-
tion of movement of a process, and thus serve as a guide
in the determination of its rate of change.

(c) Limiting Factors.—Animals live in a real world,
they are dependent upon an environment and they can
not be understood independently of it. They do not live,
as it were, ina vacuum. As Brooks (’02, p. 485) has said:
‘‘No physiologist who studies the waste and repair of
living bodies, no naturalist who knows living beings in
their homes, no embryologist who studies the influence of
external conditions upon development, can, for an in-
stant, admit that living beings are self-sufficient or self-
sustaining, or that their being is in themselves; for the
line we draw, for our study, between living beings and
the external world is not one we find in nature, but one

478 THE AMERICAN NATURALIST [VoL. LII

that we make for our own purposes.’’ We have seen that
the essence of the animal is its activity. Its life is a
continuous collision with the environment and a bombard-
ment by the environment, with changes which tend to re-
lieve the disturbances. This is particularly true of free-
living animals, and is indirectly so, even of sedentary and
sessile kinds. This radiating activity of the animal, and
the direct convergent influence of the environment on the
animal, is the basis for the friction and interaction which
exists between the organism and the environment. There
are, therefore, definite zones of influence and stimulation
about the normal or attuned environment of the animal,
and with departure from these conditions locally and
geographically there are certain definite results (Adams,
04, p. 211):

The new vital conditions are a cause of stimulation and with further
departure (beyond a certain limit) it leads to increased stimulation or
to unfavorable conditions. This results in retarded growth, development,
and reproduction of the organism as a whole. Thus the end results of
extreme departure from the optimum in either direction are similar.

(Adams, 715, pp. 8-9) :

Thus departure from the optimum toward an increase or a decrease,
are departures from the most favorable, conditions toward less favor-
able conditions and hence toward limiting conditions. ... In nature we
look upon the optimum as that complex of habitat factors which is most
favorable, and departure in any direction from this optimum intensity
is in the direction of a less favorable degree of intensity, or into unfavor-
able conditions. From this standpoint any unfavorable condition is a

limiting factor and may retard, hasten, or prevent vital and ecological
activities.

(Adams, *13, p. 98) :

The similar results of extremes of high and low... temperatures,
aridity, and the lack of oxygen may be cited as examples. Such effeets
have an important bearing upon the subject of physical and chemical
limiting factors which influence individuals. [Cf. Shelford, *11, pp-
598-599. ]

I would now modify my preceding quotations so as to

_ definitely discard the old idea of the optimum,* in harmony

6 This word has a general utility, but its technical value, like that of the

““normal, > both long considered peculiar to organic response, appears to
be limited.

Nos. 622-623] MIGRATION A FACTOR IN EVOLUTION 419

with the suggestions of Blackman and Smith (711) who
show that certain physiological processes are better ex-
plained as the ‘‘result of interacting limiting factors than
by the conception of the optima.’’ This principle js an
extension of the law of the minimum and is formulated
by them as follows (p. 411):

The identification of the particular limiting factor in any definite
case is carried out by applying experimentally the following general
principles. When the magnitude of a function is limited by one of a
set of possible factors, increase of that factor, and of that one alone,
will be found to bring about an increase of the magnitude of the func-
tion... (p. 397). When several factors are possibly controlling a
function, a small increase or decrease of the factor that is limiting, and
of that factor only, will bring about an alternation of the magnitude of
the functional activity.

Probably this formulation should be broader, and be
made to include not only a single factor, but all unfavor-
able or limiting factors, as I have indicated above, and as
both Livingston (*17, p. 8) and as Hooker (717, p. 201)
suggest.

Recent additional physiological evidence of the con-
centric zonation (gradation) of the limiting factors of
temperature and humidity have been made by Pierce
(16). He accepts the older idea of the optimum and thus
certain of his results on zonation harmonize with my
statement of 1904. He shows that for the cotton boll
weevil there is a vertical temperature gradient which in-
fluences the metabolism, growth and other activities, and
that for a given temperature there is a corresponding
horizontal humidity gradient which forms concentric
zones of less favorable conditions. These extend from
the optimum, through dormancy, on to death. It seems
likely, however, that the idea of ‘‘interacting limiting
factors’? explains his facts better than that of the
optimum.

The idea of limiting factors in experimental work is
now building up a laboratory idea of environmental com-
plexity, even under controlled conditions, which corre-
sponds closely with what the field ecologists have called

480 THE AMERICAN NATURALIST [VoL. LII

an environmental complex. This is a healthy sign as it
will greatly assist in the correlation of field and labora-
tory studies. Recently Livingston (*17, p. 8) has said:

I wish now simply to emphasize the point that we can no longer
speak of a single condition as being a cause of an observed effect, The
next generation of physiologists will have to learn to handle more than
a single variable and to deal with complexes of conditions.”

This recalls John Stuart Mill’s statement that:

It is seldom, if ever, between a consequent and a single antecedent
that this invariable sequence subsists. It is usually between a conse-
quent and the sum of several antecedents; the concurrence of all of them
being requisite to produce, that is, to be certain of being followed by
the consequent. In such eases it is very common to single out one only
of the antecedents under the denomination of cause, calling the others
merely Conditions. .. . The real cause is the whole of these antecedents;
and we have, philosophically speaking, no right to give the name of
cause to one of them, exclusively of the others. . . . All the conditions
are equally indispensable to the production of the consequent; and the
statement of the cause is incomplete unless in some shape or other we
introduce them all.

When Hooker (*17, p. 201) states that, ‘‘It is neces-
sary to get away from the custom of discussing causes,
however difficult this may be. The idea of causation in-
variably indicates incomplete analysis,” he does not ex-
press the full significance of Livingston’s remark. We
have not yet outgrown Mill’s statement.

In addition to its application to the individual animal,
Bancroft’s law applies with equal force to the dynamic
tendencies of plant and animal associations. The domi-
nance of a climax society shows that (Adams, ’08, p. 125) :

Such dominance, in general, implies extensive range, relative abund-
ance, and ability to indefinitely succeed or perpetuate itself under given
conditions. . . . The primary environmental conditions tend to encroach
upon all pike: The local conditions thus tend to become transformed in
the direction of the dominant environment and to be appropriated by it.
The associations . . . are thus given a definite dynamie trend. . . . Minor
environments tend to become encroached upon by the demiagol regional
influences and ultimately to become extinct. The succession of socie-

ge Blackman, 705, p. 293, on the limitations of control experiments.

Nos. 622-623] MIGRATION A FACTOR IN EVOLUTION 481

ties of local habitats is a declining one, while that of the geographic or
climax habitat is an increasing and ascending one. . . . That the domi-
nant geographic conditions tend to override local influences seems very
fairly established because diverse local original conditions are trans-
formed into the climax or dominant type.

To students of human economies (for except ecolo-
gists we seem to have almost no students of general eco-
nomics, including wild animals and plants) Bancroft’s
law should be a revelation. The interference or friction
seen in economics (Conant, '08) should be included under
Bancroft’s law. That these laws apply to human social
conditions as well can easily be tested by any one who
will venture to ‘‘interfere’’ with any system of social
machinery, whether it be of the family, fashion, church,
state or a political party, for very soon the pressure or
stress exerted by the ‘‘system’’ will make itself evident,
by the processes of coercion, persuasion, ridicule, pros-
elyting, threat, ostracism, or by a final crushing effort ;
for interference with a dominant system whether it is
large or small has but one tendency. Years ago Bagehot
(’73, p. 97) clearly recognized what appears to be essen-
tially the laws mentioned above, and applied them with
great skill to the development of political history, under
the names of ‘‘persecution’’ and ‘‘imitation.’’ Perse-
cution corresponding to interference and imitation to ac-
celeration. Hooker (17, p. 208) not recognizing Ban-
croft’s law, suggests what he calls the ‘‘prineiple of inte-
gration’’ to cover the interaction of the systems which he
recognizes. He says: ‘‘These systems are invariably
overcoming the effects of limiting factors.”’

(d) Interaction of Systems.—The next higher category,
above the animal system, is the interaction of the systems,
and their principles of complex action. To be sure, the
animal system can not be divorced from its environment,
so that several important features of this interrelation
have already been discussed briefly. In dealing with the
organism and the environment these two gross systems
are perhaps the most clearly recognized in biology. The

482 THE AMERICAN NATURALIST [Von. LIT

environmental complexity is so great that it is bewilder-
ing to many, particularly to those who have not followed
the most recent methods of dealing with the vegetation
and gross physical environment. For convenience in
handling, this complex may be broken up advantageously
into smaller systems: or units which are the agencies
which influence animals. This plan provides for both
their qualitative and quantitative relations, because the
agents provide for the qualitative units, and their dy-
namic relations include their quantitative intensities.

In dealing with the interaction of systems relating to
animals, one of the first points to consider is the classifi-
cation of these systems, and the recognition of the sizes
of the units. Many groupings are possible, such as the
individual animal, its plant and animal associates, and
the numerous factors of the physical environment. Fur-
ther analytical systems of the vegetational environment
can be grouped according to the recognized units current
among the students of the genetic aspects of vegetational
development (see Cowles, Clements, etc.). For the phys-
ical environment the geologists, physiographers and
geographers have already made much progress in the
analysis of unit systems, which can be used with com-
parative ease (see Chamberlin, Salisbury, Van Hise,
Davis (*09), ete.). In the study of all these systems natu-
rally more progress has been made in their recognition,
than in their complex modes of interaction; and the for-
mulation of their laws of interaction is or the greatest
importance. There are three models which really come
to mind in this connection. These are:

1. The physical model of the interaction of forces,
which leads to resultant motion.

2. The application of Bancroft’s law to the interaction
of all systems.

3. The application of the physical and chemical model
of the phase rule of Gibbs to equilibria of all kinds.

These will now be considered in their respective order:

1. The physical model will assist in keeping in mind

Nos. 622-623] MIGRATION A FACTOR IN EVOLUTION 483

the underlying relations that the stresses, exerted by
agents, will reinforce, overcome or balance one another,
and influence the end result of change. This is a quanti-
tative law. The inertia of the process of adjustment, and
the inertia of equilibria, should be recalled (cf. Newton’s
first law) in this connection. The conception of inertia
appears to have been almost neglected in biology.

2. Bancroft’s law, that systems tend to change to mini-
mize external disturbance, is a general law which appears
to apply to the interaction of all systems. This is a quali-
tative law, which should be of great practical value.

3. The phase rule, according to Henderson (’13, pp.
257-258) is that the
condition of equilibrium in any material system depends upon the num-
ber of its components, the number of its phases, temperature, pressure,
and in general, the concentrations of all the components . . . [as to] the
term “component” and “phase” it will here suffice to say that in
general the number of components increases as the number of separate
chemical individuals increases, and that a phase is any solid, liquid or
gaseous part of the whole system which possesses homogeneity of com-
position. For instance, if a system is made up of sand, salt solution,
ice and aqueous vapor, each of these separate parts in that it is homoge-
- nous, is a phase. . . . Other things being equal, the greater number of
phases, the less the tendency to change.

The quantitative character of this rule makes its appli-
cation one of great difficulty, but it will serve as a guide
or model for the organization of problems, and suggests
the form into which experimental data should be organ-
ized, and secured for testing its application and validity.

The following extracts from Findlay (’04, pp. 8-18)
will assist in gaining some of the general ideas involved
in this subject:

A heterogenous system is made up of different portions, each in itself

vapor, are three phases of the same chemical ca A
phase, however, whilst it must be physically and chemically homogeneous,
need not necessarily be chemically simple. . . . The number of phases
which ean as side by side may vary woii in different systems. In

484 THE AMERICAN NATURALIST [Vou. LIL

all cases, however, there can be but one gas or vapor phase on account
of the fact that all gases are miscible with one another in all propor-
tions. In the ease of liquid and solid phases the number is indefinite,
since the above property does not apply to them... .. It is important to
' bear in mind that equilibrium is independent of the amounts of the
phases present.

By component (p. 10) is

meant only those constituents, the concentration of which can undergo
independent variation in the different phases, and it is only with these
that we are concerned here. . . . The Phase Rule is concerned merely with
those constituents which take part in the state of real equilibrium; for
it is only to the final state, not to the processes by which that state is
reached, that the Phase Rule applies. (Pp.11-13.) It is, however, only
in the case of systems of more than one component that any difficulty
will be found; for only in this case will a choice of components be pos-
sible. . . . Now, although these constituents take part in the equilibrium,
they are not all to be regarded as components, for they are not mutually
independent. . . . In deciding the number of components in any given
system, not only must the constituents chosen be capable of independent
variation, but a further restriction is imposed, and we obtain the fol-
lowing rule: As the components of a system there are to be chosen the
SMALLEST NUMBER of independently variable constituents by means of
which the composition of each phase participating in the state “a equi-
librium can be expressed i in the form of a chemical equation. her
method may be given by which the number of components jpk ma
system can be determined. Suppose a system consisting of several
phases in equilibrium, and the composition of each phase determined by
analysis. If each phase present, regarded as a whole, has the same
composition, the system contains only one component, or is of the first
order. two phases must be mixed in suitable quantities in order that
the composition of a third’ phase may be obtained, the system is one of
two components or of the second order; and if three phases are neces-
sary to give the composition of a fourth coexisting phase, the system
is one of three components, or of the third order. . . . Again, therefore,
we see that, although the number of the components of a system is defi-
nite, a certain amount of liberty is allowed in the choice of the sub-
stances; and we also see that the choice will be influenced by the condi-
tions of experiment.

Summing up, now, we may say:

1. The components are to be chosen from among the constituents
which are present when the system is in a state of true equilibrium,
and which take part in that equilibrium.

_ 2. As components are to be chosen the smallest number of such con-
stituents necessary to express the composition of each phase partici-

Nos. 622-623] MIGRATION A FACTOR IN EVOLUTION 485

pating in the equilibrium, zero and negative quantities of the components
being permissible.

3. In any given system the number of the components is definite, but
may alter with alteration of the conditions of experiment. A certain
freedom of choice, however, is allowed in the (qualitative, not quantita-
tive) selection of the components, the choice being influenced by con-
siderations of simplicity, suitability or generality of application.

We see, therefore, that in case of some systems two, in other cases,
only one of the independent variables (temperature, pressure, concen-
tration) can be altered without destroying the nature of the system;
while in other systems, again, these variables have all fixed and definite
values. We shall therefore define the number of degrees of freedom of
a system as the number of the variable factors, temperature, pressure,
and concentration of the components, which must be on te LCT in
order that the condition of the system may be perfectly defin

A knowledge of its variability is, therefore, of essential anea in
studying the condition and behavior of a system, and it is the great
merit of the Phase Rule that the state of a system is defined entirely by
the relation existing between the number of components and the phases

resent, no account being taken of the molecular complexity of the par-

an ee ye indeed, which attaches equally to the terms “ physical ”
and “chemical ” process

The ae Rule of Gibbs, now, which defines the condition of equilibrium
by the relation between the number of coexisting phases and the com-
ponents, may be stated as follows: A system consisting of » components
can exist in n + 2 phases only when the temperature, pressure, and con-
centration have fixed and definite values; if there are n components in
n +1 phases, equilibrium can exist while one of the factors varies, and
if there are only n phases, two of the varying factors may be arbitrarily
fixed. This rule, the application of which, it is hoped, will become clear
in the sequel, may be very concisely and conveniently summarized in
the form of the equation—

P+F=C+2, o F=C+2—P

where P denotes the number of the phases, F the degrees of freedom,
and C the number of components. From the second form of the equa-
tion it ean be readily seen that the greater the number of the phases,
the fewer are the degrees of freedom. With increase in the number
of phases, therefore, the condition of the system becomes more and more
defined, or less and less variable. . . .

486 THE AMERICAN NATURALIST [Vou. LII

Systems which are apparently quite different in character may behave
in a very similar manner. Thus it was stated that the laws which govern
the equilibrium between water and its vapor are quite analogous to
those which are obeyed by the dissociation of calcium carbonate into
carbon dioxide and calcium oxide; in each case a certain temperature is
associated with a definite pressure, no matter what the relative or abso-
lute. amounts of the respective substances are. And other examples
were given of systems which were apparently similar in character, but
which nevertheless behaved in a different manner. e relations be-
tween the various systems, however, became perfectly clear and intel-
ligible in the light of the Phase Rule. In the case first mentioned, that
of water in equilibrium with its vapor, we have one component—water
—present in two phases, i. e., in two physically distinct forms, viz.,
liquid and vapor. According to the Phase Rule, therefore, since C —1,
and P= 2, the degree of freedom F is equal to 1+ 2—-2—1; the
system possesses one degree of freedom, as has already been stated. But
in the case of the second system mentioned above there are two com-
ponents, viz., calcium oxide and carbon dioxide, and three phases, viz.,
two solid phases, CaO and CaCO,, and the gaseous phase, CO..
number of degrees of freedom of the system, therefore, is 2-2 —3—=1;
this system, therefore, also possesses one degree of freedom. We can
now understand why these two systems behave in a similar manner;
both are univariant, or possess only one degree of freedom.. We shall
therefore expect a similar behavior in the case of all univariant sys-
ems, no matter how dissimilar the systems may otherwise appear.
Similarly, all bivariant systems will exhibit analogous behavior; and
generally, systems possessing the same degree of freedom will show a
like behavior. In accordance with the Phase Rule, therefore, we may
classify the different systems which may be found into invariant, uni-
variant, bivariant, multivariant, according to the relation which obtains
between the number of the components and the number of the coexist-
ing phases; and we shall expect that in each case the members of any
particular group will exhibit a uniform behavior. By this means we are
enabled to obtain an insight into the general behavior of any system, so
soon as we have determined the number of the components and the
number of the coexisting phases.

In the preceding quotations there are certain points to
which special attention should be called :

1. The phase rule is concerned with equilibria, and not
with the processes by which this state is reached. It thus
supplements Bancroft’s law in a very important manner,
| because that law is mainly concerned with the process of

developing equilibria. To make complete continuity and

Nos. 622-623] MIGRATION A FACTOR IN EVOLUTION 487

contact between these two methods and to fuse them into
one cycle Bancroft’s should also be quantitatively ex-
pressed (the interacting systems). The importance of
this is evident. I have not seen attention called to this
fact, or the intimate relation between these two laws.

2. The number of components, phases, independent
variables, concentrations, etc., to which the phase rule
applies, is also one of the most marked features of bio-
logical problems in its dealing with the relation of ani-
mals to diverse environmental conditions and media.

3. The analysis of biological problems into cycles of
action, systems, and agencies, is a necessary simplifica-
tion of the biological problems, and is preliminary to the
determination of the number of components, phases, and
concentrations which are involved in the application of
the phase rule to equilibria, and to Bancroft’s law of
their development. Even in case of biological problems
which have not been reduced to quantity this model of
dynamic relations should be of much assistance in clari-
fying working plans, especially in associational studies.

4. Improvements in the dynamic theory will probably
simplify its application to biology. The detailed non-
mathematical expression of these correlations will facili-
tate their wider use in biology, and it is also equally evi-
dent that with an adequate mathematical equipment the
biologist’s application of these ideas would be greatly
facilitated.

The phase rule has been so valuable chemically that a
special effort should be made to use it as much as possible
as a model in biological work. Mellor (’04, pp. 183-184)
says: ‘‘Gibbs’ phase rule is the best system extant for
the classification of equilibria—chemical and physical.
All changes, both physical changes of state and changes
of chemical composition, are found to depend upon the
same general laws.’’ Henderson (713, p. 260) remarks:

“«“Mhere can be no doubt that, when feasible, the ideal
method—from the physico-chemical point of view—to de-

488 THE AMERICAN NATURALIST [VoL. LII

scribe a material system is in terms of the phase rule.?”

To apply these principles to the interaction of sys-
tems is the great practical problem. It requires, as pre-
viously mentioned, the analysis of the problem to such a
degree as to distinguish its different systems and homo-
geneous units, their degrees of freedom, their directions
of change, their cycles, their dynamic status and their
quantitative relations. Many of these action systems have
long been clearly recognized by plant and animal physi-
ologists and ecologists, as cells, tissues, organs and com-
munities, and.many are recognized also in the physical
and chemical world, where much attention is given to dy-
namic relations.

The processes of integration and dominance tend to
limit this diversity of systems. It is believed, however,
that to strive consciously for the application of these con-
ceptions with some idea of what they imply, will, how-
ever, greatly hasten progress. Further, by calling atten-
tion to these general ideas it may enable some investiga-
tors to become better prepared for handling them, be-
cause we may well recall Pasteur’s remark that: ‘‘In the
fields of observation chance favors only the mind which
is prepared.” It is hoped that by emphasizing these re-
lations others better qualified than I am will give atten-
tion to this subject, and supply numerous examples in the
various specialties; for it is by this method largely that
others can become interested and extend the applications.

Up to this point the discussion has been mainly devoted

+ to the development of a dynamic conception of systems
and their methods of interaction. The migration of ani-
mals has long been recognized to include not only those
caused by the activities of the animal itself, but also by
the activity or agency of their environment; it is there-

8 Henderson (713) does not mention Bancroft’s law or attempt to relate
it to the phase rule. In my ‘‘Guide’’ (*13, p. 85) special attention was
called to both Baneroft's law and the phase rule by listing these first among
papers on dynamics. Recently, since this paper was written, I have seen
Henderson’s (717, p. 138) criticism of Spencer’s ‘‘stability of the homo-
geneous,’’ to which he applies the phase rule and refers also to Bancroft’s
law, although not as here advocated.

Nos. 622-623] MIGRATION A FACTOR IN EVOLUTION 489

fore now necessary to review briefly some of the main en-
vironmental agencies and processes operative. The geol-
ogists and physiographers have made much progress in
the dynamic interpretation of their problems, so that it
is a relatively simple matter to adapt their results to our
purpose. They have shown that the rocks below the
ocean are heavier than those of the land and that the
present shores of the oceans change as a dynamic equi-
librium is established between the heavier sea bottom and
the lighter land area (Willis, 711). We have in this a
cause for innumerable changes in the physical features of
the earth’s surface, and in the environments of animals.
This beautifully illustrates the fundamental unity and
method of interaction between the liquid sea and the solid
land systems. This is one of the huge physical cycles
which illustrate every dynamic phase, from a condition
of stress through the slow process of adjustment to strain
—retarded by the rigidity of the earth’s crust—on to a
dynamic equilibrium. All the land area which remains
above sea level is exposed to waves, the disintegrating in-
fluences of the atmosphere, and the erosion by wind and
running water, all of which tend to cut down all land to
sea level, and to deposit the heavy débris on the sea bot-
tom, thus cumulatively destroying its relative equilib-
rium, and, supplemented by radio-activity, there are in-
stituted cycles of stress, in an unending series.

In the equilibria existing between the land and the sea,
the isotatic cycle, and the cycle of erosion or base-leveling,
are found two phases of the most important gross influ-
engon; in De physical causes of animal migration (cf.
W , 94; Adams, ’01). To this must be specifi-
cally included alimak cycles (Huntington, ’14, and oth-
ers) whose influence upon animal migration is also pro-
found. Intimately related to the preceding physical fac-
tors are the cyclic changes in the vegetational covering
of the earth, particularly those recognized in recent years
by the plant ecologists (Cowles, ’11; Clements, 716, and
others). The physical a eneo all organisms,

490 THE AMERICAN NATURALIST [VoL. LII

and operate in both short and long cycles, the various
cycles traveling at diverse rates and mutually influencing
one another in their adjustments to pressure. In these
wonderful moving systems of cycles can be visualized the
essence of modern scientific conceptions. From electrons,
atoms, molecules, chemical compounds, colloids, cells, tis-
sues, organs, individuals, and culminating in the com-
munity and association, is seen in each a dynamic center
or microcosm, about which revolves other systems, in
turn revolving as a part of a larger system in ever widen-
ing expansion, each in turn subordinated to a higher or-
der of dominance, the culmination of interacting systems.

I have now completed an outline of the fundamental
dynamic principles which are necessary as a background
for my discussion of animal migration. These general
principles appear to underlie all processes of animal reg-
ulation internal and external, and are expressions of
these laws of interacting systems.?

9 Although - EERE discussion is intended to bear r on migra-
tion, it shoul t be inferred that I would limit it in this manner. It is
my belief that dins general principles are of relatively wide pescar

(To be continued)

A STUDY OF HYBRIDS IN EGYPTIAN COTTON

THOMAS H. KEARNEY anb WALTON G. WELLS

BUREAU or PLANT INDUSTRY, UNITED STATES DEPARTMENT OF
AGRICULTURE

INTRODUCTION

THe Egyptian type of cotton comprises numerous va-
rieties which have presumably originated by mutation
(Kearney, 1914). This presumption is based upon the
following facts:

1. Each variety descended from a single individual
which differed in several characters from the parent
form.

2. The absence or extreme rarity of connecting forms
and the infrequency of sterility in both the parental and
the mutant stock make it difficult to account for these
mutants on the basis of recombination, as ordinarily
understood.

3. The new characters of the mutant are uniformly ex-
pressed in the successive generations of its offspring as
long as hybridization with other forms is excluded.

The observed facts make it difficult to escape the con-
clusion that these mutants are the result of simultaneous
alteration of several factors in the egg cell after fertiliza-
tion.: Otherwise, it is necessary to assume that the mu-
tant has resulted from the union of a male and a female
gamete, in both of which similar but independent altera-
tion had taken place with respect to several factors.
Probability is so greatly against this interpretation as to
make it almost unthinkable.

The question suggests itself, what are the conditions
under which mutation occurs in Egyptian cotton? The
appearance of mutants has thus far been observed only

‘in mixed stocks (Kearney, 1918, pp. 60-61). Hence, not-

1A case of mutation has been thus explained by Hayes and Beinhart
(1914).
491

492 THE AMERICAN NATURALIST [Vou. LIT

withstanding the difficulty of interpreting the mutants as
direct recombinations, the inference can scarcely be
avoided that mutation in this group is conditioned by
heterozygosity.

In an endeavor to obtain more definite information on
this point it was decided to make simple and back-crossed
hybrids between two varieties of Egyptian cotton and
to study these hybrids in comparison with line-bred
progenies of the parent varieties. Mutants in this group
are of comparatively rare occurrence, nothing analogous
to the ‘‘mass mutation’’ observed by Bartlett (1915) in
(Enothera and by De Vries (1918) in Zea having been
observed. Very large numbers of plants of the hybrid
and parental stocks will therefore need to be examined
before we may hope to obtain reliable statistics as to the
production of mutants. In the meantime, it is believed
that what has been learned in regard to the behavior of
these hybrids in the first three generations is of sufficient
interest to warrant preliminary publication.

Most previous studies of hybrids in the genus Gos-
sypium have been made upon interspecific crosses, such
as Sea Island cotton (G. barbadense) x Upland cotton
(G. hirsutum) and Egyptian cotton? x Upland cotton.
In these cases there is very great variability in the F,
and later generations. Cotton breeders have found it
to be practically impossible to ““fix”” such hybrids, even
after selection continued during six or seven genera-
tions. On the other hand, little is known of the be-
havior of crosses between varieties within the same
species. Are such hybrids less variable and less diffi-
cult to fix by selection, and, if so, can not stable and uni-
. form new varieties be obtained by recombination? It
is believed that these questions are partly answered by
the data presented in this paper.

The investigation was conducted at the Cooperative
Testing Garden, Sacaton, Arizona, which is conducted-

* The Egyptian type of cotton, although commonly supposed to be of hy-
brid origin (Balls, 1912, pp. 3, 4) presents many analogies to a natural spe-

cies and its differences from American Upland cotton are certainly of spe-
cific magnitude.

Nos. 622-623] HYBRIDS IN EGYPTIAN COTTON 493

by the Bureau of Plant Industry in cooperation with the
Indian Service. The writers are indebted to Mr. G. N.
Collins of the Bureau of Plant Industry for many help-
ful suggestions throughout the course of the investiga-
tion.®
PLAN OF THE INVESTIGATION

The only varieties of Egyptian cotton of which ap-
proximately pure strains were available when the experi-
ment was begun were the Yuma, Gila and Pima varieties,
all of which had been developed in Arizona. These va-
rieties were described (with illustrations of the leaves,
bracts, and bolls) and an account of their origin was
given in an earlier publication (Kearney, 1914). Their
relationship may be indicated thus:

Mit e

q
Gila Yuma

The Gila and Pima varieties were chosen because they
show the greatest amount of difference in the largest
number of characters. Of the three varieties, Gila is
most similar to the common ancestor, Mit Afifi, and Pima
is the most distinct from it. Gila may, in fact, be re-
garded as representing a small portion of the area of
- variation of the extremely heterozygous Mit Afifi stock
from which all these varieties have descended, while the
characters of Pima are far outside the hitherto observed
range of variation in Mit Afifi. The hybrids described
in this paper may therefore be taken to represent, in a
measure, the result of crossing the mutant Pima with its
more remote ancestor, Mit Afifi

Several typical individuals of each variety were se-
lected in July, 1914. A number of flowers were self-
pollinated on each plant and intervarietal cross-pollina-
tions were made among them. The resulting first gen-

3It is not practicable to publish in full the voluminous data resulting
from this investigation but the writers are prep pared to supply, to any one
who may be interested, photographie copies of the original records, at t
cost of reproduction.

494 THE AMERICAN NATURALIST (Véi: LIT

eration parental and hybrid progenies were grown in
1915. Seeds produced by flowers which were selfed on
certain individuals in these progenies furnished the sec-
ond generation, which was grown in 1916. Flowers on
selected plants in the second generation parental and F,
hybrid progenies were again selfed to furnish the third
generation parental and the F, hybrid progenies, which
were grown in 1917.

Some of the flowers on F', hybrid plants in 1915 were
pollinated from plants in the first generation progenies
of the original parent plants of either variety. From
the resulting seed 34 Pima and 34 Gila back-crosses were
grown in 1916. Plants were selected in each of the 3
back-cross progenies because of their approach to the
corresponding predominant parent in respect to impor-
tant characters. The Pima back-cross plants were pol-
linated from a plant in the second selfed generation of
the Pima parent and the Gila back-cross plants were pol-
linated from a similar Gila individual. The resulting
% Pima and % Gila back-cross progenies were grown
in 1917.

Every effort was made to grow the various parental
and hybrid progenies under as nearly as possible uni-
form conditions in respect to soil, irrigation and cul-
tural treatment. All comparisons of hybrids and parents
have been made on the basis of progenies grown the
same season, in order to obviate the influence of different
weather conditions. Measurements on the different
plants were made, as far as practicable, upon organs
which occurred at the same nodes of the axis and
branches and which were in the same stage of develop-
ment. The number of plants on which most of the char-
acters were measured in each generation were, in round
numbers: .

Fi (1915) | Fe (1916) Fs (1917)
Pima Eee Ge ra PUM A Gat Se 60 200 180
GHA LLA ge as aS 40 200 100
Fima Cala eg Se a 80 400 300

Nos. 622-623] HYBRIDS IN EGYPTIAN COTTON 495

CHARACTERS MEASURED AND SIGNIFICANCE OF THE VARIETAL
DIFFERENCES

The Pima and Gila varieties, as represented by the
first and second generation progenies of the selected par-
ent individuals, differed by an amount equal to three or
more times the probable error of the difference, in re-
spect to 24 characters. Many of these are physically or
physiologically correlated, but six of the characters
showed practically no correlation inter se, in either
parent. Most of these characters are expressions of
size (e. g., the length of the internodes, leaves, floral
parts, bolls and fiber) or are ratios between two size
characters and expressive of shape. The leaf index

esis y X 100 and the boll index A y 100 showed

weeds significant differences between the two varie-
ties named, the difference in the second generation hav-
ing been 26 times the probable error in respect to leaf
index and 46 times the probable error in respect to boll
index. In length of fiber the difference between the
parents in the second generation was 22 times its prob-
able error. The only characters of diagnostic value
which could not be accurately measured and which were
therefore determined by grading, were color of the fiber,*
amount of fuzz on the seeds and roughness of the boll
surface (depending upon the depth, number and regu-
larity of distribution of the pits in which the oil glands
are situated). Even in respect to these characters the
differences were of degree rather than of kind.

The Gila variety, as represented by three successive
selfed generations of the progenies of the parent plants,
gave larger coefficients of variation for most of the char-
acters than did the Pima variety. Since no general de-
crease in the variability of either variety was observed
after three generations of selfing, it would appear that
Gila is inherently more variable than Pima.

The two varieties, as represented by the first and sec-

4 The two varieties showed no difference in the color of any part of the
ower, `

496 THE AMERICAN NATURALIST (Vou. LIT

ond generation progenies of the plants selected as par-
ents of the hybrids, showed overlapping ranges for all
characters excepting fuzziness of the seed and color of
the fiber. These two characters were measured on only
small numbers and overlapping would very likely have
been observed if larger populations had been compared.

THE SIMPLE HYBRIDS
Means

Comparing the means of the simple hybrids (Pima X
Gila) with those of the parents, a strong tendency to in-
termediacy was apparent. The means for a large ma-
jority of the characters, in both the F, and F,, lay be-
tween the parental means, and in nearly one half of the
total number of characters the hybrid means did not
differ significantly from the midpoint of the parental
means. The relative number of characters for which
the departure of the hybrid mean was towards the Pima
mean was much greater in the F, than in the F,. This
was probably due to increased vigor in the conjugate
generation, eight of the thirteen characters which showed
a significant® departure of the hybrid F, mean from the
midpoint of the parents being size characters for which
the Pima parent gave a larger mean than the Gila parent.

Eleven F, progenies of the simple hybrid were grown
in 1917 from plants which were selected in the F, in 1916
because of their approach to one or the other parent or
because of their intermediacy with respect to various
characters, especially leaf index and boll index. The F,
means for these characters in all cases fell between the
means of the third generation parental progenies, or
else did not differ significantly from the mean of one or
the other parent.

Coefficients of Variation
-mal three generations the hybrids gave significantly
larger coefficients of variation, for most of the characters,

5A difference or other quantity is here referred to as ‘‘significant’’ when
amounting to three or more times its probable error.

Nos. 622-623] HYBRIDS IN EGYPTIAN COTTON 497

than did the corresponding selfed parental progenies of
the Pima variety, but were not significantly more variable
than the corresponding Gila progenies. In length of
fiber the hybrid F, was not significantly more variable
than either parent. The hybrid F, was significantly
more variable than the F, in the leaf characters, but not
in the boll characters. This result was so surprising
that it was accepted only after repeatedly checking the
original data. The averages of the coefficients of varia-
tion, for leaf index and boll index, of the eleven F,
hybrid progenies did not differ significantly from the co-
efficients of the F, progeny from which they were de-
rived. The average variability of the F, did not exceed
that of the third selfed generation of the more variable
parent (Gila) and two of the F, progenies were not more
variable than the corresponding generation of the less
variable parent (Pima).

These facts point to the possibility of obtaining a rela-
tively uniform new variety of cotton by hybridization of
two varieties belonging to the same general type, al-
though hybrids between different types, such as Egyp-
tian and Upland, are notoriously diffieult to fix.®

Distributions

The distributions, for the important characters leaf
index, boll index and fiber length, of the parental and
simple hybrid progenies, are shown in Figs. 1 to 3.

The range of the hybrid F, for none of the characters
appreciably exceeded the combined parental ranges and
for the majority of characters it was more restricted
than the latter. The variation was therefore much
smaller than in the F, of hybrids between less closely re-
lated types of cotton, in which the range often greatly
exceeds that of both parents. (For example, the Egyp-

6 Longfield Smith (1915, p. 30) states that hybrids between the not very
dissimilar Sea Island and Sakellaridis cottons were ‘‘fairly uniform’’ in
the F, and F, Different behavior was shown by the cross between Sea Is-
land and ‘‘St. Croix Native,’’ a type more nearly resembling American Up-
land cotton. In this case, ‘‘new characters, not present in either of the
original plants, appear in the first and subsequent generations,”” and the
F, and F, ‘‘split into a mass of types.’’

498 THE AMERICAN NATURALIST [Vor. LII

tian X Kekchi hybrids described by Cook (1909, p. 12-14)
and the Egyptian X Hindi hybrids described by Marshall
(1915).)

Little or no evidence of dominance was shown by the
distributions, for the various characters, of the F, of the

E

PERCENTACE PPPEQIOLIVCE,
ee a

~-
LEGE INDEX

Leaf index: first and second generation distributions of the parental

and simple hybrid progenies. The dotted line curye represents the distribution

of the Pima parent, the broken line curve that of the Gila parent and the solid

total — of the population. The figures on the axis of abscissas indicate
the c

sie pe numbers of the respective populations were: F;, Pima 60, Gila 40,
PxG 80; F», Pima 213, Gila 203, PxG 418.

simple hybrids. The F, distributions were strictly uni-
modal for all characters excepting the highly variable
one length of axis, and even in this case the indication of
bimodality was so slight as to be probably insignificant.

Nos. 622-623] HYBRIDS IN EGYPTIAN COTTON 499

No evidence of segregation in definite ratios was ob-
tained.

It is well known that complete dominance in the F,
and 3:1 segregation in the F, are exceptional in size and
shape characters such as chiefly distinguished the Pima
and Gila varieties of cotton. On the other hand, the
question whether the behavior of the Pima X Gila hy-

A.

PENCENTACE IHEPUENCI

Fra. 2. Boll index: first and second generation distributions of the parental
and simple hybrid progenies. Details as in Fig. 1. The actual numbers of
the respective populations were: Fi, Pima 60, Gila 40, Px G 80; Fo, Pima 161,
Gila 207, PxG 419.

brids with respect to such characters might be inter-
preted by the multiple factor hypothesis is extremely
difficult to answer. The differences between the means
of the parents for most of the characters, while small,
are highly significant, but the parental ranges, in most
cases, overlap to such an extent that Mendelian analysis
would seem to be out of the question.

500 THE AMERICAN NATURALIST [Vou. LI

A few of the characters in respect to which these va-
rieties differ significantly might be termed ‘‘qualitative.’’
"These are roughness of the boll surface, color of the
fiber and fuzziness of the seed, all of which are included
in the lists of allelomorphic pairs of characters in cotton

N

e $--e
AOAN N y Nh
X ANNEANNE
yan
AIEEE? LENCG7I7

Fic. 3. Fiber length (mm.); second generation distributions of the parental
and simple hybrid progenies. Details as in Fig. 1. The actual numbers of th
respective populations were: Pima 46, Gila a PxG 49.
given by Balls (1909, p. 18) and by McLendon (1912, pp.
168, 169).7 Yet in the Pima X Gila hybrids these charac-
ters behaved like the size and shape characters, showing
unimodal distribution in the F,. It should be noted,
however, that in respect to these characters the differ-
ences between the two Egyptian varieties are much
smaller than the differences between the parents of the
wider crosses (Egyptian X Upland and Sea Island X
Upland) dealt with by Balls and McLendon. Instead of
the differences between pitted and smooth bolls, buff and
white fiber and smooth and fuzzy seeds we have, in com-
paring Pima with Gila, merely the differences between
more and less numerously and regularly pitted bolls,
lighter and darker buff-colored fiber and more and less
fuzz on the seeds.

T Complete ets is apparently a rather rare phenomenon in cotton,
even in the ¢ of color characters. It is stated, however, by Leake and
Prasad (as, p pp. 126-128) that yellow corolla color and the presence of
the petal spot are completely dominant in certain hybrids of Indian eottons.

Nos. 622-623] HYBRIDS IN EGYPTIAN COTTON 501

Correlation of Characters

It was sought to ascertain whether these hybrids show
genetic as distinguished from merely physical or physio-
logical correlations, in other words, whether there is co- -
herence in the transmission of the parental characters.
To this end, application was made of the test proposed
by Collins (1916, p. 439), 1. e., comparison of the coeffi-
cients of correlation of the F, with those of the F,. It
is assumed that if the F, coefficient significantly exceeds
that of the F',, in the direction indicated by the relation
of the parental means for the two characters, genetic
correlation or coherence of characters is indicated. For
example, the Pima parent has a lower leaf index and a
higher boll index than the Gila parent. If there is co-
herence of these characters, the hybrid should show a
negative correlation and the coefficient of correlation
should be significantly larger in the F, than in the F,.

The coefficients of correlation of 40 pairs of characters
were determined for both the F, (grown in 1915) and F,
(grown in 1916), upon the basis of one measurement of
each character on each plant. In three of these cases
the coefficient of correlation of the F, was significantly
larger than that of the F, (difference from 3.5 to 4.5
times its probable error), in the direction indicated by
the parental relation. Since, however, the coefficients of
correlation of the first generation (1915) and second gen-
eration (1916) of the parental progenies had also been
found to differ in magnitude, the possibility was consid-
ered that the difference in the F, and F, hybrid coeffi-
cients was at least partly due to variations in the weather
of the two years. The coefficients of correlation for the
three character pairs above mentioned were therefore
calculated for a new F, which was grown in the same
year as the F,.? When the F, and the new F, coefficients

8 Two characters of great practical importance and in respect to which the
parents differ very significantly, length of fiber and fuzziness of seeds,
showed no i ag correlations, either with each other or with the leaf

wr boll characters
he new F, was the meats of gl m daughters of the original
Se plants, which had been made in

502 THE AMERICAN NATURALIST [Vou. LII

were compared, it was found that the former was
significantly larger than the latter, in the indicated direc-
tion, for only one pair of characters (width of leaf and
number of teeth on the involucral bracts) and in this
case the difference was only 3.2 times its probable error.*”
It was sought to throw further light upon this problem
by determining the coefficient of correlation in the hybrid
upon the basis of progenies rather than of individual
plans. The means, for leaf index and boll index, of
eleven F, hybrid progenies were used for this purpose.
These progenies comprised from 8 to 44 plants each and
the means were based upon measurement of one leaf and
one boll on each plant. Since the Pima parent has the
smaller leaf index and the larger boll index, the correla-
tion in the hybrid, if determined by the parental relations
of the two characters, should be negative. The coefficient
obtained was in fact negative, but was no larger than its
probable error (r= — .17 +
e balance of evidence is therefore strongly against
the occurrence of coherence of characters in these hybrids
between somewhat closely related, although distinct, va-
rieties of cotton. - It does not, of course, follow that the
same result would have been obtained in the case of
hybrids between less closely related types, especially if
these differ in allelomorphic characters rather than in
the variable size and shape characters which chiefly dis-
tinguished Pima from Gila.*?
10 The correlation in question, width of leaf with number of teeth on the

involueral bracts, is doubtless physiological, large bracts being associated
with large leaves and the number of teeth being greater on the larger

1 The existence of an intervarietal correlation by no means implies that
the same correlation will be found to obtain within a y ariety. For example,
it is a matter of common observation that most varieties of cotton which
have very long fiber have relatively sparse fiber, and vice versa. But when
the correlation between fiber length and lint index (weight of fiber per 100
seeds) was plotted for 80 plants of the Pima variety, the value of r was
found to be only .07 X .07, showing the complete absence of an intravarietal
correlation,

12 baka of coherence of characters in hybrids of Egyptian with Up-
land cotton have been reported by Cook (1909, pp. 16, 17 and 1913, p. 53).
On the other hand, Marshall (1915, pp. 57, 61), describing the F, of hybrids

Nos. 622-623] HYBRIDS IN EGYPTIAN COTTON 503

THE Bacx-CrossepD HYBRIDS
The means of the 34 back-crossed hybrids for nearly
all characters showed departures from the midpoint of
the parental means which were both significant and to-
wards the mean of the respective preponderant parent
(Pima or Gila). In the % back-crosses, the mean vir-
tually coincided with those of the preponderant parent,
as the following data show:
Progeny Means of: Leaf Inde Boll Index
Pima 78.1 + .14
Pima { back-eross [(P X G) X P] XP 78.7 + .63 179 + .66

Gila 3 back-eross [(P X G) XG] XG 94.14.47 153 + .90
Gila 93.1 + .47 156 + .51

The distributions, for leaf index and boll index, of the
Pima 7% back-cross, were embraced by those of Pima and
the distributions of the Gila 7 back-cross were embraced
by those of Gila. It is therefore apparent that twice
back-crossing the simple hybrid with either of its parents
has sufficed to eliminate the influence of the other parent
in the expression of these characters.

ABSENCE OF MUTANTS

Careful examination, in 1916, of every plant in the F,
progenies of the simple hybrids and in the 3 back-eross
progenies, showed only various recombinations of the
Pima and Gila characters. A large majority of the
simple hybrid plants were approximately intermediate,
although occasional individuals showed a near approach
to one or the other parent. Most of the back-cross
plants, as compared with the simple hybrid individuals,
showed clearly the preponderating influence of the 3%
parent, but few if any plants in the 3 back-cross pro-
genies could have been classed as wholly Pima or wholly
Gila, the influence of the one-quarter parent being ob-
servable in the great majority of cases. No instance of
between the equally different Egyptian and Hindi cottons, states: **Nor
was it possible to discover any general correlations or definite associations
between any of the more important structural differences. ?”

504 THE AMERICAN NATURALIST [VoL. LII

the occurrence of new or extra-parental characters was
detected, although a few plants in the F, of the simple
hybrid slightly exceeded the range of one or the other
parent. Examination, in 1917, of the F, progenies of
the simple hybrids, and of the 7% back-cross progenies
also failed to reveal the occurrence of any extra-paren-
tal characters. Nor have any new characters been de-
tected in the first, second or third generation progenies
from selfed seed of the parent individuals. It is evident,
therefore, that nothing in the nature of a mutant has yet
appeared in any of these line-bred and hybrid stocks.

It was not, however, expected that mutants would be
detected in these small progenies, which were grown for
the purpose of studying, under controlled conditions, the
behavior of the hybrids in the earlier generations and
to provide seed for the growing of each stock on a more
extensive scale. Statistical evidence regarding the pro-
duction of mutants can scarcely be expected until much
larger numbers of plants have been examined. The
stocks resulting from repeated back-crossing should be
especially interesting to study in regard to the occur-
rence of mutation.!?

CONCLUSION

The investigation here described was undertaken in
the endeavor to ascertain the conditions under which
mutants are produced, in Egyptian cotton. Simple and
back-crossed hybrids were made between two varieties
(Pima and Gila) which differ significantly in numerous
characters. Three generations of the hybrid progenies
and of progenies from selfed seed of the parent indi-
viduals, were grown. No evidence of the appearance of
- 18‘* Variations toward Upland or Hindi characters arising in dilute hy-
brid stocks of Egyptian cotton have been found to yield progenies with
more stable expression of characters than direct hybrids between Egyptian
and Upland cotton. Such facts suggest the possibility of developing a new
method of breeding by dilute hybridization. By the use of a small pro-
portion of foreign blood as i i j
otherwise uniform stocks it may be possible to secure desired combinations

Nos. 622-623] HYBRIDS IN EGYPTIAN COTTON 505

new characters was detected in any of these progenies,
but since mutants in Egyptian cotton are comparatively
rare, it will doubtless be necessary to examine much
larger populations before definite conclusions can be
drawn as to the occurrence of mutation in these stocks.

The principal interest of the data thus far obtained
attaches to the behavior of hybrids between varieties be-
longing to the same general type, as compared with that
of the hybrids between different species of Gossypium,
which have hitherto been the principal subject of genetic
investigation in this group of plants.

The varieties used in this investigation are distin-
guished chiefly by size and shape characters, although a
few of the characters in which they differ significantly
have been found to behave as allelomorphs in hybrids be-
tween less nearly related forms of Gossypium. The
Pima Xx Gila hybrids, however, showed no evidence of
segregation in definite ratios in respect to any of the
characters measured. There was little or no evidence of
dominance in the F,, and the F, distributions were prac-
tically without exception unimodal. The means of the |
simple hybrid were in most cases intermediate between
those of the parents. The result of twice back-crossing
the simple hybrid upon either parent was to obliterate
the expression of the characters of the other parent.

It could not be demonstrated that genetic correlation
or coherence of characters occurs in these hybrids. Ap-
parently all characters which are not correlated physi-
cally or physiologically are transmitted independently.

The second and third generations of the hybrids, as
compared with the parents after two and three genera-
tions of selfing, were not more variable than Gila, and
were only a little more variable than Pima. This fact
is of practical importance in cotton breeding, since it
points to the possibility of obtaining relatively stable
and uniform recombinations of the desirable characters
of varieties belonging to the same general type, although
breeders have found this to be well nigh impossible in

506 THE AMERICAN NATURALIST [VoL. LIT

wider crosses such as those of Egyptian (or Sea Island)
with Upland cotton.

LITERATURE CITED
Balls, W. Lawren ;
1909. Some “xia aspects of cotton breeding. In Ann. Rep.
er. Breeders Assoc., Vol. 5, pp. 16-28.
1912. The cotton plant in Egypt, 202 p.
Bartlett, H. H.
1915. Mutation en masse. Amer. NAT., Vol. 49, pp. 129-139.
ass mutation in (Enothera pratincola. Bot. Gaz., Vol. 60, pp.
425—456.
Collins, G. N.
1916. Correlated characters in maize breeding. Jour. Agric. Res., Vol.
435—453, pl. 55-63.
Cook, O. F.
1909. eT and ring gh characters in cotton hybrids. U. $.
. Agri 1, Ind., Bul. 147, 27 p
1913. Heredity and otiia is U. S. Dept. Agric., Bur. Pl. Ind.,
Bul. 256, 96 p., 6 pl.
De Vries, Hugo.
1918. Mass mutation in a mays. Science, N. S., Vol. 47, pp. 465-467.
Hayes, H. K. and E. G. Bein
1914, ea in enets Science, N. 8., V. 39, pp. 34, 35.
Kearney, T.
1914, Mutation in Egyptian cotton. Jour, Agric. Res., Vol. 2, pp.
287-302, pl. 17-2.
1918. A plant industry based upon mutation. Jour. Heredity, Vol. 9,

pp. 51
, H. Martin and Ram e
1614. Studies in Indian cottons, Pt, 1. Mem, Dept. Agric. India, Bot.
Ser., Vol. 6, pp. ene pl. 1-19,
Marshall, Charles G.
1915. Perjugate cotton hybrids. Jour. Heredity, Vol. 6 (1915), pp.
57-64.

McLendon, C. A. ;
1912. Mendelian inheritance in cotton hybrids. Ga, Agr. Exp. Sta.
Bul. 99, pp. 141-228.
serie Longfield. í :
1915. Experiments with hybrid cotton. Rep. Agric. Exp. Sta. St.

Croix, for 1913-14, pp. 29-31.

GENETIC RELATIONS OF THE WINGED AND
WINGLESS FORMS TO EACH OTHER AND
TO THE SEXES IN THE APHID MACRO-
SIPHUM SOLANIFOLIL

DR. A. FRANKLIN SHULL

UNIVERSITY oF MICHIGAN, ANN ARBOR, MICHIGAN

INTRODUCTION

Lire cycles are known, in many aphid species, from
field observations alone. A number of cycles have been
determined from breeding experiments upon aphids in
confinement. Often, however, these experiments appear
not to have used the pedigree method. In the course of
some work on the potato aphid, Macrosiphum solani-
folii, I observed indications of peculiarities in the ge-
netic relations of the various forms to each other, which
could be detected only by the pedigree method. Experi-
ments designed to demonstrate these relations were in-
stituted, with the results described in this paper.

Macrosiphum solanifoli, as observed in these experi-
ments, comprises four kinds of individual: (1) the apter-
ous viviparous female, which is green; (2) the alate vivip-
arous female, which is also green; (3) the oviparous or
sexual female, which is wingless and of a yellowish-green
color until late in life, when the abdomen becomes filled
with green eggs which impart a green color to the female
herself; and (4) the male, which is winged and of a brown
or brown and green color. Of these types of individual,
the alate female can be recognized when a little more
than half grown by her wing pads. The oviparous fe-
male has thickened brown hind tibie covered with sen-
soria, which are recognizable with the unaided eye, and
which develop a few days before maturity. The male

1 Identified by Dr. Edith M. Pateh.

i 507

508 THE AMERICAN NATURALIST [Vou. LIT

is usually, though not always, distinguishable at birth
because of its pink or gray color, and those that are green
at birth usually develop the gray or pink color within a
few days. This color of the immature male has not, I
believe, been recorded in the published descriptions of
the species, and it is not impossible that it is a character-
istic of certain parthenogenetic lines only.

Miss Patch (1915) has described a pink variety of each
of the viviparous forms. I have never seen these in my
experiments, though thousands of individuals have been
examined, except in diseased animals which died shortly
after discovery. The immature pink aphids in my ex-
periments have all been males. The occurrence of pink
females is probably a characteristic of certain partheno-
genetic lines.

In my experiments the potato has been exclusively
used as the host plant. These plants were reared in pots
and were covered with lantern globes closed at the top
with muslin.

EXPERIMENTS

Relation of Winged and Wingless Forms to Each Other

Experiment 293.—Starting with sister individuals, two
lines were bred for three generations, one line from ap-
terous parents exclusively, the other from alate parents
only. As in the other experiments to be described, about
a dozen adult females were placed together on a single
plant, to become parents of the following generation.
When they began to produce young, the latter were re-
moved daily, or every two days, to young plants. Suc-
cessive groups of young were placed on one plant until
they seemed likely to become too crowded (usually not
over 150 per plant), after which a new plant was used.
As many as five plants were required in some cases to
receive the young of one lot of parents. In the tables
these plants are designated, in the columns headed “Host
Plant,” as first, second, third, ete. As the young aphids
became adult they were removed and either used for
further breeding or destroyed.

Nos. 622-623] THE SEXES IN THE APHID 509

As a rule the parents for the following generation were
taken from the first of the host plants, and were trans-
ferred to a young healthy plant, on which they produced
their young.

This particular experiment was started in June from a
stock whose stem mother hatched in the greenhouse in
the preceding January. The line had passed through a
sexual phase in that time, but had been preserved by a
small number of viviparous females. In Table I the line

rom apterous parents is represented in the upper half
of the table, the line from alate parents in the lower half.
For the sake of comparison the totals are placed together
at the bottom of the table.

TABLE I

CONTRASTING THE OFFSPRING OF APTEROUS PARENTS AND THOSE OF ALATE
PARENTS IN THE APHID Macrosiphum solanifolii

Offspring
= y Alate vi
a Parents | Host Plant Dates = ae q ra Ovip-
arous | arous | (Sexual) | Males
Females | Females

J.....| Apterous | First June 27—June 30 49 64 0 0
Second July 2 35 73 0 0
Third July 4-July 5 38 48 0 0
II....| Apterous | First July 7-July 9 22 118 0 0
Second July 11 44 88 0 0
Third July 13 21 62 0 0
Fourth July 14—July 16 2T 28 0 0
III...| Apterous | First July 18-July 21 47 29 0 0
Second July 23 vá 7 0 0
Loa Alate First fJune 28-July 2 110 21 0 0
Second July 4-July 6 32 2 0 0
II....| Alate First July 9-July 11 66 29 0 0
Second July 13-July 14 99 0 0
i July 16-July 18 62 52 0 0
Fourth July 21 41 12 0 0
Fifth July 23 14 4 0 0
III.. .| Alate First July 21 80 23 0 0
Second July 23 35 6 0 0
Third July 24—July 26 5 3 0 0
Totals from apterous parents.........-...--| 285 517 0 0
"Potala from alate paronit.. o coses 544 0 0

There is a striking preponderance of winged offspring
in the families of wingless parents, and a preponderance
of wingless offspring from winged parents.

510 THE AMERICAN NATURALIST [VoL. LIT

Experiment 294.—This was a repetition of the preced-
ing experiment, in part simultaneous with it but of
shorter duration. The method of conducting the experi-
ment was the same as in the preceding experiment.
Table II gives the results.

TABLE II

oe Two RELATED LINES oF Macrosiphum solanifolii, a REARED
M APTEROUS PARENTS, THE OTHER FROM ALATE PAREN

Offspring
Genera- es
coset) Parents | Host Plant | Dates of Moeting o | Vivipe- | rous
rous rous (Sexual) | Males

| Females | Females | Females
Ein Apterous | First July 7-July 9 22 118 0 0
Second vehi aE 44 88 0 0
d July 13 21 62 0 0
Fourth July 14-July 16 24 28 0 0
II....| Apterous | First July 18—July 21 47 29 0 0
Second July 23 y ri 0 0
PAS Alate First July are 9 98 58 0 0
Second che 93 67 0 0
| 62 18 0 0
Fourth a 1 July 16 66 46 0 0
Fifth July bi 3 0 0
II... .| Alate First July 18 65 10 0 0
| Second July 2 179 29 0 0
‘July 23 July 24 22 4 0 0
Totals from reig DAMA a 168 332 0 0
‘Totals from slate parents... 6c. Io... cans 602 235 0 0

The conclusion is the same as from Table I. Apterous
parents give birth more largely to alate offspring, alate
parents more largely to apterous offspring.

Relation of Winged and Wingless Forms to the Sexes

Experiment 303.—Just before the sexual phase of the
cycle began a line from winged parents was started from
a line being reared from wingless parents. In so far as
the two were bred simultaneously their progeny are re-
corded in Table IIT.

Of the sexual offspring, the wingless parents produced
exclusively males, while the wingless parents gave birth
to a very large majority of sexual females. It may be

Nos. 622-623] THE SEXES IN THE APHID 511

TABLE III
CONTRASTING THE PROGENY OF ALATE PARENTS WITH THE PROGENY OF
PTEROUS PARENTS, WITH SPECIAL REFERENCE TO THE SEXUAL FORMS,
IN THE APHID Macrosiphum solanifolii

Offspring
Genera-| Parents | Host Plant Dates of Isolating Apterous| Alate Ovipa-
tion Young Vivipa- | Vivipa- rous
rous rous | (Sexual) | Males

Females | Females | Females
I.....| Apterous| First Sept. 7-Sept. 9 68 2 0 0
Second | Sept. 11-Sept. 13 33 31 0 0
Third Sept. 15-Sept. 19 0 14 0 36
Fourth . Sept. 21 0 0 0 11
TT....| Apterous | First Sept. 21-Sept. 23 0 62 0 0
Second Sept. 25-Sept. 27 0 17 0 0
Third Sept. 29-Oct. 9 0 eto 0 59
Fourth ? 0 0 0 11
Tee Alate First Sept. 17-Sept. 19 6 9 0 0
Second | Sept. 21-Sept. 23 T 0 A ES e
Thir Sept. 25-Sept. 27 0 0 19 1
Fourth Sept. 29-Oct. 5 1 0 14 4
II... .| Alate First Sept. 27-Sept. 29 0 0 27 0
Second Oct. 1-Oct. 7 0 0 62 0
ct 0 0 4 0
Totals from apterous parents. ..............| 101 134 0 117
Totals from slate parents... > ..:. 66a ee 14 9 134 i7

also pointed out that the conclusion regarding the rela-
tion of the winged and wingless viviparous females to
each other that was drawn from Tables I and II is con-
firmed in Table III.

Progressive Change in the Frequency of all the Forms in
Successive Generations

Confirmation of the conclusions drawn from the pre-
ceding experiments is found in several lines which were
designed to show the normal life cycle over a consider-
able period when each generation was derived from wing-
less parents. In addition, these lines show a progressive
change in frequency of both the winged and wingless
viviparous forms and of the sexes. Owing to this pro-
gressive change it was not possible to maintain uniform
parentage, since in the late generations there were no
apterous individuals from which to breed. The princi-
pal lines were obtained in the three following experi-
ments.

512 THE AMERICAN NATURALIST (Vou. LII

Experiment 299.—This was a line reared from a wing-
less female obtained out of doors about August 10. It
was reared in the laboratory. Table IV records this line.

TABLE IV
A Re aa aren ee LINE or Macrosiphum solanifolii, PRO-
Y WINGLESS PARENTS
Note (1) ree asad transition Pei wingless to winged viviparous

mainly by wingless parents, the sexual females chiefly by winged parents.
See Experiment 299

Offspring
Oma varens | on mant| Dateng tenting | apioa] Atte | ovio
arous arous | (Sexual) | Males
Females | Females | Females

I.....| Apterous | First Aug. 24-Aug. 25 50 0 0 0
Second Aug. 27—Aug. 28 51 0 0 0

Third Aug. 30-Sept. 3 29 0 0 0

II....| Apterous! First Sept. 1 0 0 0
cond Sept. 3 149 0 0 0

Third Sept. 5-Sept. 7 49 0 0 0

Fourth Sept. 5-Sept. 15 0 2 0 37

Fifth Sept. 9-Sept. 15 58 42 0 7

Sixth Sept. 17 4 1 0 0

TIT... Apterous| First Sept. 9-Sept. 11 57 17 10) 0
Second Sept. 13-Sept. 17 9 13 0 7

Third ae 0 0 0 9

IV...) Apterous | First Sept. 17-Sept. 19 13 86 0 0
Second Sept. 21-Sept. 23 1 30 1 3

Sept. 25-Sept. 27 0 0 0 15

Fourth | Sept. 29-Oct. 10 0 0 0 11

V....| Apterous | First Sept. 27 6 81 0 0
ond Se 0 34 0 0

į hird Oct. 1-Oct. 3 0 2 14 44

Fourth be 0 0 2 39

Fifth Oct. 9-Oct. 16 0 0 0 25

VI Apterous | First Oct. 7-Oct. 9 2 17 8 0
cond Oct. 11-Oct. 16 0 1 LE 21

Third Oct. 20-Oct. 0 0 0 12

Alate First Oct. 16-Oct. 25 0 0 24 0
Second Oct. 30—-Nov. 6 0 0 4 0

VI ..| Apterous t Oct. 0 5 11 3
Second Nov. 6 0 0 0 1

Alate? First Oct. 18-Oct. 20 0 0 134 0
Second 0 0 101 0

Third Oct. -2i 0 0 44 0
Fourth | Oct. 30-Nov. 6 0 0 18 0

VIII .| Alate? First Nov. 1-Nov. 3 0 0 50 0
l Second Nov. 6 0 0 10 0

Third - Nov. 10 0 0 3 0

2 The alate parents in the seventh and eighth generations were offspring
of apterous parents.

Nos. 622-623] THE SEXES IN THE APHID 513

Experiment 298.—This line was derived from the same
female as Experiment 299, but was reared in the green-
house. See Table V. ;

TABLE V
A PARTHENOGENETICALLY PRODUCED LINE or Macrosiphum solanifolii, Pro-
DUCED CHIEFLY BY APTEROUS PARENTS

The same transitions noted in Table IV are observable here on a smaller
‘scale. See Experiment 298.

Offspring
Genera-| parents | Host Plant Dates of Isolating Apterous| Alate Ovipa-
tion Young Vivipa- | Vivipa- rous
rous rous | (Sexual) | Males
Females | Females | Females

To Apterous | First Aug. 13 rd 0 0 0
II....| Apterous | First Aug. 15-Aug. 17 100 5 0 0
III...| Apterous | First Aug. 24-Aug. 27 15 0 0 0
Second A S-Au 65 2 0 0

Sept. 1-Sept. 5 64 4 0 0

IV...| Apterous | First Sept. 7-Sept. 9 68 9 0 0
Second | Sept. 11-Sept. 13 33 37 0 0

hird Sept. 15-Sept. 19 0 14 0 36

Fourth Sept. 21 0 0 11

V....| Apterous | First Sept. 21-Sept. 23 0 62 0 0
l Second Sept. 25-Sept. 27 0 ay 0 0

Third Sept. 29-Oct. 9 0 2 0 59

Fourth 0 0 0 11

Vis. .| Alate First Oct. 13-Oct. 20 0 0 14 0
Second Oct. 23 0 0 4 0

Experiment 270.—The stem mother of this line hatched
from a fertilized egg that was laid in the greenhouse in
November and hatched in January. In March and April
the line passed through a sexual phase, but a small num-
ber of viviparous females were produced during this
period and by them the parthenogenetic line was con-
tinued. The families were not fully recorded until the
latter part of May. Table VI includes only the records
beginning May 28. Nine generations are there recorded
as if an uninterrupted line, but an explanation is neces-
sary. In the midst of the fifth generation the aphids
began to die in large numbers for an unknown reason.
In a few days every aphid out of hundreds was dead.
Fortunately two or three aphids were found on a dis-
carded plant in the greenhouse. Since only this one line
had been reared in the greenhouse up to that time I felt

514 THE AMERICAN NATURALIST [Vou. LII

TABLE VI
A PARTHENOGENETIC LINE OF THE APHID tion solanifolii REARED
CHIEFLY FROM APTEROUS PARENTS
With certain irregularities the features mentioned in Table IV are recog-
nizable here also. See Experiment 270.

Offspring
EA Hire bed ip pal cael
: rous rous (Sexual) Males
Females | Females | Females

Io Apterous First May 28-June 3 12 31 0 0

Second June 5-June 8 8 0 0 0

II....| Apterous First June 12—June 16 31 25 0 0

cond June 18-June 25 3 5 0 0

III. ..| Apterous First June 27-June 30 49 0 0

Second 73 0 0

hird July 4-July 5 33 48 0 0

IV...| Apterous| First July 7-July 9 22 118 0 0

Second July 11 44 88 0 0

Third July 13 21 62 0 0

j Fourth July 14-July 16 27 28 0 0

V....| Apterous First July 18—July 21 47 29 0 0

Second July 23 Y ra 0 0

VI...| Apterous | First Aug. 17 7 0 0 0

VII ..| Apterous | First b: TE 0 0 0

Second | Sept. 11-Sept. 17 4 0 0 4

Third Sept. 19-Oct. 5 1 0 0 10

VIII .| Apterous | First Sept. 21-Sept. 23 0 50 2 0

Second | Sept. 25-Sept. 27 0 15 5 15

Third Sept. 29-Oct. 3 0 0 0 34

'- Fourth Oct. 5-Oct. 16 0 0 0 29

IX Alate i Oct. 11-Oct. 16 0 0 76 0
Second Oct. 18-Oct. 25 0 0 4 o

safe in assuming that these were of the same line. From
one of them the sixth (?) and succeeding generations of
Table VI were obtained.

Attention is directed in Tables IV, V, and VI to the
following points:

1. The wingless viviparous females, more abundant
early in the cycle, are gradually replaced by winged
females. This is especially clear in Tables IV and V. It
is obscured in Table VI by the fact that this line is not
really a continuous one. :A catastrophe in the fifth gen-
eration made it necessary to resume this line by means
of a female from the same stock. Up to the fifth genera-
tion there is an irregular increase in the proportion of
winged ais which reaches its climax in the fourth

Nos. 622-623] THE SEXES IN THE APHID 515

generation. It is impossible to state what the fifth gen-
eration would have included, since only one fifth of the
probable progeny were produced or survived. After the
fifth generation upto the complete disappearance of vivip-
arous forms, there was again a replacement—this time
rather sudden than gradual—of the wingless females by
winged ones. The same gradual disappearance of wing-
less viviparous in favor of winged females was observed
in several other experiments of shorter duration which
are not included in this paper, and has also been found
in Microsiphum destructor by Miss Gregory (1917). It
-is therefore to be regarded as of general occurrence.

2. There is observed in two of the tables (IV and VI)
a gradual increase in the tendency of wingless females to
produce sexual females instead of males, as they most
often do when the sexual phase begins. Thus in Table
IV, generation IV, 2.5 per cent. of the sexual forms pro-
duced by apterous parents were sexual females. In gen-
eration V, 12.9 per cent. of the sexual forms were females.
In the sixth generation, of the sexual offspring of apter-
ous parents, 29.7 per cent. were females. In the seventh
generation, which is the last from apterous parents, 73.3
per cent. of the sexual offspring were females. Thus,
while the apterous parents produced mostly males during
the sexual phase, there is a gradually increasing tendency
to produce females. In Table VI is a brief indication of
this same phenomenon. Males alone (of the sexual indi-
viduals) appear in the seventh generation, but a small
number of females in the eighth generation. Unfortu-
nately no apterous parents were available for a further
generation. If it were possible to obtain wingless
females in later generations it would be interesting to
note whether they would not eventually produce only
females. |

Whether there is a similar progressive change in the
sexual offspring of winged females is not so clear, since
in none of the last three tables of this paper are there
any male offspring of alate parents. However, in the

516 THE AMERICAN NATURALIST [Vou. LIT

lower half of Table III there is an example of this kind.
In the first generation from alate parents there is a
minority of males; in the second generation no males. It
is not improbable that, if a long line had been bred from
alate parents, there would be a progressive decrease in
the proportion of male offspring in the sexual phase of
the cycle.

3. Tables IV, V, and VI also contain confirmation of
the conclusion drawn from the earlier tables, namely, that
at any given time winged viviparous parents produce
more wingless viviparous offspring than do wingless
parents, and that in the sexual phase males are produced
chiefly by the wingless parents, sexual females by winged
parents.

Discussion

Although the most striking results of the foregoing ex-
periments may appear to be the fact that winged vivip-
arous females produce mostly wingless females in the
parthenogenetic part of the cycle and sexual females in
the sexual part, whereas the wingless viviparous females
produce chiefly winged females in the parthenogenetic
phase and males in the sexual, nevertheless the clue to
the explanation of this phenomenon is more nearly dis-
coverable in the progressive change in the frequency with
which all forms occur in successive generations. Thus,
there is a transition from a preponderance of apterous
females early in the cycle to a predominance of winged
females later. There is likewise, in the sexual portion of
the cycle, a transition from males to sexual females.
This latter transition has been demonstrated in the off-
spring of wingless mothers, and is indicated as probable
in the offspring of winged females.

These transitions imply a gradual change of some sort,
presumably in the metabolism of the animals. While the
difference between a male and a sexual female, or be-
tween an apterous and an alate viviparous female, may be
a definite morphological difference such as a difference
in chromosomes, so that an individual is either the one or

Nos. 622-623] THE SEXES IN THE APHID 517

the other, not an intermediate, it is hardly possible to
escape the conclusion that the thing which brings about
or prevents the morphological alteration is a gradual
process. ¡What this gradual change may be in the present
case can not be known from the evidence, for obviously
changes in the type of metabolism may be of various
kinds.

Riddle (1917) conceives of such a change of metabolism
as a change from individuals having a high rate of metab-
olism and low energy content to individuals having a
low rate of metabolism and high energy content. In the
eggs of pigeons forced to lay eggs continuously, he finds
just such a change. The early eggs are of the former
type, the late eggs of the latter type. From the early
eggs are developed males, from the late ones females;
and on those facts, supported by other work, Riddle
bases an elaboration of the Geddes and Thompson theory
of sex.

An attempt has been made to fit the facts obtained from
aphids to Riddle’s conception of sex. The gradual
transition that occurs both in the parthenogenetic and in
the sexual phase of the cycle of Macrosiphum indicates
that one type of metabolism is prevalent early in the
eycle, and the contrasted type late in the cycle. The fact
that in this transition males precede sexual females shows
that, if Riddle’s hypothesis holds for the aphids, the pro-
gressive change is from a high rate of metabolism and
low energy content to a low rate of metabolism and high
energy content. Now it has been shown that wingless
viviparous females precede winged ones, the change from
the one form to the other taking place in part simulta-
neously with the transition from males to sexual females.
Hence in accordance with Riddle’s scheme wingless
females should represent a high rate of metabolism and
low energy content, while the winged ones should possess —
a low rate of metabolism and high energy content. With
regard to the rate of metabolism alone, this assumption is
supported by the fact that winged females require longer

518 THE AMERICAN NATURALIST [VoL. LII

to develop, that they produce fewer young per day, and
that these young are on the average smaller than in the
case of wingless females.

There are certain objections, however, to the foregoing
conclusion. First, if winged viviparous females have a
low metabolic rate while the wingless ones have a high
rate, during the parthenogenetic portion of the cycle a
parent with high rate of metabolism produces chiefly off-
spring with a low rate of metabolism, and vice versa;
for wingless females produce chiefly winged ones, and
winged females produce chiefly wingless ones. On the
other hand, in the sexual part of the cycle, parent and
offspring are both of the same metabolic type; for winged
females (with low rate of metabolism) produce mostly
sexual females (which in accordance with Riddle’s view
should possess a low rate of metabolism), whereas wing-
less females (high rate) produce mostly males (high
rate). Why parent and offspring should be of a similar
type of metabolism in the sexual phase, but of unlike type
in the parthenogenetic phase, is not clear.

Unless the withholding of food increases the rate of
metabolism, or unless the rate of metabolism is taken to
mean not the absolute rate, but the rate relative to the
food consumed, another objection to the assumption that
the winged female possesses a lower rate of metabolism
than the wingless ones is found in the work of Miss
Gregory (1917). Miss Gregory finds that in M icrosiphum
destructor starvation of the apterous mothers results in
the production of more winged offspring. It is only by
assuming that starvation increases the rate of metab-
olism, or that ‘‘rate of metabolism’’ means the relative
“shia tats relative to the amount of food consumed, not
relative to the rate in another type of individual—that
Miss Gregory’s discoveries can be interpreted in support
of Riddle’s hypothesis; providing, of course, that the
winged females have a lower metabolic rate than the
wingless females.

If, to avoid either or both of the difficulties just men:

Nos. 622-623] THE SEXES IN THE APHID 519

tioned, and notwithstanding the slower development,
smaller young and smaller daily output of young of the
alate females, these winged individuals be assumed to
have a higher rate of metabolism than the wingless ones,
other difficulties are encountered. This assumption
would have the advantage of allowing parent and off-
spring to be of opposite metabolic type in both the par-
thenogenetic and sexual portions of the cycle, instead of
being of opposite type in the parthenogenetic phase and
of like type in the sexual phase. The transitions, how-
ever, would be in opposite directions in different parts of
the cycle. In the parthenogenetic portion there would
be a transition from a low rate of metabolism to a high
rate (wingless to winged); while in the sexual part of
the cycle the transition would be from high rate to low
rate (male to sexual female). These opposite transitions
would have to occur in part simultaneously, as in Table
IV, fourth and fifth generations.

Unfortunately there has been no opportunity to deter-
mine experimentally the rate of metabolism in the various
kinds of individuals in Macrosiphum; that is part of the
program for the future. In the meantime, whether there
is a-fallacy in the foregoing argument, or a fallacy in
Riddle’s conception of the relation of metabolism to sex,
can not be asserted with any degree of confidence.

Obviously the mere rate is not the only feature of metab-
olism that may conceivably be related to sex. Tf there
are qualitative differences in the reactions that consti-
tute metabolism, it seems to me more likely that these
would influence tlie development of sexual organs than
that the production of ovaries rather than testes could
be determined by rate of metabolism alone. Qualitative
differences in the reactions might entail differences in the
rate of CO, production, and therefore be interpreted as
quantitative differences. ‘An increase in the output of
lumber from a sawmill might be taken to indicate that
the saws were running faster than formerly, whereas in
reality the saws had been replaced by a new type of saw.

520 THE AMERICAN NATURALIST [Vou. LIT

So long as rate of metabolism can be determined experi-
mentally, while the precise reactions can not, there is
every reason to continue the attempt to relate the rate
of reaction to the course of development. But when facts
come to light which do not easily fit preconceived ideas,
it is highly important that alternate possibilities be kept
in min

It is not impossible that the difficulties discussed above
may be removed by discovering that the metabolic change
that causes the transition from wingless to winged females
is different from the change that causes the transition
from males to sexual females. The two changes may be
more or less independent of each other. In that case it
may be possible to separate them experimentally. An
agent may sometimes be found which will hasten or post-
pone the sexual reproduction without in any way affect-
ing the transition from wingless to winged females in the
parthenogenetic phase. If this agent hastened the sexual
reproduction, it should act as a male-producing factor,
since sexual forms would be introduced while wingless
parthenogenetic females were more abundant. If, on the
other hand, the agent delayed sexual reproduction, it
should favor females, since the parthenogenetic mothers
would then be more largely winged.

BIBLIOGRAPHY
Gregory, Louise H.
1917. The Effect of Starvation on the Wing Development of Micro-
siphum destructor. Biol. Bull., Vol. 33, No. 4, October, pp.
296-303.
Riddle, Oscar.
1917. The Theory of Sex as Stated in Terms of Results of Studies on
Pigeons. Science, N. S., Vol. 46, No. 1175, July 6, pp. 19-24.
Pateh, Edith M.
1915. Pink and Green Aphid of the Potato. Maine Agric. Exp. Sta-
tion Bull,"242, October, 1915, pp. 205-223.

ORGANIC EVOLUTION AND THE SIGNIFICANCE
OF SOME NEW EVIDENCE BEARING
ON THE PROBLEM

PROFESSOR L. B. WALTON

KENYON COLLEGE

I

Tue biological problem recognized as having the great-
est fundamental importance at the present period is that
problem of evolution relating to the means by which the
heritable characters differentiating various organisms
from one another were first called into existence, or
granting the validity of the gene hypothesis and speaking
more concisely, how hereditary character-forming genes
have originated in the process of evolution. That the
diverse forms of life found upon the earth are only to
be explained as the result of organic evolution, is a prop-
osition which scarcely needs be mentioned at the present
period in the history of science, at least so far as indi-
viduals endowed with minds reasonably logical in evalu-
ating evidence are concerned. It is not evolution as a
process going on in the world which is being particularly
questioned nor the general method by which characters
once having originated are inherited, but the particular
method by which heritable characters first arose.

The purpose of the essay here presented is threefold.
First, that of pointing out the unsatisfactory nature of
much of the earlier evidence as a basis for sound general-
izations in connection with a clear understanding of evo-
lution. Second, that of calling attention to the serious
shortcomings of modern methodology in throwing light
on the causative factors of evolution. Third, that of pre-
senting some new evidence somewhat unique in its na-
ture, based in part on preliminary experimental work, to

521

522 THE AMERICAN NATURALIST [VoL. LE

the effect that the environment acting through long inter-
vals of time may impress characters upon an organism
which become unalterable by reversal processes.

To these propositions may be added the suggestion of
the fundamental importance which physico-chemical
methods must play during the future in solving the.prob-
lems of evolution.

II

The controversies relating to evolution have been
many. When, however, one considers the interest at-
tached to the subject, its broad bearing on various phases
of human welfare—sociology and economics in general,
animal and plant breeding in particular—together with
the difficulties of interpretation which apparently have
increased rather than diminished during the sixty or
more years seriously devoted to its elucidation, it is not
at all surprising that many different conclusions have
been reached, many dogmatic statements presented, and
many acrimonious discussions engendered.

In connection with a clearer understanding of the
points at issue, it will be well to pass certain historical
details relating to the development of the different
theories somewhat critically in review. This is done
even at the risk of a repetition of facts quite familiar to
those who have taken more than a passing interest in the
subject.

For long the theory of natural selection dependent on
the inheritance of small chance variations received gen-
eral acceptance. Championed by Weismann in his
notable controversy with Spencer to the exclusion of the
Lamarckian idea that characters acquired through en-
vironmental stimuli were heritable, it seemed at the time
entirely plausible as an explanation meeting the condi-
tions.

With the greater attention given to experimental meth-
ods, however, doubt arose concerning the fundamental
value of selection and resulted in the presentation of the
mutation theory by DeVries. Here evolution was inter-

Nos. 622-623] ORGANIC EVOLUTION 523

preted as arising from sudden and comparatively ex-
treme variations passed on by inheritance in nearly an
unchanged condition. Once more the results of experi-
mental work along the lines of the rediscovered principle
of Mendelian segregation indicated to a large number of
students of evolution that the facts set forth by DeVries
were subject to quite another explanation, in itself hav-
ing no bearing on the origin, but merely on the redistribu-
tion of the character-forming units already present in
the stock utilized. Another explanation not taking into
account the purity or impurity of the parental stock, ac-
counted for ‘‘mutations’’ through the sudden ineffective-
ness or loss of a gene.

The dissatisfaction thus arising resulted in the return
of many to the fold of “acquired characters.’’ Semon
(1912) reviving the ““mneme”” principle received the sup-
port of Wettstein, Przibram, and others. A disinclina-
tion existed, however, among most naturalists to accept
the evidence presented as seriously upholding the inheri-
tance of new characters produced by environmental
stimuli. Explanations of the results on quite other
grounds seemed more plausible. For example, the work
of Tower (1906), (1910), ete., in attempting to control the
color pattern of the potato beetle by changes in tempera-
ture and humidity, encountered the impurity of the germ-
plasm objection as well as the gene loss objection, either
one of which would be fatal to the validity of the conelu-
sions, if sustained. Commenced at a period in 1895, prior
to the rediscovery of the principles dealing with alterna-
tive inheritance, and finished in 1904 before the facts
were duly appreciated, it is not at all improbable that
genetic complications in the way of recessives, modifiers,
losses, lethals, ete., were involved. The destructive eriti-
cism presented by Cockerell, Gortner, Bateson, Castle,
and others, particularly in reference to the later studies
of Tower (1910), makes it evident that the results must
be confirmed from independent sources with more con-
sideration to the possible errors mentioned before the
conclusions are to be accepted.

524 THE AMERICAN NATURALIST [VoL. LIT

Similarly, the work of MacDougal (1907), in connec-
tion with the modification of Raimannia odorata, one of
the Patagonian primroses, may be explained. Compton,
as noted by Bateson (1912), using the same species, was
unable to obtain like results, while Humbert (1911) utiliz-
ing 7,500 pure line plants of Silene noctiflora, one of the
““pinks?” also failed to obtain so-called ‘‘mutants’’ simi-
lar to those found by MacDougal.

The investigations of Kammerer, Woltereck, Ferro-
niere, etc., are of decided interest, but to those critically
inclined they offer no evidence giving pronounced sup-
port to the proposition that environmental stimuli form
new genetic factors.

Thus, in turn, have the theories as to the method by
which evolutionary progress occurs been undermined
by doubt. Feeling the insufficiency of small chance varia-
tions, of environmental variations, and of larger germi-
nal variations, as a summation process, it is not to be
wondered that the truth-seeking pilgrim has become
wearied in his journey and longs for a more secure rest-
ing place.

Il

Let us return to the problem as suggested in the open-
ing paragraph, namely the actual origin of heritable
characters, and consider somewhat more carefully as to
whether theories exist justified by facts, which will
furnish acceptable evidence. There are two well-de-
veloped hypotheses, the general one of DeVries and the
more specific one of Morgan and his associates, founded
on discontinuous variations, and that of Castle based on
continuous variations.

Considering the views of DeVries and his followers in
the light of experimental investigations made during the
last ten years, it has become more and more evident that
by far the greater number, if not all, of the so-called mu-
tations thus obtained, were explainable on the basis of the
combinations of preexisting units of the germ cells. This
rests upon the proposition that there are present in the

Nos. 622-623] ORGANIC EVOLUTION 525

gametes certain hypothetical entities to which the name
gene or factor has been applied and which give rise to
the heritable characters of an organism. Thus it is at
once recognized that the problem relates to the origin of
the gene, rather than to the origin of the apparent char-
acters with which it is correlated, and that by far the
greater number of so-called new characters are not new,
but were performed at remote periods of time. So far
as the present arguments are concerned, it matters not
whether the results are assumed to be brought about by
material units or enzyme reactions. The prepared poten-
tialities exist in either case.

As examples of extreme types of characters which may
arise from the combinations of existing genes and which
might have been considered ‘‘mutations’’ at an earlier
period when the facts as to their origin were not fully
known, one need only mention the ‘‘blue’’ of the Andalu-
sian Fowl exhibited by the hybrid between the black and
white parental stock, or the ‘‘pink’’ presented by the
cross between the red and white ‘‘four-o-clocks’’ of Cor-
rens. A type of characters more in line with mutations
which have been described and to which there is every
reason for believing that many of them may be referred,
rests upon multiple gene effects combined with sterility,
in accordance with evidence presented by Davis, and oth-
ers. Of decided interest in this connection is a recent
paper by Muller (1917) calling attention to ‘‘An Gno-
thera-like case in Drosophila’’ where a result quite com-
parable to certain mutations of (Enothera is explained
through the action of balanced lethal genes. There are
other varying degrees of combinations from which **mu-
tant”? characters may arise and which depend on the be-
havior of the genetic material in connection with reces-
sives, modifiers, lethals, crossovers, non-disjunction, ete.

There is really nothing extraordinary in the appear-
ance and disappearance of the characters thus formed,
beyond their interpretation, and this has furnished false
premises for many erroneous conclusions, chief of which,

526 THE AMERICAN NATURALIST (Vou. LII

in the opinion of the writer, is the mutation theory as out-
lined by DeVries in so far as it may account for progres-
sive evolution.
Inasmuch as it seems probable that the results obtained
by Castle are to be explained upon the same basis as
those of DeVries, it will be well to consider them in this
connection. Here it is assumed that a continuously
variable heritable gene is involved, and that progressive
results are obtained through the selection of the ‘‘unit
characters’’ produced by such a gene. Castle, however,
stands almost alone in vigorous support of such a varia-
tion, while opposed to him are the Hagedoorns, Morgan,
Pearl, Punnett, McDowell, Muller and others equally in-
istent that genes once having originated pass on from
one generation to another unchanged except in compara-
tively rare instances where so-called ““mutations”” occur."
It is maintained by those advocating this view that the
results in connection with hooded rats on which Castle
bases his contentions, are due to an uncertain number of
modifying genes not in themselves variable, and that the
existence of such genes has been demonstrated in other
organisms presenting results similar to those obtained
in rats. The work of Little (1917) with mice where
three segregating types of spotting were found to pro-
duce varying degrees of color pattern, indicates that
multiple genes are involved. Furthermore, the analysis
by Little of the data obtained by Castle, Phillips and
Wright, points decidedly to the interpretation of their
1 Jennings (1917) has recently endeavored to show that the views of
Castle and his opponents are identical, This, however, is by no means the
ease. On the one hand there is the idea of a continually variable gene (coat-
color-producing gene in at), moved gradually along a given scale by selec-
tion. On the other hand there is the idea of a rarely mutating gene (e. g.,
sais insted eye color producing gene in Drosophila) moving abruptly from
one part to another of the scale. Its position once obtained remains for a
time constant. These differences of interpretation are at present
cilable. de
Since this note was written, Morgan (1917) has discussed the matter in
detail, presenting arguments quite similar to those mentioned above, and
arriving at a similar conclusion.

Nos. 622-623] ORGANIC EVOLUTION 527

results on the basis of multiple genes instead of a con-
tinually varying gene.

It would thus appear evident that the theory outlined
by Castle is open to quite the same objections that oc-
cur in connection with the mutation theory of DeVries,
and that there is little evidence for believing that it has
any fundamental value in explaining evolution.

The mutation theory of Morgan and his associates,
based primarily on results obtained in studies of the
small ‘‘fruit-fly’’ Drosophila, apparently presents quite
another view of the subject. Here it is clearly indicated
that evolution has taken place through the incorporation
of mutant changes, and that these changes are due to dis-
continuous ‘‘mutations’’ of genes as exemplified in mul-
tiple allelomorphs.

Assuming the validity of the arguments based on link-
age relations in respect to the localization of the genes,
the conclusion follows that the ‘‘mutation’’ results either
(1) from a change in a specific gene or (2) from the com-
plete linkage of a series of genes. If the latter proposi-
tion should be the correct interpretation, and it is by no
means clear that it is not, the objections urged against
the theories of DeVries and of Castle hold equally here.

Morgan and several others have presented evidence for
believing in the specific change of a gene. Granting that
this is the actual explanation of the facts presented in
connection with multiple allelomorphs, etc., there are two
lines of argument leading to the conclusion that these
changes are results of combinational sub-units or sub-
genes existing in the species, and that progressive evolu-
tionary changes are no more represented here than in the
previous theories of DeVries and of Castle.

The first argument (a) rests upon the recurrent ‘‘mu-
tations’’ which have been noted in a considerable number
of cases. Thus the sex-linked eye colors of Drosophila
forming the multiple allelomorph system consisting of
white, eosin, cherry, blood, tinged, and buff, and their
dominant allelomorph, red, of the wild fly, have their

528 THE AMERICAN NATURALIST [VoL. LIT

origin from a single definite area or locus in the “X”
chromosome, accepting linkage as a criterion. They have
not arisen in a continuous series but as sudden changes
from one extreme to another at comparatively long inter-
vals. The character may remain modified in one direc-
tion and then suddenly revert to an original condition.
Thus white changed to eosin and later back to white as
noted by Morgan (1916). Furthermore, the changes are
not extremely infrequent. A similar transformation has
been noted by Emerson (1917) in maize where self color
apparently changed to variegation and later back to self
color. A variation which may be of the same type has
been described by Shull (1911) for Lychnis. Quite re-
cently Zeleny (1917) in studies on Drosophila melano-
gaster Meig. (—=ampelophila Low)? has noted a reversed
mutation where full-eyed flies result from the return of
the bar gene to the original full-eyed condition. In each
of the cases mentioned the germinal purity of the stock
was believed to be without question.

Such results are not to be attributed to a continuous
series of mutations, to progressive changes, or to genetic
losses. They clearly suggest that the gene, if it is the
individual gene which is involved, is made up of smaller
combinational units which through their permutations
give rise to the characters in question. Presumptive evi-
dence is certainly furnished against the idea that any-
thing new has developed in the organism to form the par-
ticular characters. Furthermore, one may well believe
that any particular mutation under observation suffici-
ently long, will exhibit recurrent changes.

The second argument (b), to the effect that the gene is
comparatively stable and that “*mutations”? are only
transitory combinational changes, is based on the main-
tenance of apparently identical genes through long
periods of time. Thus Metz (1917 ) found that the three
mutations which had, up to that time, oceurred in Droso-
phila virilis Sturt. appeared exact duplications of the

2 Sturtevant, mss.

Nos. 622-623] ORGANIC EVOLUTION 529

mutations in Drosophila melanogaster Meig.? In both
species ‘‘confluent,’’ a modification of the wing venation,
is similar in form, dominant over ‘‘normal’’ and :
““lethal,”? when the fly is homozygous for the character.
The characters ‘‘yellow’’ and ‘‘forked’’ are sex linked
in both species and otherwise alike so far as the evidence
exists. Inasmuch as the earliest representative of Droso-
phila thus far known is a species not decidedly different
from those now existing as noted by Löw (1850), who de-
seribed it from the amber of the Baltic Sea, and belongs
to the Lower Oligocene of the Tertiary Period, with an
age of from two millions to three millions of years, one
must infer that the genes common to the two species men-
tioned have been preformed for a long period of time,
and that nature has paid little attention to such muta-
tional changes as occur in connection with multiple allelo-
morphs.

There are certain investigations widely separated as
to their content, but apparently closely correlated as to
the underlying explanatory principles involved, which
must not be overlooked in a consideration of the changes
which may take place in hereditary units. These are
concerned with the differences involved in metabolism.*

On the one hand there are studies dealing with the di-
rect effects of a changed metabolism on the developing
individual. Here may be mentioned the work of Lillie in
connection with the ““free-martin”” of cattle; Steinach on
the transplantation of the gonads in rats and guinea-
pigs; Goodale on the grafting of ovaries in male fowls;
Pearl and Surface on the degeneration of the ovary in
cattle; Riddle with pigeons, etc. On the other hand, there
are studies dealing with the indirect effects on inheri-

3 The species are distinctly separated not only in external appearance but
also by their chromosome number. D. melanogaster has four pairs, while
D. virilis has five pairs of chromosomes. ds

4 The theory has had a long historical development. Treat (1873) pub-
lished a paper on controlling sex in butterflies as a result of food supply.
Yung (1881) worked with tadpoles. Nussbaum (1897) with rotifers.
Recent evidence of an elaborate nature has ‘been presented by Goldschmidt,
- Woltereck, Whitney, Banta, Shull and others.

530 THE AMERICAN NATURALIST [Vou. LIT

tance. Among thesé may be mentioned that of Gold-
- Schmidt with moths; Woltereck with daphnids; Plough
with temperature effects on Drosophila; Hoge with the
effects of cold on Drosophila; Morgan with the effects of
moist food supply on Drosophila, ete.

As an example of the development group, the investi-
gation of Lillie may be noted. The evidence obtained
showed that the ‘‘free-martin’’ or sterile female usually
developing where the twins are of separate sexes in cat-
tle, ete., resulted from the modifying influences of the sex
hormones in the male where the two chorions had anas-
tomosed.

As an example of the inheritance group, Morgan has
found that the ‘‘mutant’’ ‘‘abnormal abdomen ”’ in Droso-
phila occurs in connection with a moist food supply. The
character is a sex-linked dominant. If an abnormal male
is bred to a normal female and the food is kept moist, the
sons are normal and the daughters abnormal. If the food
is dry both sons and daughters are normal. The recipro-
cal cross gives sons and daughters both abnormal with
moist food but normal with dry food.

It follows then that in Drosophila the gene for the ab-
normality—or the chemical preparedness for the inhibi-
tion of normality, if one so wishes to term it—is per-
formed in the ‘‘X’’ chromosome and merely awaits a
suitable environment before presenting itself as a char-
acter. Similarly, in connection with the changes occur-
ring in the development of the ““free-martin”” of cattle,
it seems necessary to admit that there are genes present
in the sex chromosome concerned with the development
of sexual characters which, however, are in a state of
equilibrium, and that the inhibition or the excitation® of
one or the other genes or groups of genes will result in
the development of the corresponding individual.

From the facts presented, one seems justified in making
the deduction that heredity hands down a framework

5 It has been shown by Chapin (1917) that the gonads of the free-martin
_ originally destined to be a female, attain a male condition.

Nos. 622-623] ORGANIC EVOLUTION 531

which within certain limits allows a plasticity in the de-
velopment, and that the direction of development is de-
termined by physico-chemical influences through the
suppression of potential units.

Thus the conclusion seems almost unavoidable that by
far the larger number, if not all, of the heritable charac-
ters making up an organism, result from combinational
units which have long been predetermined, and that the
breeder, whether the semi-scientific agriculturist or the
ultra-scientific drosophilist, is largely, if not entirely, en-
gaged in presenting new combinations of existing units.
If this be true, modern genetics has left the actual prob-
lem of evolution far to one side and deals only with re-
sults of a secondary, although none the less interesting,
nature.

One is, therefore, led to inquire as to whether there
may be available evidence which will permit a new insight
into conditions governing the formation of characters,
even though the evidence from its nature must be largely
circumstantial.

IV

Accompanying the progressive swimming movement
of most aquatic microorganisms there is a characteristic
axial rotation. This has been noted by Nigeli, Engel-
mann, Strassburger, Mast and more in detail by Jennings
(1901) who has called attention to the value which such a
compensatory motion may have for the organism in which
it exists. No explanation has been suggested other than
this as to the origin of the rotation, and without further
thought it is evident that one would be inclined to attrib-
ute it to natural selection, assuming that those individ-
uals in which it did not occur were at a disadvantage in
the struggle for existence by reason of their more con-
fined movement.

It is the phase of the question dealing with the par-
ticular causes bringing about the rotation that appears
to be of extreme significance when considered in connec-
tion with the principles underlying evolution and to be |

532 THE AMERICAN NATURALIST [Von. LII

susceptible to quite another explanation than the natural
selection implied by the term ‘‘adaptiveness’’ which, in
accordance with Jennings (1906), is based on the idea
that ‘‘it tends to preserve the life of the
animal.’? Furthermore, when the groups
of facts associated with the character-
istic rotation are brought in review, it
would seem that the explanation sug-
gested may go far toward interpreting
the origin of the fundamental activities
as well as the origin of- the characters in
general of organisms.

In connection with the preparation of
a systematic review of the order Eug-
lenoidina belonging to the class Flagel-
lata of the Protozoa (1915), it was noted
with decided interest that a large number
of the forms possessed an oblique stria-
tion ranging from almost indiscernible

markings to characters of great com-

ec plexity impressed upon a cellulose-like
eL etriotacomoction Cuvelope (E g., Euglena spirogyra
with clockwise toia- Hhrenb., Phacus pyrum (Ehrenb.), Het-
Dont) and process < FOREGO spirale Klebs, ete.), the striæ
means of anterio €Xtending forward and to the left. The
flagellum. character also appeared to be invariably
correlated with an axial rotation of the organism from
right over to left (Fig. 1). Such a movement is to be de-
scribed in physical terms as ‘‘clockwise,”’ the position of
the observer being in front of the advancing organism.

The facts took on additional interest when it was noted
that forms with a reverse striation seemed entirely ab-
- sent from the northern hemisphere, although such forms
existed in the southern hemisphere.

Inasmuch as the euglenoids are in general positively
phototactic under normal conditions, it would immedi-
ately occur to one seriously considering the question, that

Nos. 622-623] ORGANIC EVOLUTION 533

the underlying principle producing the rotation was the
turning of the earth on its axis, with the resultant appar-
ent motion of the sun from east to west. Such a hy-
pothesis would become the more tenable when it was
found that negatively phototactic microorganisms of the
northern hemisphere rotated as a rule in a reverse or
counter-clockwise direction.

Some of the evidence thus far obtained may be pre-
sented more clearly in tabular formë (Table I). Thus

TABLE I
A series of aquatic microorganisms showing in general the clockwise rotation
of positively phototaetic forms and the counter-clockwise rotation of nega-
tively phototactic forms in the northern hemisphere, with evidence for the
tendency to reverse condition in the southern hemisphere

Northern Hemisphere

peN IES. URI a aAa a Positive. Clockwise,
Eu prin ieai M ea eres Positive. Clockwise.
fist a ia DU e tec rt ae ee Positive i
Huplena epiregyra bres. ee ii ee. Positive Clockwise
Leptocinelis ovum (Ebr.). +...» ... Positive Clockwise
o a eats a oe kis os ositive Clockwise
ryptamonas ovata ERT, ..........«..... Positive Clockwise.
P 1 (MAR) Ta a Positive Clockwise
Wind elegans (BNE) IP Positive i
Volvox glo PEN A E Arms E E Positive. Clockwise.
Stentor polymorphus Ehr. (= wni.. . Positive. Counterclockwise.
Phacus longicauda Ehr. ..............-. Positive. Counterclockwise.
Stentor: curulis ET. oons eiei is ee. Negative. Counterclockwise.
amre cpeblearía Goe r-ri faite oc Negative. Counterclockwise,
Arenicola cristata (larva) ......... wipes egative. Counterclockwise,
Chilomonas paramecium Ehr.1 .......... Negative. Counterclockwise.

Leptocinclis piriformis Cun. ........ „as Positive? Counterclockwise.
Phacus bactilifer Cam, .. siei- evs cee ees Positive? Counterclockwise.
Leptocinclis mammilata Cun, ...........- Positive? Clockwise

it may be stated that so far as the facts are available,

positively phototactic forms with the exception of
6 The rotation direction and light responses noted are those taking place

under normal conditions. The conelusions presented are not altered by the

fact that as a result of stimuli under conditions imperfectly known, reverse

movements may occur, e. g., the Seealive response of catas. to intense
ight.

534 THE AMERICAN NATURALIST [VoL. LIL

Stentor polymorphus Ehrenb. (=S. viridis aut.) and
Phacus longicauda Ehrenb. rotate clockwise in the north-
ern hemisphere. Inasmuch as the phototactic relation of
the ciliates in general is negative, where a reaction exists,
it seems probable that the inclusion of the minute sym-
biotic forms of alge, Chlorella vulgaris Beyer, which
gives the species its characteristic green apperance, has
induced a change from a negatively phototactic to a
positively phototactic condition, while the organism re-
tained its original counterclockwise rotation. Small
forms like Chlorella which contain chloroplasts, are gen-
erally positively phototactic so far as their responses to
normal conditions are known.

Phacus longicauda (Ehrenb.) is an euglenoid about
100» in length, with comparatively flat wing-like expan-
sions. The striæ covering the body are longitudinal. In
swimming, however, many of the forms have the anterior
part of the right expansion turned slightly down, while
the left expansion is turned up ina similar manner. This
gives their progressive movement a counterclockwise
rotation.

In the southern hemisphere direct observation of the
characteristic rotation has not been made, but inasmuch
as the direction of the striæ indicates the direction of the
rotation, certain evidence is available. Cunha (1913) in
his studies of Protozoa from Brazil has figured several
forms showing distinctly the arrangement of the strie -
in the excellent plates drawn by himself. While it is not
impossible that a careless investigator might focus on the
lower part of the specimen, thus showing the reverse
position of the striæ, the careful work of Cunha scarcely
permits one to suggest such a criticism. It may further-
more be noted that the apparently counterclockwise ro-
tating forms described by him are species quite different
from the typical northern forms, while the forms which
_ evidently rotate clockwise are closely allied to species
from the northern hemisphere, and may have been intro-
duced comparatively recently so far as evolutionary time
is concern

Nos. 622-623] ORGANIC EVOLUTION 535

The original development of the unicellular forms in
the northern hemisphere with their subsequent introduc-
tion to the southern hemisphere by aquatic birds, ete., is
well within the range of possibility and suggests that
even should forms with reverse rotations be entirely ab-
sent from south of the equator, the hypothesis which has
been proposed would by no means be invalidated.

Having presented the general facts as to the behavior
of free-swimming microorganisms, it becomes advisable
to consider the explanations which may exist as to the
origin of the characteristic rotation. It seems impossible
to attempt to account for such a character on the ground
of ‘‘natural selection.?”? One would be compelled to be-
lieve that the reverse rotation—the counterclockwise ro-
tation of positive northern forms—possessed an elimina-
tion value, an almost indefensible proposition, particu-
larly when forms like Stentor polymorphus and Phacus
longicauda are considered as well as forms in the south-
ern hemisphere which do not rotate in agreement with
theory.

The most obvious explanation to be considered is that
based on the influence which the sun in its apparent daily
movement from east to west in the equatorial region may
be supposed to have exerted on the flagellum (Fig. 2).
This assumes that the flagellum is the orienting factor
and that the sun has induced in it an east-west rotary-like
or whip-like propelling movement. The consequent me-
chanical effect would be that in the northern hemisphere
forms with a positive light reaction would rotate clock-
wise and those with a negative light reaction would rotate
counterclockwise. Conditions would be reversed for
those which might be present in the southern hemisphere
during the evolutionary stage.

Conversely, negatively phototactic forms would de-
velop a reversed or counterclockwise rotation by means
of the influence of the light rays on the stroke of the light
avoiding flagellum and their modified organs, the cilia.
It is by no means necessary to believe that the stroke of

536 THE AMERICAN NATURALIST [VoL. LIT

the flagellum should be one of rotation, although theory
would imply a partial rotation in the primitive flagellate
forms. The method of movement by means of the flagel-
lum furnishes a problem of considerable difficulty which

o. 2. Il amr the theory as to the origin of axial rotation in aquatic

microorganisms of the northern and southern hemisphere of the earth by the

ap) peck tin pr the sun from east to west. N, etc., north, etc.; F, flagel-
+ i

e and ri
(observer in at 4 ve oe ny ot Arar pits send site
primitive rotation
has received attention from several investigators, notably
Delage and Herouard (1896), Goodspeed and Moore
(1911), Biitschli and others.
| ere are several inferences of an axiomatic nature

Nos. 622-623] ORGANIC EVOLUTION 537

that follow from such an hypothesis. Forms near the
neutral equatorial region may be assumed to possess a
slower rotation than forms near the poles and at the
same time there may be expected to occur a change in the
relative angle which the stri# make with the longitudinal
axis of the body, their direction becoming approximately
parallel with that axis.. The cosmopolitan distribution
of unicellular organisms with the evident non-selective
value of the character makes such a hypothesis difficult
of demonstration. The application of statistical meth-
ods would be of interest, however.

A second explanation of the rotation direction, appar-
ently, however, a negligible one, is on the basis of the
angular velocity of the earth so far as it may have an
influence on small bodies at its surface. With free-
swimming microorganisms oriented in accordance with
the axis of the earth during definite intervals and rotat-
ing in the same direction that the earth rotates, condi-
tions are fulfilled for such a mechanical explanation.
When, however, the relative dimensions of the earth and
the organisms as well as the relative density of the earth,
the water and the organism, are considered, it is difficult
to believe that the explanation lies in this direction.
While many of the forms are attached to some definite
surface in the water during certain periods of their de-
velopment, there are others which reproduce directly in
the water and should this have been the primitive condi-
tion of development, the rotation of the earth would have
been ineffective. |

While the possibility of electrical forces may be men- _
tioned as an influence, there are no facts known which al-
low an interpretation in this direction.

During the. past two years a considerable number of
experiments have been made in attempts to obtain some
definite evidence as to the cause of the rotation. Obvi-
ously it would be of interest to maintain a culture of
northern forms in the southern hemisphere or a culture
of southern forms in the northern hemisphere. Efforts

538 THE AMERICAN NATURALIST [Von. LIT

to obtain living cultures from desirable localities, the
Falkland Islands, New South Wales, the southern part of

outh America, etc., have thus far failed. A method,
however, was devised by which it seemed theoretically
possible to subject northern flagellates to an environment
similar to that of the southern hemisphere. A clinostat
(Fig. 3) having a clockwise rotation of fifteen minutes

Fic. 3. A clinostat arranged for the purpose of subjecting northern flagellates,
etc., to an apparent west-east movement of the sun, the covered portion being
toward the north.

was procured, a circular table ten inches in diameter
fitted to this, and the northern half covered so that the
revolving table containing slides in excavated recesses
would pass into darkness on one side and emerge moving
from east to west. Thus the apparent path of the sun
so far as the organisms were concerned would be from
west to east. The larger unstriated Euglena have been
used almost entirely in the experiments, inasmuch as it
would apparently be impossible to change the direction
of rotation in forms like Euglena spirogyra Ehrenb.,
Phacus pyrum (Ehrenb.), ete., where the strim are cari-
nate in structure with an angle almost if not entirely pre-
cluding a rotation in the direction opposite to that in
which they were accustomed to turn.

4

A

Flagellate forms

rom

a f t
(Ohio); B, Euglena spirog

C
phere illustrating the development of the left-hand s
); O, Urceolus costatus Lemm. (Europe); D, Heteronema spirale Klebs (Europe).

~

D

triæ. A, Euglena hemogranulata

[€Z9-ZZ9 “SON

NOILATOAY JINVIYO

540 THE AMERICAN NATURALIST [VoL. LIT

While there has been a large amount of data obtained,
thus far no evidence shows that either a ‘‘reversal’’ or a
““slowing up”” of the rotation may be produced in any of
the individuals utilized.

Even though it may not be possible to change the direc-
tion or the period of rotation in ‘‘adult’’ forms, may not
such changes be produced in encysted forms or during a
period when gametes are developed. Experiments are
yet to be made with individuals in an encysted condition,
and with material available it will be possible to utilize
gamete-producing forms. That Euglena has a sexual
cycle was pointed out by the writer nearly ten years ago
(Walton, 1909).7 Certain forms encyst, the cysts subdi-
viding to approximately a 16-cell stage, small flagellate
gametes emerge and conjugate. An experiment of this
nature involves a discussion of the environmental effect
on germ cells as compared with somatic cells, but does
not affect the issues with which we are concerned in the
present paper.

There are many other questions of interest which arise
in a study such as outlined. For instance, what has been
the origin of the striæ which are much specialized in many
forms, although entirely absent in other forms (Fig. 4) so
far as visibility with the microscope is concerned. The
majority of the positive northern forms have ““left-
handed” striæ, a smaller number have longitudinal
striæ, while a considerable number appear to have no
strie. None have been found with “*right-handed”
strie. At first one may be inclined to attribute such a
character to natural selection, but when one commences
to ascertain the value of the character on the basis of pro-
gression, rotation and axial angle, such a conclusion
seems less certain. There are a few facts that appear
evident. First that the development of the striw has

been at a considerably later period than that of the rota-

tion direction. Second that the relative position of the
strie has been largely dependent on the rotation. Third,
7 Paper presented at annual meeting of Ohio Academy of Science, 1908.

Nos. 622-623] ORGANIC EVOLUTION 541

that the development of the striæ has in many forms pro-
ceeded so far that a reversal rotation seems an impossi-
bility.

The presence of a considerable number of other groups
which have ‘‘left-handed’’ spirals so far as observation
goes, is of interest. The various genera of Spirochetes,
as well as Spirulina and Arthrospira among the Cyano-

Fic, 5. Flagellate forms from the southern hemisphere, Rio de Janeiro, Brazil
(from Cunha), illustrating the development of ‘sad right-hand strie. A, Phacus
bacilifer Cunha; B, Leptocinclis piriformis Cun

phycoidea (Cyanophycee) may be mentioned. The
twining of plants may, in the final analysis, be attrib-
utable to the same cause.

Other related problems are the pendulation theory of
Simroth (1912) relative to bipolar distribution, and the
tropism theory as outlined by Verworn (1894), in con-
nection with the excitation contraction of the flagellum.

542 THE AMERICAN NATURALIST [Von. LII

y

Having indicated some of the difficulties existing along
the lines of established research in the efforts to account
for the derivation of the fundamental heritable units
making up an organism and having presented a series of
observations suggesting that a new perspective may be
obtained by utilizing methods of attacking the problem
somewhat different from those hitherto employed, the
particular purpose of the paper is accomplished.

It may be asserted by some that such an attempt at a
summary disposal of the existing evidence as to the actual
origin of characters represents an unfortunate type of
destructive criticism. Furthermore, that the acceptance
of the validity of the arguments as to the long predeter-
mined nature of the genes or subgenes, leads us once
more in the direction of the somewhat antiquated theory
of preformation. It is not impossible that these views
are in part justified. Nevertheless it is well within the
bounds of propriety to occasionally inquire as to whether
the enthusiasm developed for a special discovery has
not resulted in too broad an application of its principles.
The mutation theory particularly as developed by Mor-
gan is of interest. It is circumscribed at least in part
by the phenomena of mendelian inheritance, and it is
evident that one should look farther for the facts which
may assist in explaining the real origin of the diversity
of living things.

Tf additional studies support the view suggested by the
facts here presented, namely, that characters of a physio-
logical nature may be produced by environmental causes,
and that these in turn may demand the correlation of
morphological characters regardless of the origin of the
latter, an important step will have been made in account-
ing for the primary differentiation of organisms. The
later secondary differentiation through the combinations
‘of units which have thus arisen, and which attains its
maximum in the complex multicellular animals and
plants, offers no particular difficulties in its explanation.

Nos. 622-623] ORGANIC EVOLUTION 543

Furthermore, such an idea is more in harmony with the
paleontological evidence as presented by Osborn (1912)
and others, than one based on the mutational idea, and
it is to fossil forms that one must look for the all-impor-
tant historical record.

Should one propose a hypothesis of an ultimate unit,
slightly plastic as to its immediate environment, but sub-
ject to the permutations and combinations of a mendelian
type, and possessing a definite qualitative condition de-
termined by prolonged environmental action, the picture
is not at all so fanciful as some might at first thought
insist.

The practical importance of such a viewpoint in its
application to the problems of animal and plant breed-
ing lies in the realization that new forms can not be cre-
ated, but merely new combinations uncovered during the
comparatively brief epochs of time which human intelli-
gence has for working out the processes. Thus one re-
turns to genetics.

In summarizing the paper, the following conclusions
are suggested:

I. The heritable characters in general which make up
an organism arise from preformed units in the nature of
genes or subgenes that have been in existence during long
geological periods of time. There are at present no cri-
teria available in modern genetics by which an appar-
ently new gene may be distinguished from one long in
existence; furthermore, there is doubt as to whether new
genes. are actually arising in multicellular organisms.
The change of a gene in a given direction, whether it be
considered as a morphological unit or a chemical condi-
tion followed by the return to its original condition, sug-
gests its composition of combinational sub-units, and is
an argument against the idea that anything actually new
has come into being during its series of so-called muta-
tions. Such a conclusion receives additional support
from the presence of apparently identical genes which
exist in distinet species of organisms separated during

ý

544 THE AMERICAN NATURALIST [Vov. LII

long epochs of time, as well as from the evidence of the
non-contamination of genes during diverse environments.

II. The mutations demonstrated by DeVries and oth-
ers, together with the variations obtained by Castle, are
to be interpreted as a result of the combinations of exist-
ing genes. The mutations noted by Morgan and his as-
sociates, as evidenced from recurrence and stability, are
in the nature of modal fluctuations having no definite
cumulative value.

Ill. The direction of axial rotation in aquatic micro-
organisms is best explainable on the basis of the appar-
ent east-west motion of the sun having influenced the
movement of the organs of locomotion. Thus the charac-
ter becomes one acquired from external stimuli, and the
persistence of reverse forms in both the northern and
southern hemispheres indicates the hereditary nature of
the character. Morphological characters, such as the
striæ, etc., may arise in a similar manner or through se-
lection. By correlation with the preceding characters, a
cumulative and irreversible effect is produced.

IV. The primary factors in evolution are environ-
mental and thus dynamic. The secondary factors of a
combinational nature are gradually approaching a limit-
ing Maximum value, and are thus becoming static.

BIBLIOGRAPHY
Bateson, W.
1913. Problems of Genetics. (New Haven, Yale University Press.)
eee C. B.
Non-Disjunetion as a Proof of the Chromosome Theory of
Heredity. Genetics, Vol. 1, pp. 1-52, 107-1
1917. e Genetics, vel 2, pp. 445-465,
Castle, W.
1916. taai and Eugenics. (Cambridge, Harvard University Press.)
See bibliography for earlier papers of Castle.
1917. Piebald Rats and Multiple Factors. AM. Nar., pp. 102-114.
1917. Rôle of Selection in Evolution. Journ. Wash. Acad. Sci., Vol.
6, pp. 369-387.
Cunha, A. M. da :
1913. Beitrag zur Kenntnis der Protozoenfauna Brasiliens. Mem.
Inst. Oswald Cruz, pp. 1-13.

Nos. 622-623] ORGANIC EVOLUTION 545

Davis, B. E.
1917. ‘Some Inter- pl de aten of F, Gnothera Hybrids. Genetics,
, pp. 155-185.
East, E.
1912. ei Mendelian Notation as a Description of Physiological Facts.
Am, NAT., p. 633

1917. The Rearing of Some General E ge Facts on Bud Varia-
tion. AM. NAT., pp. 129-143
Emerson, R
1917. cala Studies of Variegated Pericarp in Maize. Genetics,
Vol. 2, pp. 1-35.
Engelmann, T. W.
Ueber experimentelle Erzeugung zweckmássiger Anderung no
pflanzlichen Chromophylle durch farbiges Licht. Arch. Ana
Physiol. (Phys. Abtlg.). (Also see Gaidukov, 1902, 1906, om
Goldschmidt, R.
1916. Genetic Factors and Enzyme — ae Vol. Ea
1916. Experimental Sexuality and t oble AM
705-7 18, (See bibliography rd waited ieme ‘of senna 5
Goodspeed and Moore.
1911. Univ. Calif. Pb. on Vol. 4, p. 17.
Hagedoorn, A. C. and A. L.
1914, Can Selection Improve the pen of a Pure Strain of Plants?
Journ. Board of Agricultu
1914, pes on set and isc Zeit. Abst. Verbung, pp.
5-183. i

1917. ie and Evolution. AM. NAT., pp. 385-418.

M.
1915. The ‘Tnfivence of Temperature in the Development of a Men-
delian Character. Journ. Exp. Zool., Vol. 18.
Humbert, E. P.
1911. A Quantitative Study of Variation, Natural and Induced, in
e Lines of Silene noctiflora. Zeit. Abst. Verbung, pp.
161-286,
Jennings, H. S.
1901. On the Pa of Spiral Swimming of Organisms. AM.
Nat, pp.
1906. Behavior of =e oe Organisms. (New York, Columbia
niversity Press.)
1917. Modifying Factors and Multiple Allelomorphs in Relation to
the Results of Selection. AM. NAT., pp. 301-306.
Kammerer, P.
1910. Beweise fiir der Vererbung erworbener Eigenschaften durch
planmassige Zuchtung. Vol. 12, Flugschr. D. Ges. f. Zuchtungs-
kunde Berlin. (Also note other papers of Kammerer’s in Bib-
liographies.)
Lankester, R.
1917, The Problem of Heredity. Nature, p. 181.
Little, C. ©.
1917. Evidence of Multiple Factors in Mice and Rats. Am. NAT., pp.
457-480.

546 THE AMERICAN NATURALIST — [VoL. LII

Mast, $.
oe ‘Bites of Chemicals on Reversion in Orientation to Light in
Colonial Form Spondylomorum quaternarium. Journ.
ke Zool., Vol. 26, No. 3, p. 503-520.
McDowell, E. C.
1914. oe Inheritance in Rabbits. Carnegie Institute, Washington,
196.

1916. ee Rats and Multiple Factors. AM. NAT., pp. 719-742.

Metz, C. W.

1917. Mutation in Three Species of Drosophila. Genetics, Vol. 1, pp.
91-607.

Moore, A. R.
1916. ee Menyene of Sane in Gonium. Journ. Exp. Zool.,
, pp. 431—
Morgan, T. me
1915. The Rôle of the Environment in the Realization of a Sex-Linked
Mendelian Character in Drosophila. Am, Nar., Vol. 49, pp.
385-429.

1916. A Critique of the Theory of Evolution. (Princeton, Princeton
University Press.)

1917. An Examination of the so-called Process of the Contamination
of the Genes. Anat. Record, pp. pnp

1917. The Theory of the Gene. AM. NAT., pp. 513-544.

1918. Evolution by Mutation. Sci. Mo., pp. or VOL, TF No: L

Morgan, Sturtevant, Muller and Bridges

1915. The Moca of Heian Heredity. (New York, Henry

H

Muller, H. J.
1917. ja aan -like Case in Drosophila. Proc. N at. Acad. Sci-
e, Vol. 3, pp. 619-626.
Osborn, H. nae
1912. The Continuous Origin of certain Unit Characters as observed
by a Paleontologist. AM. NAT., pp. 185-206, 249-278.
Pearl, Raymond.
1917. The Selection Problem. AM. NAT., pp. 65-91.
Semon, R.
1910, Der Stand der Frage nach der Vererbung erworbener Eigen-
aften. Fortschr. natur. Forsch., Vol. 11.
1912, Das Problem der Vanha ‘‘erworbener’’ Eigenschaften. Pp.
1-203. (Leipzig, W. Engelmann.)

1917. "The — of Temperature upon Facet Number in the bar-eyed
of Drosophila. Abstr. Am. Zool. Soc., Dec., pp. 14-15.

E
1912. Zur Pendulationtheorie. Petermann’s Mitt., pp. 268-269.
1911. Reversible Sex Mutants in Lychnis dioica. Bot. Gaz., Vol. 52.
906. An Investigation of Evolution in the Chrysomelid Beetles of the

genus Leptinotarsa. Pp. 1-320. Carnegie Inst. No. 48.
deny The Determination of Dominance. Biol. Bull., pp. 285-337.

Nos. 622-623] ORGANIC EVOLUTION 547

Trolland, L.,

1917. Bake et ne and the hee of Enzyme Action. Am.

NAT., pp. 321

Verworn, M.

18 Allgemeine Physiologie. (Jena, G. Fischer.)
Wager, H.

1914, Movements of ego Organisms in Response to External

ees, (London.)

Walton, L. B.

1914. The Evolutionary 4 oas of Organisms and its Significance.

Science, pp.
1915. Variability de Aiks imixis. AM. NAT., pp. 649-687.
1917. The Axial Rotation of Aquatic Microorganisms and its Signifi-
cance. Ohio Journal Science, pp. 6-7.
Woltereck, R.
1911. Beitrag zur ‘‘erworbener’’ Eigenschaften. Verh. D. Zool. Ges.,
pp. 142-172. (See bibliographies as to various papers of
Woltereck.)
Zeleny, C.
j 1917. Fulleye and Emarginate Eye from Bar- -eye in Drosophila
ithout Change,in the Bar Gene. Abstr. Am. Zool. Dec. AS
1917. Selection for High-facet and for Low-facet Number in the Bar-
eyed Race of Drosophila Abstr. Am. Zool. Dec., pp. 9-10.

SHORTER ARTICLES AND DISCUSSION

RHYTHMIC SYNCHRONISM IN THE CHIRPING OF
CERTAIN CRICKETS AND LOCUSTS

SYNCHRONISM in the rhythmic chirping of the snowy tree
cricket, (Ecanthus niveus De G., has been observed and men-
tioned by a number of able observers, including Burroughs,
Thoreau, McNeill and Dolbear. The synchronous chirping of
this particular cricket has been too generally observed to be con-
sidered merely an illusion of the mind. In a previous paper’
I reported that I had observed the occurrence of rhythmic syn-
chronism in the chirping of colonies of the tiny tree crickets,
Cyrtoxipha columbiana Caudell in Georgia. During the sum-
mer of 1917 I was afforded an excellent opportunity to make
further observations of the synchronous chirping of these inter-
esting arboreal crickets near Vinson Station, Virginia. A little
colony had become located in the crown of a small black cherry
tree in my back yard, where I could readily keep them under
observation at all times. During the latter part of August,
when the chirping season was at its height, a remarkable degree
of synchronic rhythm characterized their chirpings during
warm, quiet evenings. So constant was this rhythmic syn-
chronism that only now and then would any irregularity occur.
It finally oceurred to me that I could subject their consecutive
chirpings to a fairly accurate statistical analysis in the follow-

nner. With a tablet of paper and a pencil I made a short
horita dash for those instances when the chirpings were in
unison, and a short vertical dash when they were not in unison.
In this way I was able to record the consecutive chirpings for
certain periods of time. In order to illustrate this method, I
will give a graphic expression of the first period, which included
98 consecutive chirpings, 8 of which were not in unison

— A + e

|
Fourteen dlra eres of consecutive chirpings were re-

1*“Synchronism and Synchronic Rhythm in the Behavior of Certain
Creatures,” Ax. Nan, Vol. 51, July, 1917.
: 548

Nos. 622-623] SHORTER ARTICLES AND DISCUSSION 6549

corded in this manner, the results of which may be noted in the
following table:
TABLE I
STATISTICAL ANALYSIS OF 14 DIFFERENT PERIODS OF CONSECUTIVE CHIRP-
INGS OF A COLONY OF CRICEKTS OF THE SPECIES Cyrtoxipha columbiana

Period | Consecutive Chirp- Chirpings in Per Cent. of Chirp-
ings Recorded Unison ings in Unison

98 90 91.8

BOCONG Sr. RIE A 79 72 91.1
JE E OA oa la 85 72 84

Fourth. 23 21 91.3

i Sy ee sa oy 78 72 92.3
84 79 94

Seventh 66 59 89.3

AL -n et i el 73 69 94.5

NEER Gets ent SaN A 45 97.7

OMEN Sa A T 23 19 82.6

MOURNE i PE e al 95.7

TIO. O, 21 21 100.0

Thirteenth., A eave eek 75 74 98.6

Fourteenth... 49 48 97.9

Tie ou A 870 808 92.8

From these data it would appear that out of a total of 870
chirpings observed, 92.8 per cent. were in unison. Even grant-
ing that some errors have been made in these determinations,
it is quite evident that this observed high percentage indicates
that a remarkable degree of synchronic rhythm occurs.

I kept this particular colony of four or five crickets under
observation for a long #ime, and this rhythmic synchronism was
always very noticeable. These crickets chirp most actively just
before sundown. At this time every individual chirps briskly,
and it is not long until chirping in unison is gradually estab-
lished. This rhythmic synchronism does not take place at once,
- but becomes evident after the crickets have been chirping steadily
for some time. When this rhythmic unison is fairly esta
lished it appears difficult for the crickets to chirp otherwise, for
if there is any tendency toward asynchronous chirping, it is
quickly overcome. A re markable feature of the chirping of
these crickets is a tendency now and then for the chirping
become noticeably accelerated briefly. Even though this occurs,
the entire group keeps pace, so that the same unison is main-
tained.

The mole cricket (Gryllotalpa borealis ease is not an uncom-
mon pees grounds in this bart of the country. Its notes

550 THE AMERICAN NATURALIST [Von. LIT

are low, mellow, intermittent chirpings—gur-r-r-r-r, gur-r-r-r-r,
gur-r-r-r-r—which may be kept up almost incessantly during the
active mating season. I have never been able to observe any
definite rhythmic synchronism in the chirping of these burrow-
ing crickets. However, late in August, during the season of
1918, I attempted an analysis of the notes of two of these crickets
which were stridulating at the same time in their underground
burrows in a wet spot near Vinson Station, Virginia. These
two individuals chirped very persistently, and at times I noted
some degree of synchronism.
TABLE II
STATISTICAL ANALYSIS OF 10 DIFFERENT PERIODS OF CONSECUTIVE CHIRP-
or Two MOLE CRICKETS, Gryllotalpa Borealis

Consecutive Per Cent. of
Chirpings in Chirpings Not

eras A | oaa | ea T OO

si a le 16 65.2
DOOR + bo nck cad ae es 6 18 50

Wiehe cre es 62 | 36 26 58.0
Fourth 53 | 28 25 52.8
PUGH Oe. eee AR ee 28 16 12 57.1
Sixth 48 21 27 43.7
Seventh 37 14 23 37.8
Eighth. 47 | 18 61.7
A A ote 21 10 11 47.6
Tenth.. 46 | 26 43.4
Howa cco aah L o eae 202 52.3

From these data, which indicate that out of a total of 424 chirp-
ings only 222, or 52.3 per cent., were in“unison, it would appear
that there was no particular tendency to maintain a definite
synchronism in their chirpings.

The most remarkable instances of rhythmic synchronism I
have ever heard have been afforded by the cone-headed grass-
hopper of the species Neoconocephalus exiliscanorus (Davis).
A careful study of the intermittent notes—zeet-zeet-zeet-zeet—
of these locusts was made on the sot ofa swamp near Vinson
Station, Virginia, late in August, 191

‘The characteristic habit of ¿sed for anida of

this species is to produce a certain number of consecutive notes,
followed by a brief pause. Usually, from fifteen to thirty con-
secutive notes are delivered before the pause takes place, then
stridulation is again resumed. =. and Hebard? mention this
habit as follows:
2 Rehn, James A. G., and Hebard, Moan, **Studies in American Tet-

Nos. 622-623] SHORTER ARTICLES AND DISCUSSION 6551

The number of consecutive times without pause that this sound was
produced were on one occasion counted, 26-14-20-20-17 ; usually on a
warm evening an undisturbed singer would average about as above
before ceasing a few seconds. The song is rapid, the sounds being
emitted on warm evenings about 3 to the second.

When stridulation has become fairly established in a colony
of these locusts, for the evening, it is likely to be continuous, for
if some singers cease their notes briefly, others take it up.
Rehn and Hebard, in the publication mentioned above, have also
noted this behavior and say:

When near a colony of this species on favorable evenings after dark
the air is vibrant with the sound; as several singers cease others take up
the constantly rising and falling song, but at no very great distance the
song is inaudible.

In the colony observed by the writer at Vinson Station, Va.,
three individuals which were somewhat isolated from the rest
maintained a perfect rhythmic synchronism for many minutes
at a time, including in this period many hundreds of consecutive
notes. Now and then all three would be stridulating at the
same time, then only two would produce their notes, yet the same
perfect rhythmic synchronism was always evident. Sometimes

all but one would cease to stridulate, then one or both of the
others would again take up the rhythm with a precision that was
marvelous. It did not matter how often one or another indi-
vidual joined the chorus following a pause, the notes were always
perfectly synchronous from the start and the rhythmic syne
nism was maintained.

A representation of this perfect synchronism which was evi-
dent as the different ‘‘singers’’ took up the rhythm from time to
time may be shown graphically with dashes as follows:

A A

A A A e a a M ———-— mm — e o

On several different nights I observed the same marvelous
rhythmic synchronism in this particular group of individuals.
Although other groups were ““singing”” elsewhere, it appeared —
tigoniide : A Synopsis of the Species of the Genus Neoconocephalus found
in North America, North of Mexico,’’ in Trans. Ent. pes 40, Jan. 6,
1915, pp. 365-413.

552 THE AMERICAN NATURALIST [Vor. LIT

that their notes were delivered independently of the rhythm of
this particular group. From observations of the stridulations
of other groups in this same colony, I am of the opinion that it
is not unusual for these locusts to develop a rhythmic synchro-
nism in small groups.

It would be interesting to know why some species of locusts
and crickets possessing the intermittent habit of stridulation
tend to develop a more or less perfect rhythmic synchronism
while others do not. Although this is true of the two crickets,
Ccanthus niveus and Cyrtoxipha columbiana, 1 have been unable
to note any synchronism in the chirpings of the common arboreal
ericket, Orocharis saltator. Although large colonies of these
erickets may often be heard in stridulation, each individual
appears to stridulate in its own leisurely manner independently
of its fellows.

H. A. ALLARD
WASHINGTON, D. C.

ON THE PIGMENTATION OF A CLYPEASTROID,
MELLITA SESQUIPERFORATUS LESKE'

THE common clypeastroids, Echinarachnius and Mellita, when
adult, are characteristically of a brown or (in the former spe-
cies) reddish-brown color. This seems to be general throughout
the group. Taxonomic lists contain, however, numerous refer-
ences to a greenish coloration of the test of these animals
When preserved in alcohol, or when dried, either after fixation
in aleohol or after killing with fresh water, these sand dollars
usually assume, for a time at least, a somewhat greenish color.
Clark (1899, p. 118) says that specimens of Mellita sesquiper-
foratus Leske (=sexforis A. Ag.) collected at Jamaica were
delicate olive green [when alive, I infer, though with doubt].
He also gives the coloration of specimens of this species obtained
in Porto Rican waters as ‘‘usually light olive green (rarely
brown) when alive.”? At Bermuda living individuals of this
species are, he says, invariably brown, with no hint of green
about them, either on the external surface or in the viscera.

killed in alcohol, however, they become green, and green-

ish pigment is dissolved by the fluid. This is also true of Echi-
narachnius parma (Clark, 1904, p. 564; Coe, 1912, p. 111).

Now, examination shows that there is at the bottom of this

1 Contributions from the Bermuda Biological Station for Research, No. 90.

Nos. 622-623] SHORTER ARTICLES AND DISCUSSION 5538

matter—in descriptive lists somewhat confusing—a rather in-
teresting point, which it is the purpose of this note to elucidate.

Mellita, adult, is at Bermuda undoubtedly brown; large speci-
mens (9.0-11.5 em. in transverse diameter, usually 9.5 cm.),
which I have from time to time collected by dredging upon
grass-free bottoms of fine sand or mud at Flatt’s Inlet, Spanish
Point, Two Rock Passage, and other localities, are uniformly
brown upon both aboral and oral surfaces, although the different
individuals vary somewhat as to shade. Their general hue
harmonizes well with that of the substratum. It is improbable
that light has had a direct effect in producing pigmentation,
since the oral surface, never turned toward the light, is at least
as densely pigmented as the aboral one and is frequently (in :
larger specimens) darker. Young individuals in an active,
healthy state were gotten in association with adults during the
autumn months. Up to 5 em. diameter, in one case 8 em., they
were found, with one exception in about 30, to exhibit no brown-
ish coloration whatever; they were, on the contrary, pure white,
and translucent, the yellowish stomach being easily made out
through the test. These individuals were usually 3.5 to 4.0 cm.
in transverse diameter. The one exceptional specimen, 2 cm.
in diameter, was unusual because it was of a light coffee-brown
shade.

When placed in alcohol, or in fresh water, these young white
Mellitas became bright green; in sunlight the green on alcoholic
specimens quickly disappears. Clark (1901, p. 254) notes that
some young specimens of M. pentapora examined by him were
green [in aleohol?].

When kept in aquaria for several days the small white “‘sea-
plates’’ developed, in most cases, local indications of green pig-
ment, although the animals were still quite active. This was
also true of the large brown individuals. It was noticed that
in cases where a large brown Mellita had been damaged by the
cutting edge of the dredge, a green coloration was present along
the wound when the haul was brought to the surface. Other
specimens, apparently uninjured, sometimes developed an olive-
green color on the oral surface within half an hour after being
transferred from the dredge to a tub of sea water.

Thus the green coloration of Mellita is associated with a con-
‘dition of injury or death. It is possible that the green material
may have no connection with the substance responsible for the
general brown integumentary coloration of adults.

554 THE AMERICAN NATURALIST ` [VoL. LIT

The green pigment is readily extracted with fresh water. It
is not chlorophyll. When an animal is allowed to die in fresh
water, the integument and the ordinarily white internal parts
of the skeleton of a Mellita become bright green. Putrefactive
changes decolorize the green extract, and the color can not be
restored by alkali, or with H,O,. Green extracts are also de-
composed, irreversibly, by boiling.

The green color is not seen in faintly acid fresh-water extracts
and the Mellita remains brown. If such an extract is made
alkaline, the green color promptly appears in the extract. The
substance responsible for the green hue in dead or injured parts
of Mellita is in fact a very good indicator. It is colorless in
- acid, vivid green in alkaline solutions; this color change may be
reversed many times. The ‘‘turning-point’’ of the indicator
is at an acidity of py=7.6-7.8—in a solution more alkaline
than neutrality,. but well on the acid side of the reaction of sea
water (pa=8.1=). This greening material seems to be pres-
ent throughout the body of Mellita, as freshly secured bits of the
(white) internal skeleton turn bright green in alkali. The few
available references to the coloration of clypeastroids indicate
that the alkali-greening substance regularly occurs in Clypeaster
and in other genera of this group.

The ovaries of M. sesquiperforatus are heavily sient by
a substance of the ‘‘antedonin’’-‘‘echinochrome’’ group. The
mature egg itself is light brownish yellow, heavily yolked, and
apparently larger than any echinoid egg that has been described.
It measures about 0.26 mm. in diameter, and is thus about twice
the size of the egg of M. testudinata (=pentapora), which
measures 0.11 mm., though not so large as that of the brittle-
star Ophioderma, 0. 30 mm. in diameter (Grave, 1916, p. 439).
The ovarian egg of M. sesquiperforatus is surrounded by a
gelatinous envelope, the whole being 0.35 + mm. in diameter.
This envelope bears numerous evenly scattered clumps of pre-
cipitated reddish-purple pigment, the ovarian stroma being
densely crowded with similar ‘‘chromatophores.’’ In Echi-
narachnius the ‘‘chromatophores’’ of the egg-envelope are red
rather than purple. The purple pigment of the Mellita ovaries
exhibits the acid-alkali color changes and the absorption spec-
trum of the ‘‘echinochrome’’ pigments found in sea-urchins,
holothurians, erinoids, and even in star-fishes ; a dilute extract
of a female Mellita may ewe be prepared which changes

Nos. 622-623] SHORTER ARTICLES AND DISCUSSION 565

from bright green to reddish purple at a hydrogen-ion concen-
tration of about 7.5, the ‘‘echinochrome’’ not becoming orange
until a much greater acidity is reached. This pigment becomes
blue at an alkalinity of about p= 8.2; in the ovary it is red-
dish purple, but it is not in solution.

The interest of these facts lies not so much in their supplying
some additional instances of the elaboration of similar (or even?
identical) substances—conspicuously pigments—by animals re-
lated in descent, but in the evidence which is afforded regard-
ing the reaction of intracellular fluids. Numerous cases illus-
trating the former point are available from among echinoderms,
molluses, tunicates, and so forth, and these cases have a certain
importance for the general theory of animal coloration. But I
am here chiefly concerned to point out that, if the alkali-greening
substance present in M. sesquiperforatus is closely similar to
that produced in the tissues of other clypeastroids, unequivocal
statements as to the occurrence of greenish hues in living ‘‘sand
dollars’’ may contain a suggestion as to a possible mode of origin
for certain known geographical color differences in echinoderm
species. I have not been able to find such references, but the
point is worthy of further study. The evidence afforded by
intracellular substances capable of behaving as indicators of
acidity shows plainly-that the tissues of marine animals are
much more acid (less alkaline) than sea water. In M. sesqui-
perforatus the green color is produced when the tissue fluids
become more alkaline than they customarily are; thus the in-
tegument, when injured or killed, becomes permeable to sea
water and assumes a green hue. Healthy individuals for ex-
perimental work may, incidentally, be selected (at Bermuda)
by the absence of green areas upon the test; the readiness with
which greening is induced indicates the degree of care which
must be employed in handling some marine animals. If by
some means, for example, by higher temperature, the tissues of
a Mellita population in a warmer sea were constantly main-
tained at a higher alkalinity than those at Bermuda, they might
normally appear somewhat greenish in color. The normal varia-
tions in the coloration of Chromodoris zebra, a nudibranch con-
taining an indicator favorable for such observations, strongly
suggest that such differences in the reaction of intracellular
fluids (not necessarily of the protoplasm) are entirely possible
(Crozier, 1916). Whether or not comparable changes may be

556 THE AMERICAN NATURALIST [VoL. LII

induced as a regular thing in different oceanographic regions
can not as yet be stated.

REFERENCES

Coe, W.

h aie ek: of Connecticut. State Geol. and Nat. Hist.

, Bull. 19, 152 pp., ills.

Clark, H, L:

1899. Further Notes on the Echinoderms of Bermuda. Ann. N. Y.

Acad. Sci., Vol. 12, pp.
1901. pe Witinoderns of "Porto Rico. “Bull. U. S. Fish Comm. for

. 231-263.
1904. ri eS eee of the Woods Hole Region. Ibid., for 1902,
p. 545-576.

Crozier, W. J.
19164. Some Indicators from Animal Tissues. Jour. Biol. Chem.,
24, 3-445,
1916b. Cell Penetration by Acids. II. Further Observations on the
Blue Pigment of Chromodoris zebra. Ibid., Vol. 26, pp. 217-
223.
Grave, C.
1916. Ophiura brevispina. II. An Embryological Contribution and a
of the Effect of Yolk Substance upon Development and
Developmental Processes. Jour. Morph., Vol. 27, pp. 413-
445, 3 pls

PEMBROKE,
ERMUDA,
January, 1918.
W. J. CROZIER

A CASE OF ABNORMAL INHERITANCE IN
DROSOPHILA MELANOGASTER

AMONG great numbers of cultures one is occasionally found
which gives exceptional results not explainable by the usual
mode of inheritance. Although such cases do not aid in under-
standing genetic problems unless the mechanism involved is
worked out, the present case seems to be sufficiently remarkable
to merit brief mention. The writer has no explanation to offer.

In culture 76, which was made up February 9, 1917, a large
preponderance of males was observed, the ratio being 38 males
to 3 females, and the males included unexpected classes. The
parents of the culture were one homozygous eosin ruby forked
female from stock and a male which was normal wild-type in
all respects with the exception that the posterior eross veins of
the wings were missing. His pedigree is unknown and he was

Nos. 622-623] SHORTER ARTICLES AND DISCUSSION 591

bred to determine whether the missing vein represented a genetic
characteristic. This peculiarity probably had no relation to the
exceptional nature of the offspring produced. The bottle was
kept on the desk in the laboratory and the temperature was
rather low most of the time, so that the larve developed slowly
and the bottle became moldy before the flies finished hatching.
It yielded exceedingly few flies, probably on this account as well
as owing to the fact that nearly all the females were eliminated.

A count of the flies, as they hatched, gave 38 males and 3 fe-
males, a ratio which is inexplicable. The classes obtained were
also as surprising. Owing to the cold the flies developed very
slowly, so that the first offspring were removed on March 1 and
comprised 18 eosin ruby forked sons. Four more hatched on the
second and third, making a total of 22 eosin ruby forked sons
which are of the expected class, since the three characteristics
are sex-linked and the mother was homozygous for them. The
count, continued until the thirteenth of March, gave a total of
3 normal females, which were expected in equal numbers with
the males; 30 eosin ruby forked males; 2 eosin ruby males; 1
eosin male; 3 forked males; 2 normal or wild-type males.

The count in this case was kept up for more than the usual
10 days, but that could not have had any effect on the result in
this case as no F, females were found until March ninth and
could not have produced offspring, even had the temperature
not been so low as to lengthen the incubation period beyond 13
days. The exceptional males, which are the cross-over classes
ordinarily obtained in the F, generation from such a cross, could
not be the result of a back-cross of the original mother to a son,
as the only sons with which she could have come in contact were
eosin ruby forked. In cases of primary non-disjunction, where
sons inherit the sex-linked characters of the father, they inherit
all his sex-linked characters, so that this can not be a case ex-
plainable by that means.

Contamination can hardly account for the results as the early
males were of the expected class and later males always carried
characters used in the cross, and the females were normal in ap-
pearance. Moreover, there is no known source of contamination
that would give such a sex-ratio as this.

Results from the offspring were interesting but have not sug-
gested any possible explanation of what occurred in the first
generation. The eosin ruby forked sons were crossed out and

558 THE AMERICAN NATURALIST [Vou. LIL

behaved quite normally in F, and F, generations. One forked
son bred and gave offspring which behaved normally. The other
two forked sons failed to produce any offspring, even though
transferred to new bottles. The two eosin ruby sons were mated
to bar females and afterwards rebottled with three females, but —
no offspring resulted. The eosin male also seemed to be sterile,
as he was rebottled and remated without producing offspring.

The three daughters, which should have been heterozygous for
eosin ruby and forked, and which have produced sons correspond-
ing to that constitution, were crossed to brothers and gave the fol-
lowing unexpected results:

One in culture 88, mated to an eosin ruby forked brother, pro-
duced a total of 42 normal females, 74 forked females, 73 forked
males, 37 normal males, and 13 eosin ruby forked males. A
second in culture 93, mated to one of the wild-type brothers,
produced 50 normal females and 28 forked females, 51 normal
males and 73 forked males. The third daughter, mated to an
eosin ruby forked brother, produced 2 normal, 35 forked, and 10
eosin ruby forked males; and 26 females all forked.

It is possible that the females were not virgin in these cases
but that could not affect the sex-linked characters of the sons
according to the normal mode of inheritance.

The two large classes of sons should have been the normal and
eosin ruby forked classes, while the forked class of sons, which is
the largest in all cases, should be no larger than the eosin ruby
class, which does not occur even once.

Efforts to determine what was causing this abnormal inheri-
tanee were unsuccessful, because further breeding experiments
showed the offspring of all classes to behave quite normally in all
respects.

D. E. LANCEFIELD
CoLUMBIA UNIVERSITY

INDEX

NAMES OF CONTRIBUTORS ARE PRINTED IN SMALL CAPITALS.

ADAMS, CHARLES C., Migration as a
Factor in Bvolntion: its Ecolog-
ical Dynamies, 455.

ALLARD, H. A., Rhythmic Synchro-

Certain

þe-
tween Color and other Characters
i . LOVE and W. T. Cra

in,

BABCOCK, ERNES
Factor

T B., The Rôle of
etei in Evolution,

Babcock and Claussen on Genetics in
Relation to Agriculture, E. M. E

Bacterial Phylogeny as indicated by
Modern Types, R. E. BUCHANAN,

Blue Andalusian, The Case of the,
WILLIAM A, LIPPIN

BREGGER, T.,
Pestle Modifie
a Sexinked EREA D;

46

as, gi E, Bacterial Phy-
se as indicated by Modern
Types, 233

Cancer?s Place in General Biology,
W. C. MoCarty, 395
Coloration of Planes minutus, W. J.

Coral Reefs, the Hawaiian. A Sur
me of, VAUGHAN MACCAUGHEY,

ZIER, W. J., or of Planes
tus, 262; A Land Planarian

ypte
sesquiperforatus Leske, 5:
Chromosomes, Disproof of a Certain
Type of Theories of Shwe. “wee
between, H. S. JEN
Craic, W. T. and VE, The
Relation between Color and other
Characters in Avena Crosses, 369

DEXTER, JoHN S., Inheritance in
Orthoptera, 61

TAT Melanogaster, A T of
Abnormal Inheri in
LAN 556

E. M. E., Babcock and Claussen on
Geneties in Relation to Agricul-

, The Rôle of Reproduc-

in Evolution, 273

Egg Prodnetian in the Rhode island

Red Breed of Domestic Fowl,
era Abe ci e ing, H.
, 65, 209,

Evolution, “The Rôle pu Factor Mu-
tations in, ERNES . BABCOCK,
116; Evidence from Insular Floras
as to the Method of, EDMUND W

Aphid Macrosiphum solanifolii, A
N SHULL, 507

Factors for Yellow in Mice and
Notch in Drosophila, WILLIAM A.
laige gorr 3

MARGARET M., The Uses of
Tnsect Galls, 155

Fish, The Policy of eee a
Inland “Waters with, W. M. Sm
WOOD, 32

GOLDSCHMIDT, RICHARD, Gen etie Ex-
periments concerning Evolution.

, H. L., Internal Factors in-
in

he
f Do-

©
au
7 8
EN
=

mestie Fowl, 65, 209, 301
Hot, THEO., Joan Baptista Porta,

Hybrids in Egyptian Cotton, THOMAS
KEARNEY and WALTON G. WELLS,
491

Inheritance of Number of Sagi
of the Fantail Pigeon, T. H. Mor

GAN, 5; in Orthoptera, “Jomx S.

E. | Insect Galls, The Uses of, MARGARET
M. Fagan, 155

559

X

560

ie. H. S., Disproof of a Cer-
tain Type of Theories of Crossing-
over between Chromosomes, 247
KEARNEY, THOMAS, and WALTON G.
WELLS, Hybrids in Egyptian Cot-
ton, 491

ELD, D. E., Three n

462; A Case of Abnormal Taher’.
tance in Drosophila Melanogaster,
6

LAUGHLIN, H. Modifications of
353

ue An
Factors for Yellow
Notch in Drosophila, 364
. T. Crate, The
een Color and other
Characters in Avena Crosses,
McCarty, W. C., Cancer’s Place in
General Biol 395

Love, H. H. and
Relation betw

n, H, TERAO, 5.
Moopiz, Roy L., Opisthotonos and
MORGAN, T. H., Inheritance of Num-
sae Feathers of the Fantail
Mutations in Previously Known
Loci, D. E. LANCEFIELD, 264

pe tico and yem Phenomena,

Ei <a Evolution “and cha Signifi-

Some Lo

orate on the. Problem E e
WALTON, 521

ical Problems in the Life

THE AMERICAN NATURALIST

[VoL. LIT

Pigmentation of a Clypteastroid,
oe ee sesquiperforatus Leske, W.
J

Planarian, A Land, found at Ber-
muda, W. CROZIER , 272
Porta, Joan Baptista, ridad HoLm,
455

Ratio, Modifications of the 9:3:3:1,
H. H. LAUGHLIN, 353

SHELFORD, VICTOR E., Physiological

Problems in the Life Histories of

ith Particular Refer-

Te] to their Seasonal Occurrence,
12

SHUL: , Genetic Rela-
tions of va Winged and Wingless
rms other an
Sexes in a gone Macrosiphum
solanifolii, 50
SINNOTT, rbd W., Evidence from
Insular Floras as to the Me thod
of Evolution,
Sm 00D, W. M., The Policy of
making the Inland Waters
with
SUMNER LA B., Continuous and Dis-
contin ariations and their
Tokasiikkso in Peromyscus, 177,

290, 439

Synchronism, Rhythmic, in the
Chirping of Certain Crickets and
Locusts, H. A, ALLARD, 548

TERao, H., Maternal Inheritance in
the Soy 1 Bean, 51

Variations, Continuous and Discon-

thei

Watton, L. B., Organie Evolution
and the Significance of Some New
Evidence bearing on the Problem,

WELLS, WALTON G.

KEARNEY, Hybrids ‘in Egyptian
Cotton, w